Digitized by the Internet Archive
in
2011 with funding from
Boston Library Consortium IVIember Libraries
http://www.archive.org/details/persistenceofpowOOacem
DEWEY
HB31
.M415
c?3
Massachusetts Institute of Technology
Department of Economics
Working Paper Series
PERSISTENCE OF POWER, ELITES
AND
INSTITUTIONS
Daron Acemoglu
James A. Robinson
Working Paper 06-05
February 28, 2006
Room
E52-251
50 Memorial Drive
Cambridge,
MA 021 42
This paper can be downloaded without charge from the
Social Science Research Network Paper Collection at
http://ssrn.com/abstract=8881 87
Persistence of Power, Elites and Institutions^
James A. Robinson^
Daxon Acemoglu^
This Version: February 2006.
Abstract
We
construct a model of simultaneous change and persistence in institutions.
The model
and workers, and the key economic decision concerns the form
(e.g., competitive markets versus
labor repression). The main idea is that equilibrium economic institutions are a result of the
exercise of de jure and de facto political power. A change in political institutions, for example
a move from nondemocracy to democracy, alters the distribution of de jure political power,
but the elite can intensify their investments in de facto political power, such as lobbying or
consists of landowning elites
of economic institutions regulating the transaction of labor
the use of paramilitary forces, to partially or fully offset their loss of de jure power.
In the
baseline model, equilibrium changes in political institutions have no effect on the (stochastic)
equilibrium distribution of economic institutions, leading to a particular form of persistence in
we refer to as invariance. When the model is enriched to allow
on the exercise of de facto power by the elite in democracy or for costs of changing
economic institutions, the equilibrium takes the form of a Markov regime-switching process
equilibrium institutions, which
for limits
with state dependence. Finally, when we allow
tutions
is
more
for the possibility that
insti-
than altering economic institutions, the model leads to a pattern of
difficult
captured democracy, whereby a democratic regime
tions favoring the elite.
changing political
The main
may
ideas featuring in the
survive, but choose
model are
economic
institu-
illustrated using historical
examples from the U.S. South, Latin America and Liberia.
Keywords: democracy, de facto power, de jure power, dictatorship,
elites, institutions,
labor repression, persistence, political economy.
JEL
Classification: H2, NIO, N40, P16.
*We thank Alexandre Debs for excellent research assistance and Lee Alston, Timothy Besley, Alexandre
Debs, Stanley Engerman, Michael Munger, Nathan Nunn, Torsten Persson, Konstantin Sonin, Gavin Wright and
Pierre Yared and seminar participants at Clemson, ITAM, Princeton, Rochester, and the American Economic
Association Annual Meetings for comments.
Acemoglu
gratefully acknowledges financial support
from the
National Science Foundation.
^Massachussetts Institute of Technology, Department of Economics, E52-380, 50 Memorial Drive, Cambriudge MA02142. e-mail: daron@mit.edu.
'Harvard University, Department of Government, IQSS, 1737 Cambridge St., N309, Cambridge MA02138;
e-mail: jrobinson@gov.harvard.edu.
meme
"Plus ga change plus c'est la
"The domination
The power
who
of an organized minority
any minority
of
chose." French Proverb.
is
over the unorganized majority
...
irresistible as against
is
organized for the very reason that
it
is
inevitable.
each single individual in the majority,
stands alone before the totality of the organized minority.
minoritj'
is
At the same time, the
a minority." Gaetano Mosca (1939,
p.
53).
Introduction
1
Current empirical work and theoretical discussions of the impact of institutions on economic
development either implicitly or exphcitly assume that institutions persist
Engerman and
Sokoloff, 1997,
North, 1990,
(e.g.,
Acemoglu, Johnson and Robinson, 2001, 2002). In
fact,
some
of
the most popular empirical strategies in gauging the efTect of institutions on economic perfor-
mance use the
and
persistence of institutions over centuries as part of their conceptual approach
identification strategy.
periods
much
America and
But many aspects
shorter than a century.
Africa, have
Many
changed their
of "institutions"
show substantial change over
less-developed countries, especially those in Latin
political institutions all too often over the past 100
between democracy and dictatorship
years, with frequent switches
(see, e.g.,
Acemoglu and
Robinson, 2006a) and multiple changes in constitutions.^
The same
pattern also emerges
many historians and economists
practices such as the
plantation complex,
when we turn
economic institutions. For example, while
trace the economic problems of Latin
America to
colonial labor
encomienda or the mita, and those of the Caribbean to slavery and to the
all
of these economic institutions vanished long ago.^
of change, however, economic systems often
tural labor relations in
after colonialism,
to
many
and perhaps
of the Latin
show surprising
continuity.
Beneath
this pattern
The form
of agricul-
American and Caribbean countries changed
relatedly, these societies continued to suffer various
little
economic
problems, slow growth, and economic and political instability throughout the 20th century.
Another interesting example comes from the U.S. South. Even though slavery was abolished
at the
end of the Civil War, the U.S. South maintained a remarkably similar agricultural
sys-
tem, based on large plantations and low-wage uneducated labor, and remained relatively poor
'For instance, Colombia had 8 constitutions in the 19th century (Gibson, 1948), while Bolivia had 11 (Trigo,
1958) and Peru 9 (Palacios and Guillergua, 2003).
^In Latin America, the last form of official forced labor, pongueaje, was abolished in Bolivia in 1952 (Klein,
1992, Chapter 8).
freed, for
example
Unpaid labor services lasted
in
lasted until in 1886 in
in Guatemala until 1945 (McCreery, 1994). Slaves were gradually
Colombia. In the British Caribbean slavery was abolished after 1834, though it
Cuba and 1888 in Brazil.
1850
in
until the
middle of the 20th century.
In this paper,
we provide a
possible explanation for this paradoxical pattern of the coexis-
tence of frequent changes in political institutions with the persistence in certain (important)
Our approach
aspects of economic institutions.^
The
of persistence.
baseline
model leads
illustrates the possibility of
to a pattern
which we
two
different types
refer to as invariance,
whereby
a change in political institutions from nondemocracy to democracy leads to no change in the
(stochastic) equilibrium process of
in society.
we
economic institutions and of the distribution of resources
Simple extensions of our baseline model lead to a richer form of persistence, which
refer to as state dependence; the probability that a society will
tomorrow
pro-citizen economic institutions)
The underlying
idea of our approach
is
is
be democratic (and have
a function of whether
it is
democratic today.^
that equilibrium economic institutions emerge from
the interaction between political institutions, which allocate de jure political power, and the
distribution of de facto political
power across
social groups (see
Acemoglu and Robinson, 2006a,
and Acemoglu, Johnson and Robinson, 2005b). De Facto power
by institutions (such
weapons or
as elections), but rather
is
is
power that
is
not allocated
possessed by groups as a result of their wealth,
ability to solve the collective action problem.
A
change
in political institutions
that modifies the distribution of de jure power need not lead to a change in the equilibrium
process for economic institutions
of de facto political
paramilitaries)
.
power
The
if it is
(e.g., in
central
associated with an offsetting change in the distribution
the form of bribery, capture of the political parties, or use of
argument
in this
paper
is
that there
is
a natural reason to expect
changes in the distribution of de facto political power to partially or entirely
in
offset
changes
de jure power brought about by reforms in specific political institutions as long as these
reforms do not radically alter the political structure of society, the identity of the
source of economic rents for the
To make these
and the
citizens.
ideas precise,
elites, or
the
elites.
we develop a model
The key economic
consisting of two groups, a landed ehte
institution concerns the organization of the labor mar-
^Throughout, persistence refers to the continuity of a cluster of institutions, for example, the extent of
enforcement of property rights for a broad cross-section of society. Lack of property rights enforcement may have
its roots in quite different specific economic institutions, for example, risk of expropriation by the government
or elites; extreme corruption; economic systems such as serfdom or slavery preventing large segments of the
population from selling their labor freely or from investing in most economic activities; legal rules making it
impossible for those without political connections to have their contracts enforced; or entry barriers creating a
non-level playing
field.
''We refer to this type of persistence as "state dependence" since the probability distribution over equilibrium
political
and economic institutions tomorrow depends on the "state" of the system, which
is
political institutions
today. See Page (2006) for a discussion of richer forms of persistence in political systems, where the past entire
sequence of events, rather than simply a low-dimensional state vector, might influence future outcomes.
whether wages are competitive or are repressed below
ket, in particular,
model, economic institutions are decided either by the landed
depending on who has more
In the
the citizens (workers)
elite or
Political power, in turn,
political power.
this level. ^
determined by both
is
pohtical institutions that allocate de jure power and the distribution of de facto power, which
is
derived, at least partly, from a social group's ability to solve their collective action problem.
A
key observation
that landowners, by virtue of their smaller numbers and their established
is
position, have a comparative advantage in solving the collective action
amount
Olson, 1965). This implies that the
of de facto political
librium outcome, and responds to incentives.
political
power
power
litical
is
matter
also
for
effectively,
may
democracy de jure potowards the citizens
tilted
and the
help the citizens in solving their collective action problem
thus facilitating their exercise of de facto political power.
