Institutional political dynamics:
political cycles between the liberal and the conservative
Dong-Hwan Kim, Khalid Saeed
Abstract
Although many system dynamicists point out the close relationship between institutional
economics and system dynamic, the relationship between institutional political theory and
system dynamics approach is not explicated appreciated yet. We developed a system dynamics
model to investigate theoretical propositions of institutional politics. Our system dynamics
model showed how the endogenous mechanism can explain the political changes as well as
orders. Although simple in the causal structure, our model could show a complex behavior of
political competition. Several simulation results imply that some unexpected change in election
and power competition may come from the endogenous system rather than from exogenous
factors.
1. Order and change in politics
There are lots of dynamics in political life. We observe continuing ebb and flow in the political
support of parties, in the popularity of politicians, and in public interests of political issues
(Richards 1992). It seems that political change is a rule rather than an exception.
Although political change seems to be a way of political life, it is hard to predict what kind of
change will prevail next time and who will benefit from that change. Political changes are hard
to predict. On the other hand, it is hard to think that political changes occur randomly. Although
it is difficult to predict political changes, it is rather easy to find causal factors for them
afterwards.
However, it is hard to find any political theory that explains the structural mechanism of
political change. Political changes are usually explained in terms of historical events. Usually
previous events are interpreted as a source of posterior events (Thelen 1999). However,
chronological explanation is hard to give a necessary and sufficient structure of political
changes. Historical explanation only tries to give endless story of political changes.
Traditional political theory is more likely to stress political order rather than political change.
Game theory emphasizes the equlibrium state rather than disequilibrium or changing states.
Also institutional political theories tend to focus on order rather than change. Specifically,
political theorists showed that the percentage of partisan voters will grow and the party system
will become stable, unless there is a external crisis. This tendency of growing partisan voting
behavior predict that there will be no power change in subsequent voting. (Leithner 1997)
However, recent theoretical deveoplment in the institutional political theory points out the gap
between the traditional emphasis on the order and the political reality full of change. As pointed
out by Greif and Laitin (2004), it is a central concer in political science today - that is, how to
explain both institutional stability and change with endogenous causal mechanisms in politics.
In this paper, we try to make a model that can show the endogenous mechanism of political
change.
To make a dynamic model, we use a system dynamics approach. System dynamics model is
powerful in constructing endogenous mechanism of changing system. And thus, system
dynamicists have made several models for political change. Among them the studies on
resource allocation in authoritarian regime and the resulting unstable cycles of political regime
are most famous (Saeed 1998, Pavlov, Radzicki and Saeed 2005, Saeed and Pavlov 2006). They
focus on the resource allocation mechanism as a endogenous engine for political change in
authoritarian regime. In this paper, we want to make a dynamic model for a democratic political
system focusing on its conflict resolution process.
2. Endogenous explanations of political cycle
Theories on political change can be characterized as endogenous vs. exogenous. Exogenous
theory of political change finds the driving force of political change outside political systems,
especially economic system. Endogenous approach tries to find a systematic mechanism of
change inside the political system itself.
Political changes that repeat themselves create a pattern of political cycles. Many scholars in
politics have recognized political ups and downs repeating in the political history of the world
as well as America. Periods of political cycles range from several months to hundred years.
One of the longest period of political cycle reported is by Zvi Namenwirth. He analyzed patterns
of changing values in party platforms. And he found there are long waves of 152 years length
and short waves of 48 years length. He explained these patterns of political cycles as reflecting
problems in the society, especially economic problems. In particular, he explained the short
period of political cycle as reflecting the economic long wave of 50 to 60 years, that is,
Kondratieff wave. In the period of increasing wealth, the liberal and progressive value will be
emphasized in the party platforms. However, once the expansion tums its peak and economic
contraction sets in, the conservative value begins to dominate the platform of political parties
(Namenwirth 1973).
A political cycle resulted by the economic long wave is often regarded to determine war cycles.
Goldstein points that the economic long wave is closely connected with the cyclical character of
great world war. He argues that long-term economic upswings could increase the likelihood of
war through several mechanisms; high competition for markets and resources, increase in
military expenditures, and growing aggressiveness and expansionist mood conducive to war
(Goldstein 1985).
