East European Politics and
Societies and Cultures
Volume 29 Number 1
February 2015  61­–80
© 2015 Sage Publications
10.1177/0888325414566072
http://eeps.sagepub.com
hosted at
http://online.sagepub.com
Hurricane Season:
Systems of Instability in Central
and East European Party Politics
Tim Haughton
University of Birmingham
Kevin Deegan-Krause
Wayne State University
The seemingly random triumph and demise of new political parties in Central and
Eastern Europe actually represent a durable subsystem with relevance for party systems
around the world. This article supplements existing research on volatility with new measures
of party age distribution that reveal clear patterns of disruption, turnover and restabilization.
These patterns emerge from stable and coherent party subsystems that follow
a simple model based on three dynamics: losses by established parties, rapid gains by
uncorrupted newcomers, and equally rapid newcomer losses to even newer parties.
Confirmed both by electoral evidence and computer simulations, this model offers insight
into the endurance of these subsystems, particularly since the very mechanisms that generate
new parties’ success can preclude their ability to survive in subsequent elections.
Central and Eastern European party systems offer a laboratory for understanding trends
in party system volatility that are emerging in Western Europe and across the globe.
Keywords: volatility; new parties; party systems; party subsystems; Central and
Eastern Europe
For the moment, the byword for political disaster in Central and Eastern Europe
is “earthquake.” The list of elections described in seismic terms is long and
growing: Bulgaria (2001), Poland (2001), Hungary (2010), the Czech Republic
(2010 and 2013), and Slovenia (2011 and 2014). Elections like these can involve
swings of more than 40 percent and give pluralities to parties still in their infancy.
Even “normal” elections bring tremors as parties come and go, merge and splinter.
As metaphors go, “earthquake” has certain advantages. There are few better
images of sudden, massive disruptions and little expectation of precise predictions.
Instead, both seismologists and political scientists gather evidence on deeper underlying
factors to identify hotspots and generate likelihoods.
But there is need for a new comparison. Earthquakes involve large and coherent
structures in tension with one another whereas party politics consists of multiple,
independent institutions with their own internal structures. Major shifts in voting
566072EEPXXX10.1177/0888325414566072East European Politics and SocietiesHaughton and Deegan-Krause / Hurricane Season
research-article2015
From the Editors: This article is a part of special section “Political Parties in Eastern Europe” guest-edited
by Kevin Deegan-Krause. Two of the four articles in this section, including the present one, were first
presented as papers at the “Whither Eastern Europe?” workshop, co-sponsored by EEPS, ACLS, and the
University of Florida. The workshop took place in January 2014 in Gainesville, FL, and was hosted by
Michael Bernhard, Miriam and Raymond Ehrlich Chair in Political Science at UF. A collection of papers
from the workshop will appear as a special issue of EEPS in May 2015 (volume 29 number 2).
62  East European Politics and Societies and Cultures
patterns depend not only on structural clash but also on the actions of individual
political parties in direct response to one another. The aim of this article is to advance
the state of research on rapid political shifts with improved measures of party system
change and a new conceptual model that accounts for the rapid changes found in
Central and Eastern Europe (CEE) and other regions as well.
Our new measures confirm current findings that political shifts in CEE depend
largely on the entrance of new parties and the exit of established ones and go beyond
that to demonstrate significant differences in behavior between newer and older parties,
at both the elite and voter levels. Political dissatisfaction sometimes produces
flows of voters and elites among established parties, but in CEE the flow is often to
new parties. These new parties have more fragile political appeals and organizations
and they are even more likely to suffer desertion than their more established counterparts.
Since most of those who give up on new parties do not return to the more
established parties but instead move on to other new alternatives, the result is a selfreinforcing
subsystem of ever newer parties that endures until parties succeed in
adopting stabilizing mechanisms. This subsystem model of overall party system
dynamics in the Central and Eastern European democracies explains both the variability
in the timing of electoral shifts and the persistence of these internally volatile
subsystems once they become established. The model also provides us leverage for
understanding potential dangers and evaluating potential remedies, which is particularly
important because the changes in CEE appear to be harbingers of a change in
party systems that is happening on a much wider geographical canvas.
Beginning with a discussion of current variables on party system change, the article
goes on to provide new data on party age and introduce the variable of age distribution.
The article finds an increasingly pervasive pattern of stable, older parties and a subsystem
of newer and less stable parties engaged in accelerating party-level cycles of birth,
death, and replacement. The final section of the article speculates on the proximate
explanations of and alternatives to these subsystems and their broader relevance.
Measurement: Volatility and Weighted Age
The most widely used tool to measure the degree of party system change—the
Richter scale of electoral earthquakes—is Pedersen’s index of volatility, an easily
operationalized quantitative measurement of the extent of change between elections.1
But like the Richter scale, which is no longer regarded as a sufficient measure,
Pedersen’s index can conceal as much as it reveals. As an aggregate measure,
not only does it ignore individual-level party switches that cancel one another out
but it also conflates shifts of voters within a stable institutional core of parties with
shifts of voters necessitated by the institutional comings and goings of parties themselves.2
Avoiding this conflation, Powell and Tucker find that volatility among
established CEE parties has been smaller than the extra-system volatility associated
with the appearance and disappearance of parties. But even their revised measure—
a huge step forward in systems with fluid political institutions—offers only a partial
Haughton and Deegan-Krause / Hurricane Season  63
resolution because it does not account for the component parts of the system or
allow simultaneous comparisons across multiple elections. Their measure recognizes
that the newness is important and offers snapshots of who goes in and out of
the main doors but it cannot tell us (except by accident) how long they have
remained inside.