Those with greater
field is in this contest.
is
a "contest" between the
elite
(democracy versus nondemocracy) determine how
political institutions
today and
is
in
an equi-
and de jure
addition, freedom of political organization
In the model society, in every period there
and
elite is
Nevertheless, political institutions
allocated to the majority, so the balance of power
existence of political parties
more
power of the
equihbrium outcomes. For example,
Acemoglu and Robinson, 2006a). In
(see
problem (Mosca, 1939,
The most
tomorrow.
political institutions
that, because the elite's de facto political
political
power
is
and the
level
citizens,
the playing
power determine economic institutions
framework
interesting result of our
an equilibrium outcome,
is
will partly or
it
entirely offset the effect of changes in political institutions. In particular, the elite will invest
more
de facto political power in democracy than in nondemocracy.
in their
In the baseline model, this effect
economic institutions
is
identical in
of invariance defined above. ^
is
sufficiently strong that the distribution of equilibrium
democracy and nondemocracy
This pattern shows that
it
—thus leading
could be mistaken to infer from
frequent changes in certain dimensions of political institutions that there
persistence.
The
result also starkly illustrates
political institutions
to the pattern
how changes
in
some
is little
can be undone by the greater exercise of de facto
specific
institutional
dimensions of
political
power by the
elite.
Even though
institutions
is
in this baseline
model the equilibrium probability distribution of economic
independent of whether the society
ability distribution
is still
''Although this setup
is
affected
is
democratic or nondemocratic, this prob-
by economic fundamentals. The comparative
natural from the view point of Latin American history,
it
is
static results
not essential to the
results.
^To be
institutions
precise, there are
is
changes
in
economic
invariant to political institutions.
institutions,
but the equilibrium distribution of economic
illustrate this.
for
The most
interesting
among
these
that the economic structure of the society,
example the presence of sectors competing with agriculture
effect
on the equilibrium. The more productive are these
from using repressive methods, and the more
Second, the smaller the numbers of the
will
is
elite,
likely
it
for labor, will
sectors, the less the elite
is
political
the more cohesive they are and the more able they
advantage created by a democracy
lead to a greater domination of politics
by the
advantage of the citizens creates a future cost
elite.
for the elite
and they are
become an absorbing
The
Finally,
for the citizens
cost.'^
willing to invest
may
more
in
However, when democratic
institutions create a sufficiently large political advantage for the citizens
"sufficiently strong"), the nature of the
favor.
This result follows because the democratic
de facto power to avoid this future
activities to increase their
have to gain
that institutions favor the citizens.
be to solve the collective action problem and choose the institutions they
and more paradoxically, the
have a major
(i.e.,
when they
are
equihbrium changes qualitatively, and democracy may
state.
invariance result, that the de facto political power of the elite can entirely offset
the effect of changes in political institutions,
is
special.
In the rest of the paper,
we extend
our baseline model in a number of ways to show how, more generally, the de facto political
power of the
elite
only partially undoes the effect of changes in political institutions, leading
to an equilibrium with a
Markov regime-switching
The two extensions we consider
structure.
allow democracy to place limits on the exercise of de facto power by the elite
their use of paramilitaries or co-option of politicians)
economic institutions
may be
difficult in
(e.g., limits
on
and introduce the feature that changing
the short run
(e.g.,
because the democratic regime
has already implemented some changes favoring the citizens). Both of these extensions lead
to an equilibrium structure where the society switches between
democracy and nondemocracy.
with different sets of economic institutions in the two regimes, and exhibits state dependence
(so that
nondemocracy
Finally,
is
we analyze a
more
likely to follow
richer
the sense that, in democracy,
it
model
is
more
in
nondemocracy than
which
it is
to follow democracy).
political institutions are
difficult for
may
durable, in
the elite to change political institutions
than economic institutions. This model leads to a phenomenon which we
democracy] the equilibrium
more
refer to as captured
feature the emergence and persistence of democracy for a
long span of time, but throughout the economic institutions will be those favoring the
elite.
In
^It is also interesting that in this baseline model, there is greater inefficiency in democracy than in nondemocracy, because in democracy the economic allocations are the same as in nondemocracy, but there is
greater exercise of de facto political power by the elite, which is costly. This result suggests some insights about
why certain potential reforms in specific political institutions in many less-developed countries may have failed
to generate significant economic growth and also perhaps about why the post-war economic performance of
democracies
may have been no
better than those of dictatorships
(e.g.,
Barro, 1997).
fact,
paradoxically, this extended
somewhat
of labor-repressive institutions
is
model predicts that the equilibrium probability
higher in democracy than in nondemocracy, motivating the
term captured democracy.
The model
broken.
It
also sheds
some
light
on how institutional persistence can be diminished or
suggests that an effective democracy requires both reforms in specific political
institutions (such as voting rules or electoral procedures), but also a
facto political
power of the
which can be achieved
elite,
ability to capture the political system, or indirectly
directly, for
way
of curbing the de
example, by reducing their
by reforming the economic structure so
that with reduced land rents, they have less incentive to thwart democracy.
The model's
insights enable us to interpret the experience of
in a different light. For
diuring the colonial era,
less
developed countries
example, in the Americas, labor repression was of central importance
and was achieved by various means including the encomienda, the
Yet repression did not end when the mita and slavery were abolished.
mita, and slavery.
It
many
continued with domination of politics by local landed
elites,
with the creation of labor
market monopsonies (Solberg, 1969, McGreevey, 1971, Coatsworth, 1974, McCreery, 1986),
and the systematic threat
of violence against peasants in rural areas.
Similarly, in the sugar
plantations of the British Caribbean, Natal or Mauritius, slavery was replaced by the use of
cheap indentured laborers from the Indian subcontinent (Tinker, 1974, Northrup, 1995). In the
impede
U.S. South, slavery was replaced by monopsonistic arrangements, policies designed to
labor mobility, political disenfranchisement, intimidation, violence and lynching.^
Our paper
(e.g.,
is
related to the literature on the persistence of institutions in political science
Steinmo, Thelen and Longstreth, 1992, Pierson, 2004, Thelen, 2004), though
this literature focuses
it
follows works
on
on how
'hysteresis'
much
specific institutions persist over long periods of time.
by David (1985) and Arthur (1989) on the lock-in of
of
In this
specific
technologies based on increasing returns. In addition to these approaches, persistence of institutions can arise in models in which social conventions or
and learning
(e.g..
norms emerge from
Young, 1998, Bednar and Page, 2006), and
specific investments in activities
in
models
in
local interactions
which agents make
whose value would be destroyed by changes
in social arrange-
in the paper beg the question of how the elite are able to exercise
democracy. This is also discussed in detail in Section 7, where we present a number
of historical case studies illustrating the pattern of persistence modeled here and also emphasize two specific
channels: the capture of the party system by the elites and the threat of violence. Both these methods were
This discussion and the general approach
de facto political power
in
extensively used in the U.S. South after the Civil War and are still present in many Latin American countries
such as Brazil, Bolivia or Colombia. For the U.S. South after the Civil War, see Key (1949), Woodward (1955),
Wright (1986), Alston and Ferric (1999), and Ransom and Sutch (2001), for Colombia, see Dix (1967), Wilde
(1978), Hartlyn (1988)
and Kline (1999), and
for Brazil, see Chilcote (1990)
and Hagopian (1996).
ments
(Dixit, 1989a,b,
Coate and Morris, 1999). Institutions could also
existence of multiple steady-state equilibria
(e.g.,
Krugman,
1991,
persist because of the
Matsuyama,
The
1991).
popular idea that economic inequality or certain forms of natural resource endowments
the balance towards bad institutions
and state dependence), since
is
also different
None
2005).
from our notion of persistence (invariance
this idea stresses the persistence of
then lead to the persistence of institutions
(e.g.,
tilt
Engerman and
of these approaches have addressed the issues
we
economic characteristics that
Sokoloff, 1997,
Benabou, 2000,
discuss here, in particular, the
coexistence of persistence and change.
From a modeling
point of view, this paper extends the framework in Acemoglu and Robin-
son (2000, 2001, 2006a), where de facto political power drives changes in political institutions
and the future distribution of de jure
model the process of the
elite investing in their
significant differences in the results.
more
"pro-citizen"
,
political power. ^
The major
difference
is
that
we now
de facto political power, which leads to some
While our previous work emphasized that democracy
the analysis here shows this
may
not be the case
if
the elite are able to
garner sufficient de facto political power in democracy. ^'^ In this respect, the current paper
and Sala-i-Martin (2004) and Mulligan and Tsui
related to Mulligan, Gil
on similarity
of various policies
(2005),
of our terminology, they explain this similarity by lack of significant de jure
in
is
which focus
between democracies and nondemocracies, though,
between regimes, while our model emphasizes how changes
is
in
terms
power differences
de facto power can undo real
changes in de jure power.