Endogenous approach focuses on political mechanisms as a source of creating pattems of
political changes, that is, political cycle. One of the most famous arguments is that of
Huntington (1981). He observes that A merican politics experienced cyclic changes with interval
of sixty years. At sixty-year interval in American history, significant numbers of Americans
become angry at ‘the ideals-versus-institutions gap’ and engage in collective behavior to restore
the seemingly corrupted institutions to the ethical level of the creed. He calls these time of
political changes as reform periods and he observed them in the 1770s, the 1830s, the 1900s and
the 1960s. Huntington concludes that the main force of political changes is a ‘creedal passion’
resulting from the gap between liberal ideals and conservative institutions. His perspective on
political change focus on the reform periods. Smith observed that his data is consistent with the
cycle of reform hypothesis (Smith 1990).
Although reform hypothesis gives an endogenous perspective on political cycles, it is not a
sufficient mechanism for cycles to take place. While reform hypothesis can explain political
change driven by the creedal passion, it alone cannot explain why the passion for reform
disappears and substituted by political order.
The interest group cycle theory raised by McFarland gives why reform and order are alternating
each other. Political interest group can be decomposed generally into the liberal (reform) and the
conservative (order). Political cycles are generated by interactions between the liberal (reform)
and the conservative. The focal point in the interest group cycle theory is the excess resulted by
power group, whether or not it is the liberal or the conservative. A fter the widespread perception
of excess (exploitation by the power group), the public, the media and politicians become
involved to correct the problems (p.263). The concept of the excess explains why the reform
periods stops and the period of order succeeds and vice versa.
“During reform periods, egalitarian social movements eventually produce actions
that are viewed as excesses by a great majority of the politically active public. Calls
for violence, flag burning and anti-Americanism, radical socialism or even
trumpeting the need for zero economic growth act as a ‘turn off’ to general public
support and participation. (McFarland 1991 p.268)”
McFarland observed rather short periods of political cycle. He points out that the liberal value
and the conservative value alternates at every 30 years.
“Interest-group cycles theory, then, means that during the 1890s, 1920s, 1950s and
1980s the power of economic producers was gaining within hundreds of issue areas
at the expense of countervailing power groups. As stated here, cycle theory implies
that during the reform periods - the 1900s, the 1930s, the 1960s and possible the
1990 - business and producer power weakens in hundreds of issue areas.
(McFarland 1991 p.265)”
While these cycle theory offers endogenous mechanism for political cycles, recently
institutionalist political theorists tried to find a structural mechasnism of political change. Firstly,
Paul Pierson points out that institution itself is a reinforcing mechanism, because institutions is
likely to continue and to reinforce certain kinds of political thoughts and values. He said if there
is a force that weaken the institution, there comes the political change (Pierson 2000).
Again his theory is not sufficient to produce self-generating cyclic change. There needs a
negative feedback loop that limit the force of positive feedback loop of reinforcing institutions
(Greif & Laitin 2004). Lieberman argued that to explain political change one should relax
emphasis on order in institutional perspective. He points out that political conflict as well as
order is institutionalized in the political system, and they conflict, friction among multiple
political order can give a starting point for institutional political change. His theory resembles to
that of McFarland and Schlesinger.
Our study on building system dynamics model starts on these verbal theory of political change
and cycle.
3. A system dynamics model for political change
Although Pierson argued institution as an order enforcing mechanism, all institutions are not
made for maintaining previous order. Some institutions are designed to introduce change.
Political party system is one of the institutions that institutionalize conflict, friction, and
endogenous political changes.
We assume that there are two competing parties, conservative and liberal. Citizen can change
their support toward the conservative and the liberal party. Also we included in our model the
deprived and the radical, whose number is affected by the pro-capitalist of the conservative
party.
Actors Role / Goal Affected By
Conservative | - Support pro-capitalist policy, Supporting from citizen
Party - Decrease economic regulation
Liberal - Increase redistribution polity that | Supporting from citizen
Party will decrease the number of the
deprived,
- Increase economic regulation
Citizen - Change their support between the | If there are too many deprived, more
conservative party and the liberal | citizen will support the liberal. If there
party are too many regulation, they will be
afraid of economic decline and change
their support to the conservative.
The deprived | - Enhance sympathy in the citizen | The pro-capitalist policy will increase
to increase support for the liberal | the number of the deprived, while the
party redistribution policy will decrease it.