Kreuzer and Pettai’s study of parties in the Baltic states seeks to remedy this gap
with the Weighted Party Age Index (WPAI), a supplementary measure of institutional
novelty that sets aside questions about shifts among parties and looks instead
at the length of time since a party’s establishment.3 To the extent that “newness” is
important, as many scholars including Sikk, Mainwaring, Bolleyer, and Powell and
Tucker4 agree, Kreuzer and Pettai’s measure provides a useful indicator of the relative
newness in a system. Figure 1, applying the WPAI to ten CEE countries, indicates
extremely wide variation ranging from Romania, whose party system grew
older by roughly one year for every year of its existence, suggesting extreme stability,
to the Baltic states, where constant replacement and reinvention gave the party
systems a perpetual youth. In the beginning of the 2010s, however, the values began
a mild convergence. The graph shows that beginning in 2008, the average age range
between the highest and lowest stopped growing and actually began to decline as the
older systems of Romania, the Czech Republic, and Hungary suddenly experienced
a drop in average age, while the parties of the Baltics began to survive elections.
Figure 1
Weighted party average age for Eastern and Central Europe, 1990–2014
Source: Authors’ calculations are based on M. Kreuzer and V. Pettai, “Persistence and Decline versus
Transience and Genesis” (paper presented at the Joint Sessions of the European Consortium for Political
Research, St. Gallen, Switzerland, 2011); S. Berglund et al., The Handbook of Political Change in
Eastern Europe (Cheltenham, UK: Edward Elgar Publishing, 2013); and Political Data Yearbook:
Interactive, n.d., http://politicaldatayearbook.com/ (accessed 10 July 2014).
64  East European Politics and Societies and Cultures
From a mathematical perspective, average age and extra-system volatility should
be related, at least at the top and bottom of the scales,5 but between these extremes
there is no direct correspondence, since new parties may replace old parties and produce
a large drop in age, or they may replace new parties for only a slight shift.
Empirically, the average age in Central and Eastern Europe tends to move with extrasystem
volatility and the correlation for each country is positive (more volatility
correlated with younger systems), but the two are not identical, and the correlations
for individual countries range between 0.2 and 0.8.
Classification: Patterns of Disruption, Turnover, and
(Re)stabilization
Since summary measures can only tell us so much, it is useful to turn to richer
but less simply summarized data—the actual distribution of parties according to
age in any given electoral period for any given country. The graphic depiction of
party age in Figure 2 divides the share of votes in each election according to the
period in which a party first appeared. An entirely black column represents an
election in which all parties receiving votes were founded during the first election
under consideration. An entirely white column represents a system in which all
parties receiving votes emerged in the most recent election. Shades of grey represent
parties of intermediate age; the darker the shade, the earlier the party was
founded.
These data reveal additional distinctions. The countries in the middle of the average
age range actually demonstrate quite significant differences in configuration. In
2013, both Poland and Bulgaria produced an average age of 11 years, but the Polish
system consisted largely of 11-year-old parties whereas Bulgaria’s party system
average resulted from a combination of some parties more than 20 years old and
others less than 5 years old. Poland’s big changes were systemwide: nearly the
entire party system changed radically in the early 2000s as early-generation parties
were replaced by middle-generation parties and subsequently developed a degree of
renewed stability and fewer newcomers. Bulgaria’s changes cut the system in half:
one segment of the party system remained stable while the other half experienced
constant churn.
While the visual distribution is a useful technique for quickly grasping the differences,
it does not permit easy quantitative comparisons between systems. It does,
however, identify the variable for which a quantitative measure is necessary: the
relative weight of parties from various periods. Fortunately, measurement of distribution
is well established, and we adapt a summary measure for distribution developed
by Esteban and Ray6 as an improvement to basic measures of polarization of
income. Because this measure incorporates the relative weights and distances of all
data points, it is responsive to widely spaced peaks (bi- and multimodal distributions)
and produces higher values as the peaks move away from one another.
Haughton and Deegan-Krause / Hurricane Season  65
The framework formula as proposed by Esteban and Ray is
P y T I a y y
i
n
j
n
i j i i jπ π π π δ, , ,( ) = ( ) ( )( )( )
= =
∑∑1 1
which can be adapted to the measurement of party age distribution as follows:
Figure 2
Share of vote for parties in each election according to age of party founding
(darkest = oldest)
Source: Authors’ calculations are based on M. Kreuzer and V. Pettai, “Persistence and Decline versus
Transience and Genesis” (paper presented at the Joint Sessions of the European Consortium for Political
Research, St. Gallen, Switzerland, 2011); S. Berglund et al., The Handbook of Political Change in
Eastern Europe (Cheltenham, UK: Edward Elgar Publishing, 2013); and Political Data Yearbook:
Interactive, n.d., http://politicaldatayearbook.com/ (accessed 10 July 2014).
66  East European Politics and Societies and Cultures
the total share of vote held by parties first appearing in election i πi( ) * Total share
of vote for parties first appearing in election j πj( ) * Duration of the period (in
years) separating i and j, summed for all i and j.
The panels in Figure 3 graphs the weighted party age index against the age distribution
measures for each country at each election, with panels highlighting specific
patterns. These patterns are summarized without country detail in Figure 4.
Figure 3A shows the stability path, in which early-generation parties endure and
few new parties emerge, resulting in lines covering significant vertical distance and
tending toward the low-polarization side of the graph. As might be expected from its
high average age and the absence of major new entrants, the Romanian party system
progresses straight upward with age, increasing by four years in each four-year
Figure 3
Weighted party system age and age polarization over time in Central and
Eastern European party system
Haughton and Deegan-Krause / Hurricane Season  67
period, and polarization remains low because all parties fall on the “old” end of the
spectrum. Hungary, the Czech Republic and Slovenia followed a similar upward path
until the early 2010s, with systems composed almost entirely of old parties. Bulgaria
followed this path until 2001.