The
rest of the
political
paper
environment.
is
organized as follows. Section 2 outlines the basic economic and
Section 3 characterizes the equilibria of the baseline model, and
tablishes the invariance result
framework
in
and the main comparative
a number of directions and shows
partial offset will occur,
and the equilibrium
will
Section 4 generalizes this
statics.
how under more
general circumstances, only
correspond to a Markov regime- switching
model, with fluctuations between democracy and nondemocracy.
model
in
stitutions
which changing
political institutions
is
more
difficult
Section 5 introduces the
than influencing economic
and shows how an equilibrium pattern of captured democracy can
elites dictating their favorite
how simultaneous reforms
economic institutions
in multiple
es-
in
arise
in-
with landed
democracy. Section 6 briefly discusses
dimensions of political institutions or economic institu-
^See also Ticchi and Vindigni (2005), Jack and Lagunoff (2006), and Lagunoff (2006) for related approaches.
among others, Austen-Smith (1987), Baron (1994) and Grossman and Helpman (1996) on models
models where the equilibrium policy in a democracy is affected by lobbying. Our approach is more reducedform, but explicitly models the incentives of individual agents to contribute to lobbying- type activities, is
dynamic and endogenizes not just policies but also institutions.
'"See,
tions can be effective in breaking the cycle of persistence in economic institutions. Section 7
discusses a
number
of historical case studies that
both motivate and substantiate the ideas
in
the paper. Section 8 concludes.
Model
Baseline
2
Demographics, Preferences and Production Structure
2.1
Consider an infinite-horizon society in discrete time with a unique
by a continuum
1
of worker/citizens
and
(a finite)
M>
number
the same risk-neutral preferences with discount factor
/3,
1
good and populated
final
of the elites. All agents have
given by
oo
at time
We
t
where
cj_^
•
use the notation
All workers
elite
i
where
denotes consumption of agent
i E.
own one
£
to denote an elite agent,
unit of labor,
i
at time
and
which they supply
to
L
produce the unique
is
a
in the
a normalization)
economy, with no alternative use, and each
.
There
elite
is
owns
F
maximum
which each producer runs into severe diminishing returns (where the
is
citizen.
Each member
denotes land and N^^ denotes labor used by this producer, and
returns start after land size of L/Ad
to
j in terms of the final good.
inelastically.
returns to scale. This production function implies that there
after
-|-
£ C to denote a
i
E £ has access to the following production function
U
f
final
of the
good:
exhibits constant
land size of
L/M
fact that diminishing
a total supply of land equal
L/M
units of land (and no
labor)."
The
final
good can
also
be produced with an alternative technology, which can be
in-
terpreted as small-scale production by the laborers themselves (or a low productivity proto-
industry technology). This alternative technology exhibits constant retinrns to scale to labor:
Ya "The
market to save space.
is
introduced to prevent an allocation in which
free-rider
same reason, if initially
become concentrated in
(3)
all land is owned by one individual,
problem in investment in de facto political power, explained below. For the
there were M' >
land owners, given the production function in (2), land would
the hands of
land owners. We do not explicitly discuss transactions in the land
diminishing returns
which would solve the
ANA.
M
M
good
Clearly, total output of the unique final
the market clearing condition for labor
economy
in the
will
be
y=
Ylies '^l'^^a, and
is
Y.NI + Na<1.
The main
this
role of the alternative technology, (3), will
(4)
be to
restrict
In the
first,
how low wages can
fall
in
economy.
We
consider two different economic institutions.
Given
tive}'^
since
F
When
(2),
each
elite will hire
N1 — Nl/M
exhibits constant returns to scale,
we denote by r
is
== 1,
Nl =
The return
c denotes "competitive".
1
—
iV^,
and
as:
the wage rate (and the
of each worker), as a function of labor allocated to this sector,
where the superscript
markets
units of labor, where
we can write per capita output
there are competitive labor markets, which
wage earnings
labor markets are competi-
Ni,
is
therefore:
to landowners with competitive
similarly
R^m^f\-^^,
with each landowner receiving
Assumption
(7)
R^L/M.
1
/(L)-L/'(L)>A
This assumption implies that even when
wage
in this sector
is
Nl =
1
(i.e.,
when L/Nl =
L), the competitive
greater than the marginal product of labor in the alternative technology.
Therefore, both the efficient allocation and the competitive equilibrium allocation will have
workers allocated to the land sector,
wage and
i.e.,
N^ =
all
In light of this, the relevant competitive
1.
rental return on land will be
w''
=
w'^INl
=
I]
=
f (L)
-
Lf {L)
(8)
and
R''
'This implies that, by law, landowning
= R'[Nl =
elites
cannot
l]
=
f' {L)
restrict their labor
(9)
demand
to affect prices.
Consequently, factor prices at time
w
{Tt
=
The
1)
^w"
and Rt
= R {n =
t
1)
as a function of
=
/^^ with
economic institutions are given by Wt
w" and
R'' as
defined in (8) and {9)}^
alternative set of economic institutions are labor repressive {rt
landowning
elite to
slavery
(i.e.,
is
institutions are labored oppressive, the lowest
Nl >
0)
and allow the
0, is
They
not allowed), so workers always have
wage that the
when economic
Consequently,
access to the alternative small-scale production technology.
ensuring that
=
use their political power to reduce wages below competitive levels.
cannot, however, force workers to work
—
elite
can pay the workers, while
still
A. This imphes that factor prices under these economic institutions
are
=
w''
A,
(10)
and
R'^f'-j-^.
(11)
Ju
(Recall that the landed elite are paying the
wage of
economic institutions are labor repressive, then we
R {rt —
0)
— R^ Assumption
.
1
A
will
to a total of
have wt
immediately imphes that R^
economic institutions wages are kept
greater rents. For future reference,
artificially low, i.e., w'^
we
Nl =
— w [rt —
When
— w^ and Rt =
workers).
1
0)
>
R'^,
since with labor repressive
<
w'^,
so that land owners enjoy
define
AR = R'-R"
f{L)-A f (L) >
L
One
feature to note
tive labor
is
0.
(12)
that the simple environment outlined here implies that both competi-
markets and labor repression
will
generate the same total output, and will differ only
in terms of their distributional implications.
Naturally,
costs from labor repressive economic institutions,
distortions or other costs involved in monitoring
it
is
possible to introduce additional
which may include standard monopsony
and forcing laborers
to
work
at
below market-
clearing wages (such as wasteful expenditures on monitoring, paramilitaries, or lower efficiency
of workers because of the lower
"More
payments they
formally, the second welfare
equilibrium
is
receive). Incorporating such costs has
theorem combined with preferences
in (1) implies that
no
effect
a competitive
a solution to the following program:
max
/(^^Nl + ANa
subject to (4) and L < L. Assumption 1 ensures that the solution involves
equilibrium factor prices are given by the shadow prices of this program.
Nl =
1
and L
=
L,
and the
on the
analysis,
and throughout, one may wish to consider the labor repressive
institutions as
corresponding to "worse economic institutions"
Political
2.2
Regimes and De Facto
There are two possible
political regimes,
denoted by
and nondemocracy. The distribution of de jure
regimes.
At any point
which designates the
Power
Politiccd
D
and
corresponding to democracy
A^,
power
political
in time, the "state" of this society will
will
vary between these two
be represented by
S(
e {D,N},
regime that applies at that date. Importantly, irrespective of
political
the political regime (state), the identity of landowners and workers does not change; the
M individuals control the land, and have the potential to exercise additional
Overall political power
power. Since there
is
is
political
determined by the interaction of de facto and de jure
a continuum of
citizens,
power.
political
they will have difficulty in solving the collective
we
action problem to exercise de facto political power. Consequently,
as being exogenous rather than
same
stemming from
their
own
treat their de facto
power
contributions.
In contrast, elites can spend part of their earnings to gather further de facto political power.
In particular, suppose that
elite
Then
increasing their group's de facto power.
Zt
—
Y2ie£
^t
'
^^^
assume that
"^6
>
(p
The reason why
0.
their de facto political
activities is that there
that their
An
own
is
a finite number,
activities will
for the elite is the
Even though the
is
(13)
choose to spend a positive amount on such
of them, so each of
them
will
take into account
same
in
is
that the technology for generating de facto political
democracy and nondemocracy.^"*
citizens cannot solve the collective action
problem to invest
facto political power, since they form the majority in society they always possess
power.
The
be
contribution to total spending, Zt, will have an effect on equilibrium outcomes.
important assumption implicit in (13)
power
M,
power
on such
</>Zt,
may
the elite
as & contribution to activities
total elite spending
Pf =
where
9\>Q
^ £ spends an amount
i
extent of this power depends on whether the political regime
nondemocratic.
We
model the
citizens' total political
power
in
is
in their
some
de
political
democratic or
a reduced-form manner as
follows:
Pf =
"There may be a number
taries
may be more
of reasons for
why
ujt
+
vl
{st
= D)
,
(14)
the ehte's ability to lobby and bribe politicians or use paramili-
restricted in democracy, so in Section 4,
and nondemocracy.