The radical - Increase the oppression by the
conservative
- Decrease the economic efficiency
<table 1>Actors and their roles in the model
As usually assumed by political scientists (Smith 1990), we assume that the liberal party can be
characterized by reformist, regulatory and interventionist, and egalitarian. In contrast, the
conservative is assumed to be less concerned about equality but to try to establish the adequate
incentive for capital investment (Boix 1998).
In this paper, we assume the excess pointed by McFarland as a driving force to change political
support by citizen. The excess by the conservative party is represented by the number the
deprived that is increased by the pro-capitalist policy of the conservative. As the number of the
deprived increase, people will have more sympathy for them, perceive the necessity of the
redistribution policy and support the liberal party. As the liberal party gets more power, their
redistribution policy will be reinforced to reduce the number of the deprived. But there is
another excess coming from the liberal party, an increase in the economic regulation and the
violence by the radical group, which decrease economic efficiency. As the economic efficiency
goes down, public interest moves from human rights to economic stability and development.
Thus, the conservative party gets more supporters and powers.
These excess mechanism forms two negative feedback loop, each for the conservative and the
liberal. With these negative feedback loops, they recover their power when their power is low.
But with the same feedback loops, their strong power may be constrained to reduce the excess
created by their own power.
There are another feedback loops that accelerate the changing process. As many political
thinkers pointed out, violence is oftentimes an important source of political dynamics (Saeed
1998, Morris 1993). Violence in political regime is a reciprocal process between the oppressor
and the oppressed. In this sense, violence is reciprocal and forms a positive feedback loop
(Kritzer 1977). As violence grows, oppression against them increase, which will increase the
violence further.
In our model, we introduce two positive feedback loops of violence. The first is the reinforcing
violence loop that works to increase the number of supporters of the liberal party. This loop is
emotional. As the number of the deprived group grows, more people become radical, and more
violence will take place. This makes the conservative unstable, and fear of violence increase. So,
they begin to oppress the deprived, and the number of the deprived will increase. And this
makes the positive feedback loop, a vicious circle for the conservative.
The other feedback loop around violence is its effect on the economic efficiency. As there are
more violence in the society, economic efficiency will decrease. And thus more people think
that economic efficiency is more important than the human rights. As more and more people
support the conservative party, their power and their policy for pro-capitalist will be stronger.
Finally, the number of deprived and the resulting violence will be increased. This positive
feedback loop works to increase the power of the conservative rather than the liberal party.
Over all, the party competition system is composed of several negative and positive feedback
loops. Firgure 1 shows the feedback structure of our model.
{number of the
Lo (+)
oppression by the "umber of the supporters of the
wast conservative radical liberal party
pro- capi
ci
Fon (4) Nf + redistribution
Np policy
supporters of the f
conservative party public interest in
+ economic efficiency
(-) a
economic
regulation +
Figure 1. Feedback loops in the party competition model
The causal diagram is converted to simulation model composed of five stocks that show the
states of political system and flows that change the system states. Citizen consists of supporters
for the conservatives and the liberals. And other stocks are related with the excess created by the
power group. The deprived and radical group describe the system states changed by the excess
resulted by the power of the conservative, while the regulation is resulted by the power of the
liberals.
iv-dep
Se Ss
exploitation ratio
PH The deprived
Increasing rate of|
the deprived
Decreasing rate of
increase in radicals
violence
the deprived
change in
exploitation = Sar
very Sympathy forthe maim
Raclical al wa depived —“* thedeprived
change in violence
oppression —_ St ofthe deprived
A =, ‘in ra ‘on sympathy
miniomm sspobeeeee ldelay in change for
conservative’ hatha
eee ‘eaistribution policy
pro-capitalist policy v
political powerof tical ow
‘Supporters for the| the conservative Pe oh ihe [supporters forthe
conservative Meee
Party population.
me =
delay in change for ‘minima liberal change in equality
the conservative supporters
level of
from Mberal to: , ea ‘effect of violence maximum!
Sisal te regulation ies
perception of need. effect of regulation ba
foreficiency Pee in on the economy
economy
so
ont Regulations 2
Decreasing rate of Increasing rate of
regulation regulation
Figure 2. Flow-Stock diagram of simulation model for party competition
System dynamics model in figure 2 is a theoretical model rather than a model for describing real
world. This is a highly simplified and abstracted model.This model can show how the complex
behavior of political change can take place over and over again without exogenous factors.