Figure 3B shows the disruption path, with upward movement stopping or even
reversing and significant movement toward the right side of the graph, as new parties
emerge to rival enduring older parties and produce a bimodal distribution of ages.
The disruption lines are most pronounced where party systems have low initial age
distributions. The panel clearly shows the impact of the Popular Party of Dan
Diaconescu in Romania; that of Jobbik and Politics Can Be Different in Hungary;
Figure 4
Schematized patterns of party age and party polarization
68  East European Politics and Societies and Cultures
PublicAffairs and TOP ‘09 in the Czech Republic; and Positive Slovenia and Virant’s
List in Slovenia, each of which disrupted the uniform aging of the party system and
split the party system between a substantial but diminished group of old parties and
a smaller but significant group of new parties. The same pattern is clear as early as
2001 in the Bulgarian party system, giving a sense of the overwhelming change created
by the Movement of Simeon II (NDSV).
Figure 3C shows the turnover path in which new parties yield to even newer parties.
The partial-turnover configuration appears in the center right of the graph,
reflecting the shift from new parties to newer ones in systems that also retain significant
numbers of older parties. The full-turnover configuration appears in the lower
left of the graph, where the replacements across the board keep both age and polarization
relatively low. Bulgaria offers a representative example of partial turnover:
support for the new NDSV gradually evaporated over two election cycles, but other
new parties emerged to replace it, so both the average age and the bipolar age distribution
stayed essentially unchanged. Slovakia’s political party system after 2006
shows a similar pattern, with new parties emerging to supplant old ones on the country’s
right and in its Hungarian minority. Recent elections in the Czech Republic,
Hungary, and Slovenia likewise exhibit partial turnover as some or all of the new
parties faded and newer parties emerged with similar levels of support. Latvia and
Lithuania come close to full turnover thanks to elections in which new parties of
previous elections lose to even newer entrants, resulting in an essentially new party
system every four to eight years. Along with their low average age, these systems
exhibit a narrower distribution of age between young and old simply because few
parties survived to enjoy old age.
Finally, Figure 3D shows the (re)stabilization path in which turnover stops and
new parties survive. The paths of Poland, Slovakia, and Estonia (and potentially
Lithuania and Latvia) show what happens when new parties survive their early years
and allow party age to grow again. These cases are relatively few in number, however,
and as the example of Slovakia shows, the restabilization proves to be merely
an interlude before the resumption of turnover.
Conceptualization: Systems, Subsystems, and Cycles of
Reinforcement
As Kreuzer and Pettai observe, “instability can have patterns just as distinct as
those of stability,”7 and the previous section offers evidence both that the apparently
random electoral upheavals in CEE have a discernable, quantifiable shape and that
the resulting patterns are fairly stable over time and across the region. The tendency
of the same patterns to emerge and persist across the region—even in party systems
once thought immune8—suggests that rapid political changes in CEE may benefit
from a new metaphor. The self-sustaining, cyclical patterns we see here are less like
Haughton and Deegan-Krause / Hurricane Season  69
earthquakes than they are like hurricanes. Within the broad framework of the party
system, the ever-rotating roster of newcomers has systemic characteristics in its own
right and may usefully be understood as a subsystem.
Sartori’s understanding of “systems” provides a useful starting point. His work
defines them as “bounded, patterned and self-maintaining interdependencies”9 and
sets two basic conditions: “(i) the system displays properties that do not belong to a
separate consideration of its component elements and (ii) the system results from,
and consists of, the patterned interactions of its component parts, thereby implying
that such interactions provide the boundaries, or at least the boundedness, of the
system.”10 He also sees systems as nested, and refers to parties as constituting a
“party subsystem” within the broader political system. While Sartori does not refer
to subsystems within party systems, his conceptualization opens space for such entities,
as long as they fulfill the same conditions: significant relational elements (which
make the whole greater than the sum), patterns, and boundedness.
Within the literature on political parties, the notion of a “subsystem” has made occasional
appearances to describe the regular interactions among groups of parties sharing
a common programmatic (and fairly distinctive) ideology and voting base. Hanley uses
it to describe the “semi-permanent constellation of parties” around the French Socialist
Party that he refers to as the “plural left.”11 Strmiska introduces the concept to describe
regard to parties in ethnoregional enclaves—Spain’s Basque and Catalonian regions,
Italy’s Sardinia and South Tyrol, Montenegro in Serbia-dominated Yugoslavia—which
compete primarily against one another and without much reference to parties in the rest
of the country.12 These usages fit neatly into Sartori’s framework.
The use of “subsystem” to describe the new-and-newer party phenomenon of turnover
requires a bit more explanation. Unlike the “plural left” or “ethnonational” subsystems,
the new-and-newer parties do not always compete directly against one
another, since new parties often do not emerge in the space until a previous incarnation
is in severe decline. The patterns of competition thus occur not in party space but
in party time, an intertemporal dimension that is more akin to the notion of “system”
as used in discussions of “systems of succession” that establish patterns by which
political offices change hands.13 The notion of a “new party subsystem” thus combines
the boundedness of the “minority” subsystem discussed above—multiple parties
sharing a common and distinct pool of ideas, voters, and elites—but the whole
pattern becomes visible only with multiple observations that reveal the relationship
between the decline of one party and the emergence of another as more than a series
of unrelated parties. The use of the subsystem concept is particularly useful because
the intertemporal change of names and personalities obscures continuity in the appeals
and electorates of these parties. Treating the succession of similar but distinct parties
as a subsystem is a way of capturing the elements that resemble elements of a single
party without making the conceptual stretch of actually treating a succession of new
parties as a single undifferentiated unit. It thus neatly fills the conceptual role of “system”
by describing something that is neither a unitary actor nor a random collection.