10
we allow
this technology to differ
between democracy
where
is
utt
distribution
Sj
=
a random variable drawn independently and identically over time from a given
F {)
and measures
D, such that I
measuring
citizens'
= D) =
[st
I
while /
[st
— N) =
and
0;
n
is
rj
is
an indicator function
for
strictly positive pai'ameter
de jure power in democracy.
There are two important assumptions embedded
facto pohtical
— D)
their de facto power; I {st
power of the
The second assumption
is
and
citizens fluctuates over time,
that
when
The
in equation (14).
the pohtical regime
hard to predict
is
democratic,
is
first is
i.e.,
in advance.-'^
=
St
that the de
D,
citizens
have greater political power. This represents in a very simple way the fact that democracy
allocates de jure political
power
in favor of the majority.
This
will
be both because of the
formal rules of democracy and also because in democratic politics, parties
Put
the collective action problem of the citizens.
democracy the
To
stochastic dominance.
Assumption
power of the
political
F
2
is
simplify the discussion,
defined over (w, oo) for
everywhere).
w* such that
/' (w)
lim^^^cxD
=
/
(i^)
All of the features
Moreover, /
>
for all
partly solve
equation (14) implies that in
citizens shifts to the right in the sense of first-order
uj
{cu)
we make the
some w <
twice continuously differentiable (so that
/', exist
differently,
may
its
is
0, is
following assumptions
everywhere
on F:
strictly increasing
and
density / and the derivative of the density,
single
< w* and
peaked
<
/' (w)
(in the sense that there exists
for all
ui
>
uj*)
and
satisfies
0.
embedded
in
Assumption
2 are for convenience,
and how relaxing them
affects the equilibrium is discussed below.
We
introduce the variable nt G {0, 1} to deiiote whether the elite have more (total) political
power. In particular,
and
will
more
make the key
political
decisions are.
We
ttj
—
t,
and the
In contrast, whenever
will
make
description of the environment,
Tt,
and what the
Pf <
elite
Pf,
have more political power
nt
=
I
and
citizens have
the key decisions.
assume that the group with greater
institutions at time
When
decisions.
power, and they
To complete the
sions,
when Pf > Pf, we have
political
it
remains to specify what these key
political
power
will decide
both economic
regime wih be in the following period, St+i-
the elite have more political power, a representative elite agent makes the key deci-
and when
citizens
have more political power, a representative citizen does
political preferences of all elites
and
all
so.
Since the
citizens are the same, these representative agents will
always make the decisions favored by their group.
is used extensively in Acemoglu and Robinson (2006a), and defended there. Briefly, given
whether and how effectively citizens will be able to organize is difficult to predict in advance,
change from time to time. The randomness of ujt captures this in a simple way.
'"This assumption
their large numbers,
and
will
11
Timing of Events
2.3
We now
environment.
briefly recap the timing of events in this basic
At each date
t,
society starts with a state variable
G {D, N}. Given
St
the following
this,
sequence of events take place:
1.
Each
power
political
2.
agent
elite
£ £ simultaneously chooses how much to spend to acquire de facto
i
>
for their group, 9]
The random variable cot
0,
and
drawn from the
is
Pf
determined according to
is
distribution F,
and
(13).
Pf is determined according
to (14).
3.
4.
If
Pf > Pf
(i.e., TTt
—
a representative
0),
Pf < Pf
—
randomly chosen)
(e.g.,
elite
agent chooses
(Tt,st-|_i),
and
Given
transactions in the labor market take place, Rt and Wt are paid to elites and
r^,
if
(i.e., Tit
1),
a representative citizen chooses (rt,St+i).
workers respectively, and consumption takes place.
5.
The
following date,
i
+
starts with state st+i-
1,
Analysis of Baseline
3
We now
Model
analyze the baseline model described in the previous section.
symmetric R-Iarkov Perfect Equihbria (MPE).
strategies are
mappings from payoff-relevant
particular, in
an
MPE
above the influence of
this past history
(tt)
and
.
s' (tt)
and equilibrium
political power,
^ for
elite
Symmetric
form 6
s'
(n)
MPE
(s), i.e.,
€
{D,N}
s
G {D,N}. In
on the past history of the game over and
s.
An
MPE will
consist of
agent as a function of the political state, and
as a function of
tt
factor prices as given
6
{0, 1}
by
denoting which side has more
Here the function r
(8)-(ll).-^^
determines the equilibrium decision about labor repression conditional on
the function
focus on the
imposes the restriction that equilibrium
on the payoff-relevant state
each
first
which here only include
states,
strategies are not conditioned
contribution functions {^' (s)}
decision variables r
An MPE
We
(tt)
who has power and
determines the pohtical state at the start of the next period.
will in addition
impose the condition that contribution functions take the
do not depend on the identity of the individual
elite,
i.
Symmetry
is
a natural
since
it is
(tt, 5) and s' {n,s), so that the choice of economic institutions and future
on which party has political power, tt, and the current state, s. Nevertheless,
clear that the current state will have no effect on these decisions, we use the more economical notation
T
and
s' {n).
^^More
generally,
we could have r
political institutions are conditioned
(tt)
12
We
feature here, and simplifies the analysis.
A
more formal
The
focus on
elite
AiPE
MPE
is
the
elite.
MPE
completeness telow.
for
also given below.
natural in this context as a
is
among
action problem
the
an
definition of
discuss asymmetric
Looking
at
way
subgame
of
modehng
the potential collective
perfect equilibrium (SPE) will allow
greater latitude in solving the collective action problem by using implicit punishment
strategies.
We
briefly analyze
SPEs
in subsection 3.3.
Mciin Results
3.1
The
date
MPE can be characterized by backward induction within the stage game at some arbitrary
t,
given the state s 6 {D, TV}. At the last stage of the game, clearly whenever the
political power,
and a
political
whenever
i.e., tt
=
0,
they will choose economic institutions that favor them,
system that gives them more power in the future,
citizens have political power,
i.e.,
tt
=
1,
=
s'
i.e.,
=
they will choose r
i.e.,
r
have
=
0,
In contrast,
A''.
1
elite
and
s'
=
D}'^
This implies that choices over economic institutions and political states are straightforward.
Moreover the determination of market prices under
been specified above
(recall
equations (8)-(ll)).
different
Thus the only remaining
contributions of each elite agent to their de facto power,
be summarized by a
to characterize the
V (A)
for
d].
MPE
by writing the payoff to
elite
agents recursively, and for this reason,
value of an elite agent in state s by
V {s)
(i.e.,
V {D)
to 6 (A). Consequently,
i E.
when agent
have political power
P^
We
{e\
MPE,
democracy
suppose that
£, have chosen a level of contribution to de facto
i
E £ chooses
6^,
their total
power
P^ {9\e{N)\N)^cj){{M-l)e{N) +
elite will
for
nondemocracy).
other elite agents, except
The
MPE can
be convenient
(s). It will
Let us begin with nondemocracy. Since we are focusing on symmetric
all
decisions are the
Therefore, a symmetric
level of contribution as a function of the state
we denote the equilibrium
and
economic institutions has already
will
power equal
be
9').
if
9{N)\N)
=(l)
((M -1)9 (A) +
9')
>
wj.
is the same in
and s' = N. Throughout, we
use the tie-breaking rule that, when indifferent, citizens choose s' = D, and we impose this in the analysis.
Alternatively, in Section 5, equation (49) introduces more general preferences for the citizens, whereby they
receive other benefits from democracy, denoted by v{D). In that case for any u (D) > 0, s' = D is always
strictly preferred for the citizens. We do not introduce these preferences now to simplify the analj'sis until
will see in
Proposition
1
democracy and nondemocracy, so
Section
that the equilibrium distribution over economic institutions
citizens will
be indifferent between
5.
13
s'
=
D
Expressed
differently, the probability that the ehte will
have political power in this state
p{0\eiN)\N)=F{4>{{M-l)e{N) +
We
can then write the net present discounted value of agent
i
is
9^)).
(15)
e £ recursively as
= max I -9' +p(9\e{N)\N)(^ + l3ViN\e(N), 9(D))]
V{N\e{N),9{D))
'
e'>o
M
\
t
J
+ {l-p{9\9[N)\N)) (^^+13V{D\9{N),9{D))^Y
where
is
R^
recall that
is
the rate of return on land in competitive markets, given by
(9)
(16)
and K^
The
the rate of return on land under labor repressive economic institutions, given by (11).
function
V {N
when
other elite agents choose contributions 9 {N) in nondemocracy and 9 (D) in democracy.
all
V {D
Similarly,
The form
{N) 9 (D)) recursively defines the value of an
9
\
\
{N) 9 (D))
9
,
given his contribution
remain
will
the value in democracy under the same circumstances.
is
of the value function in (16)
because of the expenditure
in the
9^
0'',
is
intuitive. It consists of the forgone
plus the revenues and the continuation values.
and those of other
hands of the
agent in nondemocracy
elite
,
elite
institutions will be labor repressive,
elite
In particular,
agents in nondemocracy, 9 {N), political power
with probability p [9\d {N)
and
consumption
this elite agent receives
|
A'')
,
which case economic
in
revenue equal to K" L/M (rate
L/M) and
of return under labor repressive economic institutions, BJ' times his land holdings,
,
the discounted continuation value of remaining in nondemocracy, jiV
probability
— p [9\9{N)
1
and labor markets are competitive. In
to
R^L/M
Agent
i
change the
E £ chooses
9^
to
g
pTV
j^
Since
is
p (0\ 9
^^^
F
^
^^>jj
-g g^j^
\
political
maximize
pohcy function (correspondence)
Qi
this case a
and continuation value j3V [D
citizens will choose to
for the
\
(N) 9 {D)). With
9
,
9 (TV)
,
member
9 {D)), since with
system to St+i
his net
power
in their hands, the
D.
expected present discounted
maximization
optimal policy
—
(TV)
TV),
in (16)
be
T^
[9
{N) 9
,
utility.