4. Model behavior
In system dynamics model, only flows can change stocks. When all value of flows are zero, the
system cease to be changed. That is the equilibrium in the system dynamics model. The
equilibrium point of our model can be found when the value of the stock and parameters are set
as follows.
Supporters of the conservative party = 1500
Supporters of the liberal party = 1500
The deprived =5
Radical Group =5
Regulations =5
exploitation =5
equality =5
violence = 1/3
The existence of equilibrium point does not guarantee a stability of the system. If the
equilibrium is not stable, system can be deviated from the equilibrium even with the small
change in the parameters or stocks.
1) Instability with changes in initial values
The power competition of two party system is sensitive to the initial value of stock variables. If
we increase the initial value of the supporters of the conservative party by 10%, there comes the
fluctuating behavior in the supporters of the conservative and the liberals. It will take more than
30 years to return to the equilibrium state as shown in figure 3.
This means that if there are some external changes such as technological or economical
development and thus some change in distribution of political supporters, the political system
will evolve different way. We found that the political competition system is unstable and very
sensitive to the all kinds of stocks. This result confirms that political systems are full of
endowment effect and path dependency, as asserted by historical institutionalism in politics.
10
0 180 360 540 720 900 1080 1260 1440 1620 1800
Time (Month)
political power of the conservative :-de-
political power of the liberal : dd ---------------------+-++-2------------
Figure 3. Political competition with 10% increase in the number of supporters for the
conservative
2) Limit cycles
When one changes parameters in the model, the equilibrium can disappear completely. For
example, when we change the exploitation by the conservative from 5 to 4, there comes the
continuing cyclic behavior in the number of supporters as in figure 4. This limit cycle in the two
party system implies that there will be endless alternation of power between the conservative
and the liberal.
We must emphasize that this cyclic behavior comes from internal structure. The cyclic behavior
of political supporters imply that timing of election will determine who will take power position.
If the election day is scheduled at the time of high level of the conservative, the liberal party
will loose the election.
When we reduce the exploitation further to 3, we can find that the equilibrium will come back
after a long period of cycle in the power as in figure 5. But the resulting equilibrium is different
from the previous one. In the new equilibrium, the conservative party will hold the power
position continually. That is because the low exploitation will reduce the resistance from the
deprived and the radical.
11
0 180 360 540 720 900 1080 1260 1440 1620 1800
Time (Month)
political power of the conservative : dd
political power of the liberal : dd-----------=----------+--20-0-0-0-seenoee
Figure 4. exploitation =5 + step( -1, 60)
0 180 360 540 720 900 1080 1260 1440 1620 1800
Time (Month)
political power of the conservative :-¢ 44H
political power of the liberal : dd -----------+-=-+-=-+-+-------000-es0s
Figure 5. exploitation =5 + step( -2, 60)
3) Chaotic Behavior
The political system can show more complex behavior. Politics in reality is not in equilibrium
state nor they show the regular pattern of cycles. If we cannot show this complex or chaotic
behavior, the model can be judged too simple and lack important mechanisms.
We found that our model also shows complex and chaotic behavior. When we introduced a
slight change in the amount of exploitation, the system showed cycles of amplification and
contraction as in figure 6. This implies that a slight change from the equilibrium state can
produce some kind of chaotic fluctuation.
12
The system shows several long period of stability. Also, the period of order in figure 6 is similar
to the punctuated equilibrium. But big fluctuation comes after the silent period of stability. In
figure 6 one can find that a period of political stability continues for a long time around 40 years
(540 month). Gradually, the politics becomes unstable. One party gets power at the expense of
other party’s defeat. This period of fluctuating power also continues around 40 years.
0 180 360 540 720 900 1080 1260 1440 1620 1800
Time (Month)
political power of the conservative : dé-
political power of the liberal : dd-------------------------22-20-2-000200->
Figure 6. Alternation of stable period and unstable period.
exploitation =5 + step(-0.1, 60)
Alternating system behavior between stability and instability in figure 6 is similar to the ‘reform
period’ suggested by Huntington. He pointed out that reform period come every 60 years. Figure
6 shows that the period of political instability comes every 80 years (900 month). Our
simulation results offers analytical basis for the reform cycle observed by Huntington.