70  East European Politics and Societies and Cultures
Quantifying the New Party Subsystems
There are several good reasons for believing that new parties across the region
exhibit specific patterns of interaction and continuity in voting base and programmatic
position that fits the characteristics of a subsystem. Unfortunately, the main
patterns of intra-subsystem flow lack conveniently comparable measurements that
would allow comparison of these developments across the region. At the level of
programmatic offerings, we are fortunate to have numerous surveys of experts,
elite attitudes, and party programs, but these surveys consistently omit the programmatic
positions most relevant for the subsystem: a focus on corruption.14 Its
omission is understandable since it is only recently that observers have come to
see the question of corruption as something more than a valence issue on which
all parties agree, leaving voters to evaluate their sincerity and capacity. In CEE,
the question of corruption more closely resembles a genuine, if sometimes vague
dimension of programmatic competition in which the “uncorrupted but inexperienced”
challenge the “tainted but experienced.” Without systematic measures,
however, we can hope at best for qualitative assessments of party positions or rely
on voters’ perceptions or the desires of party voters to give us a sense of parties’
position on this dimension. At the level of voters, we have numerous surveys of
party preferences but almost no panel data that tracks how voters change their
minds over time. Again the best we can find is indirect data: ecological analysis of
vote flows in local communities, and retrospective survey questions in which
respondents specify their current and past electoral choices, and even these are
surprisingly rare.
Although limited, currently available evidence give strong support to the newparty-subsystem
pattern in CEE countries:
•	 Bulgaria: Arguably the first country in the region to experience a full-scale newparty
subsystem, Bulgaria experienced the sudden emergence of the new National
Movement of Simeon II (NDSV) in 2001 followed by its collapse and the emergence
eight years later of the newer Citizens for the European Development of
Bulgaria (GERB). Bochsler notes clear “voter shifts in 2009, when GERB is the big
beneficiary of previous antiestablishment votes, winning a substantial part of previous
Ataka votes [a radical right party that emerged in 2004] and of previous Simeon
votes.”15
•	 Czech Republic: Both survey and electoral data show strong links between the
“new” parties in 2010—TOP ‘09 and Public Affairs (VV)—and the “newer” parties
in 2013—Action of Dissatisfied Voters (ANO) and Dawn of New Democracy
(Úsvit). A significant majority of votes from new parties in 2010 either remained
with those new parties, left the electorate, or flowed to the even newer parties of
2013, whereas a significant majority of “old” party voters remained within the
“old” camp. It is significant that preelection opinion polls and ecological inference
based on election results find almost the same results: patterns of voter outflow
show about 65 percent of “old party” voters staying with old parties; while among
Haughton and Deegan-Krause / Hurricane Season  71
voters of 2010’s new parties, only about 30 percent stayed while another 10–20
percent left the electorate and 30–40 percent opted for a newer party.16 Likewise,
patterns of voter inflow show that newer parties took 30 percent of their voters from
2010’s new parties and 15–30 percent from previous nonvoters.17
•	 Slovenia: Preliminary evidence from the 2014 election indicates that a disproportionate
share of those who voted for new parties in 2011 subsequently voted for
even newer parties (about 60 percent for the Party of Miro Cerar and about 10
percent for the United Left). Voter inflow shows the same patterns, with about half
of the newer parties’ votes coming from the parties that had been new in 2010.18
•	 Lithuania: Bochsler finds that in 2008 the new Party of National Resurrection (TPP)
gained votes across the board both from strongholds of older parties and those of
newer ones.19 When TPP folded, however, many of its voters did not return to the
old parties but opted instead for yet another newcomer: Way of Courage (DK).
Raimonte’s analysis of retrospective survey data shows that a disproportionate
share of voters from the new party voters opted for the newer party, and that the
two-thirds of DK voters came from TPP (24 percent), from the ranks of nonvoters
(34 percent) or from the ranks of other small parties (12 percent).20
In addition to evidence from these high-volatility elections that reshaped entire
party systems, there is also evidence of smaller, ongoing cycles of “new and newer”
in which a series of “clean” parties with a particular programmatic orientation
replaced otherwise similar parties with more unsavory histories. Even before the
major shifts in 2010, the Czech Republic had already experienced a cycle of party
replacement among parties occupying the moderately pro-market, social-liberal
space: the Civic Democratic Alliance was supplanted by the Freedom Union,
which was in turn supplanted by the Greens, which bequeathed many voters to the
new parties of 2010.21 In Slovakia, retrospective analyses of voting show significant
shifts in votes from the new party the Alliance of the New Citizen to the newer
party Free Forum between 2002 and 2006 and again from the new party Freedom
and Solidarity (SaS) to the newer party OLaNO between 2010 and 2012. Recent
evidence suggests a similar flow from OLaNO to the newly created party Net
(Siet’).22
Modeling the New Party Subsystem
As direct empirical evidence for the existence of voter flow in new party subsystems
mounts, it is also possible to demonstrate that outcomes in CEE party competition
are consistent with the results that a subsystem would produce. In the subsystem
described above, disillusioned individual voters opt for new parties, become disillusioned,
and then opt for still newer parties, but it is precisely these kinds of longterm
individual-level tracking data that we lack. It is possible to reformulate this
subsystem to describe party outcomes rather than individual behavior. The resulting
party-level subsystem can be described with three sets of dynamics:
72  East European Politics and Societies and Cultures
1.	 Parties sometimes face the loss of significant numbers of supporters.
2.	 New parties are disproportionately likely to benefit from losses by more established
parties.