(£*)],
Let the
so that- any
for the value function in (16) (in state s
which implies a particularly simple
1
of the elite receives revenue equal
—
TV).
continuously difTerentiable and everywhere increasing (from Assumption
is
=
have greater political power, so they choose r
TV), citizens
\
{N
2), so
first-order necessary condition for (16):
I
4>f{cP{{M-l)9{N) +
9'))
(^^ + (3(V{N\9{N),9{D))-V{D\9iN),9iD)))^
<
1,
(17)
and
(12),
0*
>
0,
and /
^*That
is,
with complementary slackness,'^ where
is
recall that
AT?
=
R^ —
the density function of the distribution function F. Moreover,
either 5'
=
or (17) holds as equahty.
14
R"^ is
it is
defined in
clear that
we
need the additional second-order condition that
wliy the maximization problem for individual
6'
that
V (N
does not affect
maximand
is
any
differently,
9^
consumption, which
which
from
this contribution,
more
political
6*'))
<
in this recursive formulation
V [D
{N) ,9{D)),so
9
\
The reason
0.-^^
is
so simple
is
differentiability of the
from
this political power,
focusing on a symmetric
de facto power by the
(17)
,
first-order condition
side of (17),
citizens,
which
ARL/M plus
(N) 9 [D)] must solve
is
and
satisfy the correspond-
quite intuitive:
must be equal to
the cost of forgone
(or less than) the benefit
the marginal increase in the probability of the
is
power than the
[9
hand
the right
is
T^
e
The
ing second-order condition.
in
i
+
1)9 {N)
guaranteed.
Expressed
direct benefit
(N) ,6 (D)) or
9
\
—
/' (0 (^{M
is
i.e.,
(pf
{•),
elite
and the benefit that the agent
will derive
the second term on the left-hand side, consisting of the
the benefit in terms of continuation value. Moreover, since
MPE,
9^
having
>
is
>
equivalent to 9 {N)
0,
so
if
there
is
we
are
any investment
then (17) must hold as an equality.
elite,
Next, consider the society starting in democracy. With the same argument as above, the
elite will
have political power
P^
which only
differs
if
9{D)\D)^ct> {{M -1)9 {D) +
{9\
from the above expression because with
tional advantage represented by the positive parameter
will
capture political power in democracy
p {9\
=
+
ut
D, the
Then the
rj.
77,
citizens
have an addi-
probability that the elite
1) e
+ 9') -
{D)
ry)
(18)
,
as before, the value function for elite agent
i
£ £
is
= max!.-9'+p(9\9{D)\D)(^ + f3V{N\9{N).9{D))
V{D\9{N),9iD))
which has
st
>
is
9{D)\D)=F (0 ((M -
and using the same reasoning
9')
w>o
[
+
-p{9\9{D) D))
(1
'
\
M
\
+ PV (D
(^^
\9
iN),9{D)))\
(19)
first-order necessary condition
cpf{cp{{M-l)9{D) +
9')-il)(^^ + /3{V{N\9{N),9{D))-V{D\9{N),9{D)))) <
1,
(20)
and
/'
0'
[(f>
>
0,
((M —
again with complementary slackness and with second-order condition
1)6' (A'^)
+
0')
—
the maximization in (19) by
'"The condition /' (0 ((M
necessary but not sufficient.
?])
T^
<
[9
0.
Denote the policy function (correspondence) imphed by
(N) 9 (D)], so that any
,
-1)9 {N) + 9')) <
We
impose the
is
sufficient,
sufficient condition
15
9'
6
T^
[9
[N)
,
9 {D)] solves (20).
< would be
{(j> ((M -1)6 (N) + 9'))
throughout to simplify the discussion.
while /'
Consequently, denoting the decision of current economic institutions by r
system by
political
Definition 1
e
A
we can have the following
s' (tt),
MPE
symmetric
T^
[9
0)
— N,
T
=
{tt
=
1)
1
and
5' {tt
—
T^
(N) 9 (D)] and similarly 9 {D) e
,
In addition, economic and political decisions r
=
symmetric MPE:~°
consists of a pair of contribution levels for elite agents
(N) and 9 (D), such that 9 {N) e
s' {tt
definition of a
and future
(tt)
I)
=
(tt)
and
D, and
s' (tt)
[9
are such that t
factor prices are given
(N) 9
,
{tt
by
—
0)
(D)].
—
0,
(8)-(ll) as a
function of r € {0, 1}.
This definition highlights that the main economic actions, in particular, the investments
by
in de facto power, are taken
elite agents,
so the characterization of the
MPE
will involve
solving for their optimal behavior.
MPE,
In a symmetric
9 {N)
,
must be given
9^
that solves (17) must equal 9{N), thus
{^ +(5V{N\9{N),9
and similarly the equilibrium condition
{<PM9 {D)
Given Definition
and 9 {D) >
-
1,
strictly positive,
by:
0/ {cf>M9 {N))
4>f
when
r/)
for 9
{D))
-
{D) (when
(iV
{D \9{N)
,9
{D))^
=
1,
(21)
=
1.
(22)
strictly positive) is
(^^ +PV{N\9{N),9 {D)) -PV{D\9{N),9 {D))\
these two equations completely characterize symmetric
MPEs with 9 (TV) >
0.
Comparison
of (21)
and
(22)
immediately implies that
9{D) = 9{N)
Moreover inspection of
(21)
and
(22),
+
^.
combined with the
(23)
fact that
F
is
continuously difFeren-
tiable, yields the invariance result:
p{D)=p{9{D),9{D)\D)=p{9{N),9{N)\N)=p{N),
which also defines p {D) and p {N) as the respective probabilities of the
taining) political
Intuitively, in
power that they
power
in
(tt)
and
democracy the
elite invest sufficiently
more
entirely offset the advantage of the citizens
s' (tt),
gaining (or main-
democracy and nondemocracy.
^"This definition incorporates the best responses of
tutional, T
elite
(24)
elites
for convenience.
16
to increase their de facto political
coming from
their de jure power.
and citizens regarding economic and political
insti-
A
more
technical intuition for this result
is
that the optimal contribution conditions for elite
agents both in nondemocracy and democracy equate the marginal cost of contribution, which
is
always equal to
1,
to the marginal benefit. Since the marginal costs are equal, equilibrium
The marginal
benefits in the two regimes also have to be equal.
ARL/M,
mediate gain of economic rents,
benefits consist of the im-
plus the gain in continuation value, which
also
is
independent of current regime. Consequently, marginal costs and benefits can only be equated
a p{D)=p{N)
It is also
as in {24).
straightforward to specify
when there
In particular, the following assumption
be positive investment in de facto power.
will
sufficient to
is
ensure that the equilibrium will have
positive contribution by elite agents to de facto power:
Assumption 3
ARL]
,,„, ARL
,,
mm|^/(O)^^,0/(-^)-^|>l.
.
Since
this
V (N) — V {D) >
r
,
,
,
,
(by virtue of the fact that the elite choose nondemocracy),
assumption ensures that in both regimes, an individual would
contribution even
if
nobody
also exist equilibria in
else is
which the
doing
elite
assumption
If this
so.^^
make no
is
like to
make a
positive
may
not satisfied, there
contribution to increasing their de facto power
(see Corollary 1).
Proposition
1
(Invariance) Suppose Assumptions
there exists a unique symmetric
MPE.
1-3 hold.
is
=
and 6
Since Assumption 2 implies that / (w)
The
comparison of these two
follows from
Assumption
result that
equalities,
2,
isfy (21)
and
Assumption
2,
/ (w)
is
p {D)
=
p (N)
F
is
F {4>Me [D) -
(A^)
some
which establishes
which imposes that
so for any interior 9 {D) and 9 {N),
from Assumption
6 (0,1), so
non- degenerate and independent
=
cannot be part of an equi-
continuous and lim^^^oo /
is
conditions (21) and (22) must hold as equalities for
establishing existence.
is
— p{N)
model,
democratic or nondemocratic.
Proof. Assumption 3 ensures that 6 {D)
librium.
in the baseline
This equilibrium involves p{D)
that the probability distribution over economic institutions
of whether the society
Then
>
interior values of 9
(24).