The 40 years period of stable politics shown in figure 6 can be interpreted as a political order
resulted by institutional force. The succeeding abrupt destruction of peaceful stability might be
interpreted as being introduced by external forces. That is because one cannot find any clue of
the instability for more than 40 years. People tend to think there will be something outside the
peaceful political regime that triggered the instability. But the unexpected unstable period of
power change comes from internal structure not from external factors.
Fluctuating behavior in figure 6 can be decomposed into two cycles. One cycle is a short cycle,
period of which is approximately 3 years (36 month). This cycle can be interpreted as being
13
resulted by negative feedback loops around the excess created by both parties.
The other cycle is a long cycle, period of which is about 80 years. We found that this cycle is
affected by one of the positive feedback loop of violence. That is the feedback loop of “Radical
group => violence => economic efficiency => supporters for conservative => power of the
conservative => pro-capitalist policy => radical group”. This feedback loop is activated when
the number of Radical group is increased over time as shown in figure 7.
ed
0 180 360 540 720 900 1080 1260 1440 1620 1800
Time (Month)
political power of the conservative : dd ——-——_————_—
political power of the liberal : de
Radical. group 9 dees ersenceincee=eenieisve=aetaremaers
Figure 7. Radical group as a source of long term cycle
exploitation =5 + step(-0.1, 60)
From this simulation result, we can conjecture that the long cycle of reform comes from the
effect of violence on the economic efficiency. As the effect of violence by Radical group on the
economic efficiency increases, the gain of the positive feedback loop increases, and this positive
loop amplify the small fluctuation. In the peaceful days, the number of radical group is small
and thus their violence. But in the days of reform, lots of radical group is accumulated and their
violence will become critical in power shifting between the liberal and the conservative.
5. Discussion
Our system dynamics model shows why there are political changes as well as order from the
endogenous factors. This model also can be used to look into the strategic implication of public
policies.
14
1) Asymmetry of tolerance
Sometimes tolerance is said as the virtue of political parties toward opponents. In our simulation
model, tolerance for the conservative means less exploitative policy, while it means less
egalitarian policy for the liberal party.
With several experiments of introducing tolerance for both parties, we found that there are
asymmetry of the tolerance effect between the conservative party and the liberal party. Figure 8
shows the simulation result of reducing the exploitation of the conservative from 5 to 3. Less
exploitative policy benefits the conservative in the long run. In the new equilibrium, the
conservative party will dominate the liberal party.
0 180 360 540 720 900 1080 1260 1440 1620 1800
Time (Month)
political power of the conservative : dé
political power of the liberal : dd-------------------+------rererreeeeneeesnenees
Figure 8. Reducing the exploitation of the conservative from 5 to 3 at 60 month.
On the contrary to the beneficial result for the conservative, the effect of tolerance is ambiguous
for the liberal party. When we reduce the equality policy from 5 to 3 at 60 month, the liberal
party could get great support at the first time period of around 15 years (180 months). But after
that initial period, the conservative and the liberal compete for the power.
0 180 360 540 720 900 1080 1260 1440 1620 1800
Time (Month)
political power of the conservative : dd
political power of the liberal : dd -------------2------2+-+---2--0-eeeseeeeseesenes
Figure 9. Reducing the equality policy of the liberal from 5 to 3 at 60 month.
Tolerance benefits the liberal party only in short term, while it benefits the conservative in the
long run. In this sense, tolerance can be regarded as a virtue of the conservative that can be used
in the long time. For the liberal, tolerance is a strategy that should be used in the short term.
2) Violence tendency of the society
If the tolerance is the virtue of the power taker, the non-violence is a virtue required for the
people, especially the deprived and the radical group. From the civil right movement in 1960s
non-violence is regarded as a crucial leverage for the success of equality policy.
But experiments with our system dynamics model shows that decerase in the violence tendency
will benefit the conservative rather than the liberal. If the violence is decreased, the oppression
of the conservative will be reduced also. And thus the number of the deprived will be decreased,
which will reduce the sympathy for the weak people and thus the number of supporters for the
liberal party.
16
0 180 360 540 720 900 1080 1260 1440 1620 1800
Time (Month)
political power of the conservative :-d
political power of the liberal : dd ------------------=----------------—-—-
Figure 10. Effect of reducing violence tendency from 0.33 to 0.13 at 60 moth.