3.	 The newer a party is, the more likely it is to suffer rapid, significant losses of sup-
port.
Figure 5 presents these three rules as a flow chart showing the cascade effect created
when all three rules are in operation.
It is a simple matter to use these three dynamics to create a basic computer model
of the emergence of a new-party subsystem that exists parallel to but distinct from
Figure 5
Flowchart of support flows according to new-party subsystem dynamics 1–3
Haughton and Deegan-Krause / Hurricane Season  73
the established parties (which together could be conceived of as an “established party
subsystem,” though its parties have little in common except their greater age and
reluctance or inability to use the appeals based on corruption).23 Simulations based
on the three dynamics (with small amount of random chance shaping the initial relative
sizes, and timing of the gains and losses, growth, and decline of individual parties)
produces patterns indistinguishable from the empirical data above not only in
terms of volatility but also party age and even distribution of party age. Not only do
the four randomly chosen simulations in Figure 6 (panels A–D) follow the hypothesized
pattern, but each one comes quite close to resembling a particular country in
CEE. The dominance of third election parties in Simulation A (while some first election
parties endure and new parties continue to replace one another) bears a striking
resemblance to Poland and to Slovakia, as does Simulation B (though the major
disruption of a new party in the eighth election bears closer resemblance to recent
developments in Slovenia and the Czech Republic). The rapid decline of first and
second election parties and subsequent periodic replacement of much of the party
system in Simulation C resembles Latvia and Lithuania. Simulation D, produced
with the exactly same dynamics, gives evidence of what can happen, as in the
Romanian case, when the first-generation parties do not give way.
These results, furthermore, are highly dependent on all three dynamics working
together. Any other combination produces different results. The final three panels in
Figure 6 show the results (again the first of a series of random iterations) resulting
from incomplete combinations of the same dynamics. Without the outflow from old
parties of dynamic 1, Simulation E leaves voters static and produces no significant
volatility (a model common to much of Western Europe during the postwar period).24
Without the flow to new parties of dynamic 2, Simulation F produces a simple backand-forth
cycle between established parties E1 and E2. Without the new party fragility
of dynamic 3, Simulation G produces a rapid crowding of the party system with
parties of all ages rather than the sharp polarization of age distribution that is characteristic
of actual cases in CEE.
Escaping the new party subsystem.  The subsystem model above is useful not only
because its outputs bear close resemblance to the existing party system dynamics but
also because it gives us a more precise frame of reference for understanding alternative
paths. The model describes party politics in Lithuania, Bulgaria, or the Czech
Republic in 2013 better than in Estonia, Romania, or the Czech Republic in 2003.
Such variation is to be expected. The point of the model is not that the dynamics are
inevitable but that circumstances make them more likely than alternatives. These
subsystems are shaped by actors’ choices, and if underlying incentive structures
change (or if actors perceive them to have changed), the model’s paths may become
less appealing, alternatives may prevail, and the cycle may slow or stop. The exceptions
thus help to illuminate the contours and limits of the rule. The dynamics of the
model frame the types of possible deviations.
74  East European Politics and Societies and Cultures
First, instead of losing significant support, some established parties may retain
voter loyalty. Many parties demonstrate that ephemerality is not the destiny of all
parties in CEE: nearly all parties in Romania, Czech Republic, Hungary, and Slovenia
until the early 2010s as well as segments of the Bulgarian left, the Hungarian centerright,
and ethnic-minority communities endured. These parties did not always avoid
defeat but they managed always to recover: in the garden of party systems, they are
the hardy perennials that can endure political winters that short-lived annuals cannot
survive. Preliminary evidence suggests that these parties’ abilities to enhance their
endurance depends on three strategies: investing in organization, becoming a standard
bearer on a major issue divide of programmatic competition, and finding ways
to balance benefits of a vibrant leader with mechanisms for shedding leaders who
have become net liabilities. These strategies work,25 but they have unappealingly
high upfront costs, requiring significant expenditure and limiting party flexibility in
Figure 6
Simulations of party system age distribution using new-party subsystem
dynamics 1–3 (dynamics held constant, specific party birth, death, growth,
and decline subject to randomization, first simulation results chosen, none
excluded)
Haughton and Deegan-Krause / Hurricane Season  75
the short run. Nor do these choices bring a guarantee of survival. The collapse of
some well-established parties after the 2008 economic crisis reinforce Roberts’s
view that electorates hold parties hyperaccountable,26 and voter disappointment can
become desertion when poor performance is accompanied by spectacular mistakes:
a prime minister’s taped secret admission that “We lied in the morning, noon and at
night” (Hungary in 2006) or the corruption-related arrest of a top administrative aide
who is also the prime minister’s mistress (the Czech Republic in 2013).
The new-party subsystem may also be avoided if newly unattached voters do not
flow en masse to new parties. Some of the barriers to new party success are well
known in the literature: party rules, electoral system, and party funding rules27 and
have been in evidence in Estonia, Hungary, and Romania. Electoral systems
(Hungary) and electoral law (Estonia and Romania) impose significant legal or psychological
deterrents on those who would like to start new parties, but as the experience
of Hungary suggests, these may simply delay new party creation rather than
prevent it outright. It is also possible—though so far it seems surprisingly rare—that
a string of failures by successive new parties may cause many in the subsystem electorate
to move to inactive status and stay there until something can recapture their
enthusiasm. The flow of voters to the new party subsystem may thereby leak away
after three or four cycles, though there is nothing to prevent another subsystem from
emerging in other segments of a country’s political system where voters are less
jaded by new-party failure.