The
fact that
3 also implies that
ry
< — oj
<
and
(where
/'
^ F {(pM9 {N)) <
{(pM9 {D)
recall that
17
w <
0,
both
{D) and 0{A^),
-
77)
0); see
<
p{D)
= p{N) <
throughout
single peaked, so only a unique pair of 9 (D)
(22) with /' {(f>M9 [N])
=
then follows immediately from the
strictly increasing
i])
(^')
1.
its
1
support,
In addition, again
and 9 {N) could
for given
sat-
V (N) - V {D).
condition (26) below.
The
V (N) — V {D) —
fact that
i]/ {(f)M)
estabhshes the uniqueness of the symmetric
This proposition
is
uniquely determined (from equation (24)) then
is
MPE.
one of the main results of the paper.
shows that there
It
librium changes from democracy to nondemocracy and the other
the fact that the equihbrium probability distribution
non-degenerate,
is
(this follows
i.e.,
of de jure power, but they
do not translate into changes
tutions and economic allocations,
invariance in equilibrium; even
p{D) = p[N) e
i.e.,
in the
we have p [D) — p
law of motion of economic
{N)?"^ This
when shocks change the
insti-
the sense in which there
is
political institutions, the probability
distribution over equilibrium economic institutions remains unchanged. This result also
trates
from
Moreover, by assumption these changes in political institutions affect the distribution
(0, 1)).
is
way round
be equi-
will
how
institutional change
and persistence can coexist
—while
illus-
change
political institutions
frequently, the equilibrium process for economic institutions remains unchanged.
Remark
1
As
will
be discussed further below, the invariance
assumptions. Section 4 will show that
de facto power
for the elite in
when
Other assumptions implicit
result are: (1) that
on functional form
there are differences in the technology of generating
democracy and nondemocracy or when economic
costly to change in the short run, de facto
partially.
result relies
democracy
power
will only offset the
in our analysis that are
shifts the
power of the
change
institutions are
power
in de jure
important for the invariance
than
cj
being
Fq
first-
order stochastically dominating F/v); (2) that the technology of de facto power for the
elite,
drawn from general
equation (13),
is
nondemocracy and Fp
distributions F/v in
linear.
When
citizens additively (rather
in
democracy, with
either of these assumptions are relaxed,
we continue
to obtain
the general insight that endogenous changes in de facto power (at least partially) offset declines
in the
de jure power of the
Remark
elite,
but not necessarily the invariance
1.
For example,
sions in Proposition
1
if
we
relax the single-peakedness assumption
would continue to apply, except that the symmetric
be unique. Multiple equilibria here are of potential
which expectations of future behavior
if
relaxed,
See Section
4.
2 Assumptions 2 and 3 can be relaxed without affecting the basic conclusions in
Proposition
Also,
result.
affects current
the parts of Assumption 2 that
we may obtain corner
transitions to
from investing
F
solutions,
is
behavior
whereby p{A^)
the conclu-
MPE may no longer
they correspond to situations in
(see, e.g.,
Hassler et
= p[D) =
(essentially
1,
al.,
when
18
total
2003).
=
because returns to individual
power
shifts
are
and there would be no
power may remain high, while the probability of a
'Yet, naturally, economic institutions will change
(cj),
increasing everywhere and limu,^oo / (w)
democracy from nondemocracy
in de facto
interest, as
on /
elites
sufficiently high
from one group to another.
level of
LJ
becomes
(^N)
=
p {D)
result
is
interesting in this context.
p
Corollary
1
=
Alternatively,
0).
Assumptions
0.
Suppose there
that Assumptions
1
and
2
Assumption 3
if
and 3
exists ^(A'')
2 hold,
>
equilibria with
The
following
such that
and that
> -u.
(26)
model, there exists a symmetric
in the baseline
we can have
relaxed,
rule out these "corner" equilibria.
r?
Then
is
MPE in which p (N)
G
(0, 1)
and p {D)
—
0.
Proof. Suppose there
R^L/
{{1
—
P) M), while V"
{N)
condition for
obtain
exists a
e{N) =
>
is
(A'')
Now
using (21) and (22),
to de facto power in
/
(-77)
=
0,
'
we
(
jj^
^
Combining
this
Then we have
0.
with the expression
^ F[cj>Me{N))^RL/M -
'
V (D) =
V {D),
for
see that (25)
nondemocracy
is
e[N)
sufficient to ensure that positive contribution
is
optimal for
elite agents.
Moreover, (26) implies that
thus
F (—77) =
p{D) = 0. m
(26),
Therefore,
sorbing state
when
0,
if
we relax part
may
rj
is
is
also optimal for the elite.
which establishes the existence of a symmetric
arise.
of
Assumption
3,
Clearly, Condition (26),
high. This implies that
in favor of the citizens,
it
may
if
symmetric
Moreover, again from
MPE with p (N)
MPEs
e
(0, 1)
which leads to this outcome,
is
more
likely
democrac}' in fact creates a substantial advantage
destroy the incentives of the
elite to
engage
in activities that
and
institutions.
It is also
p {D)
and
with democracy as an ab-
increase their de facto power, and thus change the future distribution of political regimes
economic
we
- PF{(pM9{N))
I
so that zero contribution in democracy
to hold
=
and
y
y (j^n
with p [D)
given by (16), and the relevant first-order necessary
given by (21).
9 {N) as in (25),
^
still
is
MPE
symmetric
interesting to note that even
— p {N) >
when Condition
characterized in Proposition
1
19
may
(26) holds, the equilibrium with
still exist,
leading to a symmetric
MPE
= p (N).
with p {D)
Consequently, whether democracy becomes an absorbmg state
may depend on
consolidated),
expectations.
Finally, inspection of the proof of Corollary 1
Assumption
With
3
A
There
exists
tion 3 since, despite being
Non-Symmetric
3.2
We now
more
do
this,
we
first
is
and
the results continue to hold, though
restrictive,
it is
to:
we
prefer
Assump-
simpler and more transparent.
MPE
result obtains without the restriction to
MPE
extend our treatment above and define an
Without symmetry, the power
i
all
shows that Assumption 3 can be relaxed
satisfying (25),
show that the same invariance
MPE. To
agent
0{N) >
assumption,
this modified
fully
(i.e.,
of the elite in
nondemocracy
£ £ and the distribution of contributions by
all
symmetric
more
generally.
as a function of contribution
other agents, 9~' (TV)
=
{^-^
by
6'^
{N)] .^
.
,^,
given by
Similar to before, in nondemocracy the elite will have political power with probability
p{e\e-^{N)\N)^F[<p(
In democracy, with the
same reasoning
0^{N)
+ 6A\.
as before, this probability
p{9\e-^[D)\D)^F(<pi
V
The
E
E
given by
is
+ e^\-7A.
9^[D)
\j?:£,]¥^^
J
elite implies
be indexed by
i.
that value functions can also differ across
(28)
J
possibility that different individuals will contribute different
power of the
(27)
amounts
agents,
elite
Therefore, the net present discounted value of agent
i
to the de facto
G
and must
£^ is
V'{N\9'^{N),e-'{D))
= max
9'
+ {l-p{9\9-'
Here
elite
V (TV
|
_,
+ p {9\ 9-'
{N)
,.„
(29)
,
(N)
...
N)
.
i
I
TV))
I
also
R'L
——
+ 13V' {N
(^ +/3V^ {D
0"' (TV) ,0~' (D)) denotes the value of agent
9'' (TV)
9-' (A^)
,
0-
I
i
in
agents choose contributions 9"' (N) in nondemocracy and
20
\
,
9-' (D))
(Z?)))
|
.
nondemocracy when
9"''
(D)
in
all
other
democracy. Similarly,
9~^ (N)
V^ (D
,
0^' (D))
the corresponding value in democracy for agent
is
equation
for this
Agent
is
£ £ chooses
i
6^
maximize
to
{N) 0-' (D)] so that any
[e-'
symmetric
similar to that for (16) in the
his net
pohcy function (correspondence) of agent
rf
i.
The
intuition
I
,
€
6'
,
expected present discounted
utility.
Let the
that solves the maximization in (29) be given by
i
[9''
{N) 9"' (D)]
is
,
an optimal policy
the value
for
we have
function in (29). Similarly,
9'' {N)
V' [D
Pf
case.
,
9-' (D))
(30)
I
= m^L9'+p{9',9~^{D)\D)(^^+pv'{N\9-HN),9-'{D))'^
+ {l-p {9\ 9-^
and
the set of maximizers of this problem be
let
general definition of
Pf
MPE
that r
(tt
=
0~' (Z?)]
,
,
0)
=
9-^ (D)))
I
{N) 9~^
[9~^
,
(L>)]
Then we have the more
.
i
e
S, 9'
{N) G
Pf
for elite agents
0, s' (tt
=
=
0)
A',
r
(tt
=
—
1)
1
[9'^
i^)}ip£
[9'' {N) ,9'' {D)] and similarly 9' {D) G
In addition, economic and political decisions r
.
|
as:
{9' (D)}^^^, such that for all
[(9~' (A'^)
Ff
{D r' [N]
An MPE consists of a pair of contribution distributions
Definition 2
and
\D))[j^+ aV^
(D)
and
=
s' (tt
1)
=
and
(tt)
are such
s' (tt)
D, and factor prices are
given by (8)-(ll) as a function of r £ {0, 1}.