Our simulation shows two kinds of system behavior of reducing violence. The first is the long
term benefit for the conservative. If we consider this effect only, we can say that non-violence is
not beneficial to the liberal party. But the second effect of reducing violence is the cyclic power
shift between the conservative and the liberal. This means that decrease in the violence may
create new opportunity for the liberal to pursue their reform. By reducing the violence tendency
of the people, one can create a chance for re-building power structure between the conservative
and the liberal party.
6. Conclusion
Our system dynamics model showed how the endogenous mechanism can explain the political
changes as well as orders. Although simple in the causal structure, our model could show a
complex behavior of political competition. Several simulation results imply that some
unexpected change in election and power competition may come from the endogenous system
rather than from exogenous factors.
Our model incorporated some essential ideas raised by the institutional political theories.
Although many system dynamicists point out the close relationship between institutional
economics and system dynamics (Atkinson 2004, Radzicki 1988), the relationship between
institutional political theory and system dynamics approach is not explicated appreciated yet.
We hope that our system dynamics model can give a linking pin between institutional political
theory and system dynamics.
References
Atkinson, G., 2004, “Common ground for institutional economics and system dynamics
modeling,” System Dynamics Review, vol. 20, no. 4, pp.275-286.
Boix C., 1998, Political Parties, Growth and Equality, Cambridge University Press.
Crouch H, J.W. Morley, 1993, Driven by Growth, An East Gate Book.
Gasiorowski, MJ., 1995, “Economic crisis and political regime change: an event history
analysis,” American Political Science Review, Vol. 89, no. 4, 882-897.
Goldstein, J.S., 1985, “Kondratieff waves as war cycles,” International Studies Quarterly, vol.
29, pp.411-444.
Greif A., D.D. Laitin, 2004, “A theory of endogenous institutional change,” American Political
Science Review, vol. 98, no.4, pp.633-652.
Huntington, S.P., 1981, American Politics: The Promise of Disharmony, Cambridge: Harvard
University Press.
Kritzer, H.M., 1977, “Political protest and political violence: A nonrecursive causal model,”
Social Forces, vol. 55:3, pp.630-640.
Lieberman R.C., 2002, “Ideas, institutions, and political order: explaining political change,”
American Political Science Review, vol. 96, no.4, pp.697-712.
McFarland A.S., 1991, “Interest groups and political time: cycles in America,” British Journal
of Political Science, Vol. 21, Nol. 3, pp.257-284.
Morris A.D., 1993, “Birmingham confrontation reconsidered: An analysis of the dynamics and
tactics of mobilization,” American Sociological Review, Vol. 58, pp.621-636.
Leithner, C., 1997, “Of time and partisan stability revisited: Australia and New Zealand 1905-
90,” American Journal of Political Science, Vol. 41, No.4, pp.1104-1127.
Namenwirth, J.Z., 1973, “Wheels of time and the interdependence of value change in America,”
Journal of Interdisciplinary History, 3:4, pp.649-683.
Pavlov, O.V., M. Radzicki, K. Saeed, 2005, “Stability in a superpower-dominated global
economic system,” Vol. 39 no. 2, pp.491-500.
Pierson, P., 2000, “Increasing returns, path dependence, and the study of politics,” American
Political Science Review, vol. 94, no.2, pp.251-267.
Radzicki, MJ., 1988, “Institutional Dynamics: An extension of the institutionlist approach to
socioeconomic analysis,” Journal of Economic Issues 22, no. 3, 633-666.
Richards D., 1992, “Spatial correlation test for chaotic dynamics in political science,” American
Joumal of Political Science, vol. 36, no. 4, pp.1047- 1069.
Saeed K., 0.V. Pavlov, 2006, “Dynastic cycle: A generic structure describing resource allocation
in political economies, markets and firms,” working paper.
18
Saeed, K. 1998. Towards Sustainable Development, 2nd Edition: Essays on System Analysis of
National Policy. Aldershot, England: Ashgate Publishing Company.
Schlesinger, A.M., 1986, The Cycles of American History, Houghton Mifflin Company.
Smith, T.W., 1990, “Liberal and conservative trends in the united states since world war II,” The
Public Opinion Quarterly, Vol. 54, No. 4, pp.479-507.
Thelen K., 1999, “Historical institutionalism in comparative politics,” Annual Review of
Political Science, Vol. 2. pp.369-404.
19