Finally, new parties that do break through are not necessarily condemned to a
short life as they can adopt the voter-maintenance strategies mentioned above. But
the circumstances under which new parties emerge are often directly, even intrinsically,
opposed to those strategies. The rapid startup period of many new parties only
allows for rudimentary organizational structures, and electoral success exacerbates
the problem since winners must divert their already small membership core away
from local- and regional-party building toward filling new staff positions in parliament
(and the problem is far worse for new parties participating in government).28
Furthermore, the anticorruption appeals that propel new parties are difficult to sustain
after electoral success. In an environment in which voters expect politicians to
be corrupt, successful new parties may lose their anticorruption credentials (and
again the problem is far worse for those who enter government). In addition, leaderdriven
parties tend to fall as quickly as they rise. The celebrity leadership that allows
some new parties to jump rapidly into contention becomes a liability when party
leaders prove corrupt, erratic, or ineffective and the infusions of outside cash that
allow some new parties to emerge may make the new party visibly reliant on donors
and therefore even more prone to corruption allegations. Finally, the young and
uncommitted voters who are most available to new parties are also the same voters
who are likely to change their mind before the next election.29
In countries where the cycles have stopped, the shift can sometimes be traced to
particular new parties that have made deliberate choices to eschew short-term
76  East European Politics and Societies and Cultures
gains and invest in party organization and effective ideological positioning (away
from short-lived anticorruption appeals). Some benefit, too, from circumstances
that make the hard choices easier. Among the survivors are several parties that
failed to make it into government on the first try, found themselves in opposition
before subsequent elections (or even maneuvered themselves into an opposition
role). It is therefore little surprise that few new parties—perhaps Smer in Slovakia
is the best recent example—have made a successful transition from novelty to
establishment.30
We have sketched out here the characteristics of parties, party systems, and subparty
systems in CEE. Lack of space precludes a systematic analysis of why some
parties and party systems show a greater or lesser capacity to escape the new party
subsystem predicament. Further empirical research is needed, but our research highlights
a number of propositions worthy of more detailed and thorough analysis linked
to three key dimensions: appeals, organization, and leadership. First, the greater the
preponderance of the parties in the subsystem to rely on ephemeral appeals linked to
novelty and anticorruption, the less likely that the party can escape from the subsystem
predicament. Moreover, this trend is only exacerbated if the parties trumpeting
their novelty are thrust into government. Novelty does not last forever, and few parties
prove to be whiter than white angels when accorded the trappings and temptations
of power. Second, the less parties in the subsystem engage in party-building
activities, particularly developing and nurturing local party branches, the more likely
that the party subsystem will endure. Third, the more parties are dependent on their
leaders and do not develop mechanisms to facilitate a (relatively) smooth passing of
power without the entire edifice of the party crashing down, the harder it is to break
out of the cycle of new party birth and death.
Extrapolation: The East and the Future of Party Systems
If the parties of Central and Eastern Europe seem always one step away from
disaster, it may provide at least a little comfort to learn that the rapid changes in the
region share an approximate shape and a trajectory. Coherent subsystems and cycles
of change can guide the research strategies on parties in the region, suggesting more
emphasis on party organization and leadership, voter flows between elections, and
nonstandard issue dimensions. For those interested in changing the outcomes in the
region, it can also provide guidance as to “what works.” From a party strategy perspective,
that means organizational development, sustainable programmatic positions,
mechanisms for ousting changing leaders, and staying out of government until
absolutely necessary. From a policy perspective, the new party subsystem might be
reduced with more effective regulation on the source and transparency of party funding
and Estonian-style insistence on large numbers of founding members, but the
efficacy of these strategies is highly questionable.
Haughton and Deegan-Krause / Hurricane Season  77
Another benefit of the new-party-subsystem model and the related measurements
described above is that these approaches can integrate the study of parties in CEE
into broader studies of the political changes worldwide. Central and Eastern Europe
is not the only region where party systems display evidence of similar subsystems.
Emerging democracies in Latin America,31 Asia,32 and Africa33 have witnessed similar
patterns as have more established democracies. The patterns in Greece and Italy
bear striking similarity to the initial stages of the CEE pattern: significant new anticorruption
parties with celebrity leaders and rudimentary organization. Japan and
Israel (and, some would argue, the Netherlands, Belgium, and Austria) have already
seen the fading of an initial generation of newcomers and the emergence of “newer”
parties.
The similarity of patterns across a wide range of democracies calls attention to
underlying forces that may help to explain such a widespread (if still uneven) development.
The seemingly inexorable underlying trend toward dealignment of voters
from parties and the increasing role of celebrity combines in these cases with the
growing ease of organizing electronically for one-off events. Major success by one
new party creates the impression that others might succeed as well, and the resulting
subculture of political innovation may generate a steady stream of “startup parties”
that bear closer resemblance to their technology-industry counterparts than to traditional
political parties. Together, these developments simultaneously make it easier
to create a new party and harder to keep a party alive. Nor are these shifts unique to
political parties. Naím’s The End of Power sees similarly accelerating cycles of disruption
and replacement across the world in institutional realms ranging from the
software firms to civic associations to religions.34
The key to the variability of party system fortunes lies in the interaction between
these broader contemporary forces of change and the strength of local institutions.
The full effects of the forces producing party change may not be felt until events that
significantly weaken an established party or significantly strengthen a new one.
Where these shifts will happen—that is, which country and which political stream
within a country—is not easily predictable since the process is not a linear one but
proceeds in fits and starts because of thresholds in electoral law and in electoral psychology
that differ significantly over time and across borders and the role of accidents
and misjudgments. The global forces that push toward new-party subsystems
have little impact where established parties are strong enough to restrict new party
entry and avoid or recover from accidental damage. In such cases, an opening simply
may not appear, or it may not be big enough to create a self-sustaining subsystem.