Proposition 2
Then
(Non- Symmetric
in the baseline model,
any
MPE and Invariance)
MPE involves p (D) = p (A)
Suppose Assumptions 1-3 hold.
G
(0, 1).
Proof. See Appendix.
The only
we
know
also
that the total contributions
agents
elite
between symmetric and non-symmetric
In non-symmetric
elite agent.
some
difference
may be
MPE,
expected
implies that in non-symmetric
this
may
the
elite will
in
symmetric
MPE
be equally divided among each
not be the case, and depending on expectations,
and consequently do, contribute more than others. This
to,
MPE,
made by
MPE is that
different levels of
p (D)
= p (A)
can
arise in equilibrium.
Nevertheless, the important conclusion that the probability of the elite dominating political
power and imposing
form of
political institutions remains.
symmetric
MPE.
CoroUciry 2
ff'
(D)
their favorite
=
9''
Before doing
so,
economic institutions
Given
+
ri/(f),
and
for
independent of the underlying
this result, in the rest of the
paper we focus on
however, we can also note the following result:
Among non-symmetric MPEs,
(A)
is
alH G ^ and
the following maximizes p (A)
i
7^
21
i\ 9'
{D)
=
9'
(A)
=
0.
= p (D):
for
i'
G
S,
The proof
Proof.
dition (59) in the
i
Appendix holds
(61-"'
{N),e-' (D))
(D))
=
that an equihbrium with
9''
AV'
and
(e-'' {N),e-''
^
i
note
exists,
i'
AV'
have
(0"'
(ct>
any such
(9^'
9''
+
(N)
i'
for
alH G
for
ri/cf)
—
(A)
9''
and
£"
i
and
z'
j^
9''
(A)
+
i
(A)) )=</./ (0
e
-
r?)
fARL
=
[-^
(D))
[9^'
-
7?)
<
i'
achieves the highest
and
9'
p (A)
{D)
=
=
(A)
9'
= p (D) among
Intuitively, the equilibrium that
de facto power means that
{D)
=
9''
z
£
+
=
(N)
9'
for all
(N)
if
is
given by
ri/(p.
To
for all
and
^
i
2
see
6 f
we
i',
Second, from (59) and (60) in the
^-'^
+ PAV^'
>0 =
for
all
make
ah
has a
i
,
r'''
(D))
V' {d'' (A) 9'' (D)), implies
,
,
6*- (D))
zero contributions. Hence,
e £ and
i
^
i'
is
9^
(D)
—
an equilibrium and
MPEs.
makes only one
this agent
(r'' (A)
(^^ + /3AF' {9-^ (A)
£, establishing that they prefer to
1]/^ for
^
{D)
=
i'.
,
for all other
9'
(D)) can take
as
G £,
(A)) )=#(</> [9' (D))
[9^'
9''
to maxi-
Appendix hold
(60) in the
(6*^* (A'') ,9~'^
which, in view of the fact that AV'' (9^'' (A) 9"'' (D))
</>/ (<^
and
same
so does (60) for the
£,
p{N) = p{D), we need
(59)
AV^
£
i
together with
i'
that, since 0' (D)
=
{N),d-' (D))
for
=
(D)
whom
for
some
for
ri/cf)
first
Appendix, we have that
cPf
e £
i
Clearly, the highest value
equahties.
some
as equality for
aW
for
any equihbrium where con-
clear that in
it
This implies that to maximize
e £, and vice versa.
mize AV"'
makes
of Proposition 2
elite
lot to lose
agent
make
all
of the investment in
from democracy (because of the higher
investment in de facto power involved in this regime), and maximizes investments in de facto
power.
Subgcime Perfect Equilibria
3.3
The
analysis so far has focused
they
may be
on
MPE.
able to achieve a better equilibrium than the
against each other. In this subsection,
the above game.
also take the
Since the landed elite form a small cohesive group,
The main
same form
result
as the
is
we
briefly discuss
MPE by using threats of punishments
SPEs (subgame
perfect equilibria) of
that for sufficiently large discount factors, the "best"
MPE characterized in Propositions
1
and
2,
SPEs
with the equilibrium
probability distribution over economic institutions independent of the political regime.
In characterizing the SPEs, we allow
and coordination, except that
labor
demand
in
elite
agents to use any kind of punishment strategies
competitive labor markets, they cannot (by law) restrict their
in order to affect factor prices.
To
22
define an
SPE,
let 9 {st,t)
=
{^-^
(s(,i)}
.^
be the vector of contributions by
{9{so,0) ,no,TQ,si,
An SPE
Definition 3
Z+x {N,D}
R+
up
Ti.*^^
when
t,
the state
be the history of contributions,
denoting the set of possible histories at time
H*'
and economic and
are best responses to {9^}ic£>
in
This
^^^ factor prices are given by
in which, at
agent being
if it is
is
time
made worse
f
=
0,
off.^'^
Pareto optimal and
no
elite
agent can be
In addition,
all elite
We
we
SPEs
made
an
£,
9'^
:
state st
t,
for
s'
-^ {A^,
D}
:
specify-
alH 6 £ and r and
s'
— time
to see
—
=
t
how
perfect equilibria in
utility of
the
better off without
SPE
elite.
the ability of the
elite
SPEs
define best or "Pareto optimal"
refer to
agents use the
&
has political power,
many subgame
exist
are motivated to look at the
to coordinate their actions changes the results.
SPEs
i
(8)-(ll) as a function of r £ {0, 1}.
on SPEs that maximize the ex ante
is
we
natural since
is
a best response to 0~*, r and
most repeated and dynamic games, there
model. Our focus
this
0' is
who
=
outcomes
political decision functions r
ing economic and political institution decisions as a function of time,
the state and history, such that
/i*
t.
specifying their contribution as a function of time
to that point
Let
S(.
political
Z+x{0,l}x{7V,L>}x W*-i ^{0,1} ands':Z+x{0,l}x{Af,D}x W*-i
As
is
consists of contribution functions for each elite agent
-^
X H^"^
and the history
with
t,
agents at time
,TTt,Tt,st+i)
...,9 {st,t)
and actions up to time
elite
some other
as
elite
as "symmetric Pareto optimal",
same equilibrium
strategy.
The main
result
the following:
Proposition 3 (Subgame Perfect Equilibrium and Invariance) Suppose Assumptions
Then there
1-3 hold.
SPE
exists
^€
[0,
1)
such that that for
all /3
>
^, the
symmetric Pareto optimal
induces equilibrium probabilities of labor repressive institutions
Moreover, as
/3
—
>
1,
any Pareto optimal
SPE
p (D)
involves
= p (N)
G
p{D) = p{N) €
(0,1).
(0, 1).
Proof. See Appendix.
This proposition therefore shows that as long as the discount factor
"best"
SPEs
also give the
discount factor, the
the
MPE
same invariance
elite act totally cohesively, as
MPEs.
Intuitively,
large enough, the
with a high enough
a single agent, and the same calculus as in
applies for equating the marginal cost of greater contributions to de facto
the marginal benefits, again implying
small, however, this result
may be
result as the
is
may no
p{D)
—
p{N).
When
the discount factor
is
to
sufficiently
longer be true, because a different pattern of contributions
necessary to ensure "incentive compatibility" on the side of the
Clearly, here Pareto optimality
is
power
only
among
elite
agents
(i.e.,
the elite and does not consider the utility of the citizens.
23
to
ensure that certain
make along
3.4
elite
make
agents are willing to
the contributions they are supposed to
the equilibrium path).
Comparative Statics
We now
return to the symmetric
Comparative
MPE
and derive a number of comparative
static results.
statics are straightforward in this case, since equations (16), (19)
and
(23), im-
mediately imply that
V{N)-V{D)^^>Q,
(31)
where we have dropped the conditioning of the value functions on the equilibrium 9 (D) and
6
(N) to simplify the notation. Equation
(31)
is
intuitive. In the equilibrium of Proposition 1,
the only difference between democracy and nondemocracy for the
elite is
that in democracy
they have to spend more in contributions in order to retain the same de facto pohtical power.
In particular, the per
elite
additional spending
is
equal to r]/(j)M, which
is
increasing in the de
jure political power advantage that democracy creates for the citizens (since, in equilibrium,
the elite totally offset this advantage).
Using
(21)
and
(31)
and denoting the equilibrium
*/WM«.(iV))(^
+
level of 6
(N) by
9* {N),
we
have:
|^)=l.
Similarly, denoting the equilibrium level of 9 {D) by 9* (Z)),
we
(32)
also have
^/(^Mr(Z))-,)(^ + ^)=l.
Finally, let us denote the probability that the elite will have political
p{N), and
recall that this probability
control political power,
economic institutions
and
will
also the probability that the society will
An
we have the
= p (D) =
elite will
be nondemocratic and
be labor repressive rather than competitive. Thus
this probability
of the model.