But in systems with a major party collapse or the sudden intrusion of a new party—
such as in the Netherlands, Greece, Israel, Japan, or Italy—the situation has not
returned “to normal” and has experienced higher levels of subsequent volatility, particularly
extra-system volatility, and key elements of the new-party subsystem model.
Systems of older parties may have an advantage in holding off the forces of change,
but they seem to have no particular advantage in sustaining new parties.
78  East European Politics and Societies and Cultures
So is the world doomed to follow the Central and Eastern European party system
into a period of high volatility and ever newer parties? Given the constant possibility
of established-party failure and the fragility of new parties, we are likely to see a
certain degree of churn in almost every party system over the coming decades.
Whether this spells “doom” is a more complicated question. A certain degree of volatility
can actually be a welcome relief in systems like those of Western Europe where
the inertia has prevented some much needed change and there is no clear evidence
that larger new-party subsystems fare worse in the short run: the disruption in
Bulgaria caused by Simeon II does not seem to have made it worse off than the more
stable Romanian system, and volatility in the 2000s did not set Slovakia behind the
Czech Republic. More dangerous are the subtle losses inherent in new parties’shorter
time horizons and the difficulty of imposing accountability on a party that knows it
might well disappear soon anyway. These problems are by no means insoluble, however,
and citizens aware of the problem can try to find other mechanisms for ongoing
accountability and the search for long-term policy solutions. One advantage of hurricanes
over earthquakes is that you can see them coming.
Acknowledgments
Tim Haughton’s contribution to this article was partly written during a fellowship at the
Johns Hopkins University’s School of Advanced International Studies in 2011–2012.
Financial support accorded by the Austrian Marshall Plan Foundation for the fellowship is
gratefully acknowledged as are funds granted from the POLSIS departmental budget for a
further wave of research. The authors extend their thanks to many who assisted us in improving
our understanding including Daunis Auers, Nicole Bolleyer, Karin Bottom, Michael
Bernhard, Fernando Casal Bértoa, Blagovesta Cholova, Kris Deschouwer, Zsolt Enyedi, Jana
Grittersová, Anna Grzymała-Busse, Vlastmil Havlík, Seán Hanley, Isabelle Hertner,
Raimondas Ibenskas, Krzysztof Jasiewicz, Jeffrey Kopstein, Alenka Krašovec, Kathleen
Montgomery, Conor O’Dwyer, Mitchell Orenstein, Jan Rovný, Allan Sikk, Maria Spirova,
Milada Anna Vachudova, Emilia Zankina, and this journal’s anonymous reviewers.
Notes
 1. M. Pedersen, “The Dynamics of European Party Systems,” European Journal of Political
Research 7 (1979): 1–26; S. Bartolini and P. Mair, Identity, Competition and Electoral Availability
(Colchester: ECPR Press, 2007); J. E. Lane and S. Ersson, “Parties and Voters,” Scandinavian Political
Studies 20 (1999): 179–96; R. Dassoneville and M. Hooghe, “Mapping Electoral Volatility in Europe,”
(paper presented at the European Conference on Comparative Electoral Research, Sofia, Bulgaria,
December 1–3, 2012), 1–3; E. N. Powell and J. A. Tucker, “Revisiting Electoral Volatility in Postcommunist
Countries,” British Journal of Political Science 44 (2014): 123–47.
  2. A. España-Nájera, S. Mainwaring, and C. Gervasoni, “The Vote Share of New and Young Parties”
(Working Paper, Kellogg Institute, University of Notre Dame, July 2010, Notre Dame, IN); Powell and
Tucker, “Revisiting Electoral Volatility.”
Haughton and Deegan-Krause / Hurricane Season  79
  3. M. Kreutzer and V. Pettai, “Persistence and Decline versus Transience and Genesis” (paper presented
at the Joint Sessions of the European Consortium for Political Research, St. Gallen, Switzerland,
2011).
  4. A. Sikk, “Newness as a Winning Formula for New Political Parties,” Party Politics 18 (2012);
España-Nájera et al., “The Vote Share of New and Young Parties”; Powell and Tucker, “Revisiting
Electoral Volatility”; N. Bolleyer, New Parties in Old Party Systems: Persistence and Decline in
Seventeen Democracies (Oxford: Oxford University Press, 2013).
  5. Volatility of 0.0 will produce an increase in average age equal to the full period between elections;
extra-system volatility of 1.0 will produce a drop in average age to 0.0.
  6. J.-M. Esteban and D. Ray, “On the Measurement of Polarization,” Econometrica 62 (1994).
  7. Kreuzer and Pettai, “Persistence and Decline,” 78.
  8. F. Casal Bértoa, “Parties, Regime and Cleavages,” East European Politics 28 (2012).
  9. Giovanni Sartori, Parties and Party Systems: A Framework for Analysis (Colchester: ECPR Press,
2005), 39.
10. Ibid.
11. D. Hanley, “Managing the plural left,” in The French Party System, ed. J. Evans (Manchester:
Manchester University Press, 2003), 87.
12. M. Strmiska, “A Study on Conceptualization of (Ethno-) Regional Parties,” Středoevropské politické
studie 4 (2002).
13. J. Goody, Succession to High Office (Cambridge: Cambridge University Press, 1979); A.
Przeworski, “The Mechanics of Regime Instability in Latin America,” Journal of Politics in Latin
America 1 (2009).
14. T. Haughton, T. Novotná, and K. Deegan-Krause, “The 2010 Czech and Slovak Parliamentary
Elections,” West European Politics 34 (2011). Some degree of “corruption” coding is available in the
Comparative Manifestos Project, but in practice the results prove quite inconsistent.
15. D. Bochsler and I. Szekely, “We Do the Same, but We Do It Better. Post-ideological Electoral
Competition in Post-communist Democracies” (paper presented at the ECPR Joint Sessions, Münster,
2010), 21.
16. K. Gregor, “Jak se přesouvají voliči mezi volbami (ekologická Inference),” http://kohovolit.eu/cs/
jak-se-presouvaji-volici-mezi-volbami-ekologicka-inference/ (accessed 10 July 2014).
17. CVVM, “Naše společnost 2013 – říjen” [computer datafile/počítačový soubor], Český
sociálněvědní datový archiv - Sociologický ústav, Akademie věd ČR, v.v.i., ČSDA - reg. č. V1310 n.b.,
http://nesstar.soc.cas.cz/webview/index.jsp?v=2&node=0&study=http%3A%2F%2F147.231.52.118%3
A80%2Fobj%2FfStudy%2FV1310&mode=table&top=yes (accessed 10 July 2014).
18. A. Krašovec (University of Ljubljana, Slovenia), personal communication, 13 July 2014. On those
elections, see A. Krašovec and T. Haughton, “Predictably Unpredictable: The 2014 Parliamentary
Elections in Slovenia,” http://epern.wordpress.com/2014/07/17/predictably-unpredictable-the-2014-parliamentary-elections-in-slovenia/
(accessed 28 July 2014).
19. Bochsler and Szekely, “We Do the Same,” 19.
20. A. Ramonaitė, ed., Idėjos, Interesai Ir Įvaizdžiai Politikoje: Kaip Renkasi Lietuvos Rinkėjas?
(University of Vilnius, forthcoming).
21. Haughton et al., “The 2010 Czech and Slovak Parliamentary Elections”; S. Hanley, “Dynamics of
New Party Formation in the Czech Republic 1996-2010: Looking for the Origins of a ‘political
Earthquake,’” East European Politics 35 (2012).
22. “Procházkova Sieť Rastie, Kradne Voličov Najmä Matovičovi,” eTrend.sk, n.b. http://ekonomika.
etrend.sk/ekonomika-slovensko/prochazka-kradne-volicov-najma-matovicovi-a-smeru.html (accessed 20
July 2014).
23. The flowchart for the simulation can be found at http://www.pozorblog.com/2014/11/party-
replacement-simulation/.
24. Bartolini and Mair, Identity, Competition.
80  East European Politics and Societies and Cultures
25. K. Deegan-Krause and T. Haughton, “In with the New (Again): ‘Annuals,’ ‘Perennials’ and the
Patterns of Party Politics in Central and Eastern Europe” (paper presented at Midwest Political Science
Association Conference, Chicago, 12-15 April 2012). On organization, see M. Tavits, Post-Communist
Democracies and Party Organization (Cambridge University Press, 2013), and M. Spirova, Political
Parties in Post-communist Societies (London: Palgrave Macmillan, 2007).
26. A. Roberts, The Quality of Democracy in Eastern Europe (Cambridge: Cambridge University
Press, 2010).
27. S. Hug, Altering Party Systems (Minnesota: University of Michigan Press, 2001); A. Hino, New
Challenger Parties in Western Europe (London: Routledge, 2012); and Bolleyer, New Parties in Old
Party Systems.
28. M. Poliačik (Freedom and Solidarity Party, Slovakia), Interview. Bratislava, 9 August 2010.
29. R. Dassonneville, “The Impact of Short-Term Voting Motives for the Younger Generation. An
Analysis of Party Preference Determinants in Flanders (Belgium)” (paper presented at the Dutch-Belgian
Political Science Conference, Amsterdam, 2011).
30. T. Haughton, “Exit, Choice and Legacy: Explaining the Patterns of Party Politics in PostCommunist
Slovakia,” East European Politics 30 (2014).
31. N. Lupu, “Brand Dilution and the Breakdown of Political Parties in Latin America,” World
Politics, forthcoming, n.b. http://www.noamlupu.com/breakdown.pdf (accessed 10 July 2014).
32. A. Hicken and E. M. Kuhonta, eds., Party System Institutionalization in Asia (Cambridge:
Cambridge University Press, forthcoming).
33. S. I. Lindberg, “Institutionalization of Party Systems? Stability and Fluidity among Legislative
Parties in Africa’s Democracies,” Government and Opposition 42 (2007).
34. M. Naím, The End of Power (New York: Basic Books, 2013).
Author Biographies
Kevin Deegan-Krause is an Associate Professor of Political Science at Wayne State University. His
research interests span numerous fields of comparative politics, including European politics, political
parties, democratic institutions and national identity. He is the author of Elected Affinities: Democracy
and Party Competition in Slovakia and the Czech Republic (Stanford University Press, 2006), coeditor of
The Structure of Political Competition in Western Europe (Routledge, 2010), and coeditor of the
European Journal of Political Research’s Political Data Yearbook with responsibility for its interactive
data site.
Tim Haughton is Reader (Associate Professor) in European Politics at the University of Birmingham.
His current research projects explore various aspects of electoral and party politics, especially in Central
and Eastern Europe, the relationship between politics and government, the interaction between domestic
and European sources of change, and the dynamics of domestic politics in Slovakia, Slovenia, and the
Czech Republic. He edited Party Politics in Central and Eastern Europe: Does EU Membership Matter?
(Routledge, 2011) and coedits the Journal of Common Market Studies’ Annual Review of the European
Union.