Proposition 4 (Comparative Statics) Suppose that Assumptions
1.
power by p*
corresponds both to the probabihty that the
summarizes most of the economic implications
baseline model,
(33)
1-3 hold.
Then
in the
following comparative static results:
increase in the economic rents that the elite can obtain by controlling political power
will increase their contributions
and the probability that they control
i.e.,
09* (N)
„
89* {D)
24
„
dp*
political
power,
2.
An
increase in the discount factor will increase the elite's contributions and the proba-
bility that
they control political power,
de* {N)
3.
An
them, and
power
4-
An
number
increase in the
will
„
i.e.,
de* {D)
^
dp*
^
^
of the elite will intensify the collective action
reduce their contributions and the probability that they control political
i.e.,
increase in the advantage of the citizens in
democracy
will increase the elite's contri-
butions and the probability that they control political power,
de*(N)'
^^
>
„
89* (D)'
0,
J"
or]
5.
An
>
„
^
dp*
,
and
0,
>
i.e.,
^
0.
orj
or]
increase in the effectiveness of the de facto political power technology of the elite has
ambiguous
political
effects
power,
on their contributions, but increases the probability that they control
i.e.,
dp*
Proof. All of these comparative
Function Theorem
(e.g.,
„
from
static results follow
Simon and Blume, 1994, Theorem
Function Theorem, since /
is
differentiable everywhere
ensure that the equilibrium
is
always at an interior point.
some
problem among
of these results. For example, for 39* {N)
/dAR,
and
(32)
(33) using the Implicit
We
15.2).
can use the Implicit
and moreover. Assumptions
We
2
and 3
argument
briefly sketch the
use the Implicit Function
for
Theorem on
(32) to obtain
39* jN)
f{(PM9*{N))L
3AR ~ ~ f {<pM9*
since /'
<
fi^om the
establishes 39* {D)
is
everywhere
(TV))
^
M {(pARL +
/St?)
second order condition. Using the Implicit Function Theorem on (33)
/dAR >
0.
To obtain dp* /dAR >
0,
note that p*
= F {(j)M9* {N))
and
F
strictly increasing.
The comparative
statics in part 2
with respect to
/3
are identical.
Using the Implicit Function Theorem with respect to
d9*{N)/dM <
the effect on p* at
and 39*
first
{D)/dM <
M
as claimed in part
3.
also immediately estabhshes
Since p*
=
appears ambiguous. However, note from (32) that as
M increases,
second term on the left-hand side declines, so f {4>M9* {N)) has to increase.
25
F{<pM9*{N)),
Since /'
the
<
0,
this
only possible
is
=
4>M6* [N] declines, so p*
if
F{(f)M9* {N)) also declines (given the
monotonicity of F).
Next, the Implicit Function Theorem also gives the results in part
dO* (N)
/3/(#//r(7V))
^
similarly,
<
Since /'
The argument
0.
side of (32) increases as
this implies that
0,
>
establishing dp* j drj
Finally,
>
dO* {D) /drj
term on the left-hand
it is
M {(pARL +
(pf {(t>Me* {N))
dri
and
4, in particular,
for
prj)
>
dp* /drj
increases, so
77
4>M6* {N) increases, so p*
/
The second
again similar.
is
{(f)M9* (N)) has to decline.
= F {<pM9*
(N)) also increases,
0.
straightforward to verify that the effect of
(f>
on
6*
(N) and
9*
(D)
is
ambiguous.
However, writing (32) as
we
see that an increase in
has to decline. Since /'
must
increases the second term
(p
<
>
political power.
The
greater
is
is
economic
<
more
willing
is
each
elite
when A
Since Ai? will be high
and economic institutions are more
institutions are
more
relations rather than competitive labor markets.
fact that a higher
repressive institutions
is
/3
elite
is
also interesting.
In
likely to
likely to
be
by the
many models,
^^The
institutions.
also in line with the empirical literature
the benefits are higher
fact that the elite
(e.g.,
(e.g.,
Wade,
proto-industrial)
elite
can obtain
be controlled by the
tilted
elite,
towards repressive labor
^^
also increases contributions
and labor repressive economic
^""This finding
when
agent to contribute to
low. Proposition 4 also
is
elite
and the likelihood of labor
a higher discount factor leads to
better allocations. Here, in contrast, a higher discount factor leads to
by the
terms
they have to gain by controlling
developed and where, by repressing labor the
activities are less
and consequently economic
The
in
Ai? induces both greater
In other words, in a society where alternative
0.
large rents, political
how much
a measure of
this gain, the
their collective political power.""*
implies dp* /dA
(TV))
agents and also increases the probability that they control political
elite
very intuitive, since A/?,
is
= F {(j)M9*
and yet quite useful
of economic implications. For example, the fact that an increase in
power
/ {(pM9* {N))
0.
of the comparative statics in Proposition 4 are intuitive,
contributions by
side, so
implies that 4>M9* (N) increases, and p*
0, this
also increase, estabhshing dp* /d(j)
Many
on the left-hand
The reason
more wasteful
is
activities
that the main pivotal
on collective action which finds that
it is
more
likelj'
1988).
can never impose slavery, thus taking away the outside option of the workers,
naturally important for the result that dp* /dA
<
0.
26
is
agents in this model are the eUte, which, by virtue of their smaller nmnbers, take the effect of
on equilibrium allocations into account. Contributing to de facto
their contributions
power
a form of investment, and some of the returns accrue to the
is
they secure nondemocracy instead of democracy).
them
to invest
more
in their political
economic institutions more
The
problem
show that when
in the model.
since each elite agent realizes that
A
elites.
greater
problem (the
free-rider
M
/3
encourages
elite
This highlights the
worse.
agent contributes to the group's
suboptimal from the viewpoint of the group,
is still
by contributing more he creates a positive externality on
M increases the extent of
this positive externality
collective action problem). This
and
intensifies the
comparative static therefore suggests
that nondemocracy and labor repressive economic institutions are more likely to emerge
there
is
more
increases so that there are
among them becomes
Even though each
political power, their level of contribution
other
Therefore a higher level of
likely.
the collective action problem
latent free-rider
the future (when
elite in
power and makes nondemocracy and labor repressive
third set of comparative statics
elite agents,
political
when
a relatively small and cohesive group of elite land owners, a pattern consistent with
the historical case studies discussed below.
The most
higher
77
surprising comparative static results are those with respect to
corresponds to democracy giving more de jure power to the citizens.
therefore expected a greater
that higher
77
still
costly for the elite, so
political
to lead to better
77
outcomes
The reason
holds).
for this is that
in the interest of
it is
power to avoid democracy. This
each
effect is
elite
institutions
a higher
Corollary
1
is
non-monotonic:
applies
if
77
Remark
1.^^
M reduces the
77
(i.e.,
makes democracy
on the likelihood of
an absorbing
political
we
state).
power of the
believe that the baseline
both the simplest and the most natural one, and highlights important
The
uj
in
elite
model we have
first-order effects.
democracy and the fact that (13) is
on p* is ambiguous in general
fact that the effect of
M
and Ray (2001), who emphasize that the effect of an increase in the number of agents
contribution of a lobby is ambiguous because, while each agent contributes less, there are more of
related to Esteban
to the total
likely (as
power depend on the functional form assumptions
^^In particular, both the additive shift of the distribution function of
is
find
that Assumption 3 no longer holds, then
fully consolidated
Nevertheless,
linear are important for these comparative statics.
77
more
we
have
strong enough to increase the probabihty
much
results that higher
increases their political
already highlighted in
is
increases so
and democracy may become
Note however that both the
and that higher
77
We may
agent to invest more in the group's
that they will maintain political power. However, the overall impact of
democracy
Recall that a
for the citizens. In contrast,
makes nondemocracj^ and labor repressive economic
long as Assumption 3
more
77.
them.
27
when
Finally,
elite
power
increases, the technology of garnering de facto political
(f>
may
improves. This
for the
reduce their contributions to the group's de facto power, but
will
it
always increase the equilibrium probability of a nondemocratic regime and labor repressive
institutions.
4
Generalizations:
Markov Regime-Switching Models and
State
Dependence
The model
in the previous section yielded stark results,
assumptions that the
elite
had the same technology
which were partly driven by the
power
to generate de facto political
in
both
regimes and were able to change economic institutions immediately after they took control
addition to the functional form assumption noted in footnote 26)
leads to a richer form of persistence, in the form of a
These
state dependence.
way
of our analysis of
Relaxing these assumptions
Markov regime-switching model with
issues are discussed in the next
how
.
(in
two subsections, and also pave the
to effectively reform equilibrium institutions in Section 6.
Another special feature of the model
in the previous section
was that
it
implicitly
assumed
that changing economic institutions and changing the political system were equally easy (or
An
difficult).
extension in which changing pohtical institutions
power than influencing economic
Throughout the
Our
first
on the
activities of the elite, their
(l)^
£
(0, (p)
and Zt
elite for increasing their
This
is
require greater political
discussed in Section
we
5.
focus on symmetric
MPE.
of the Elite
generalization of the above framework assumes that in democracy, because of limits
technology for gathering de facto political power changes to
.
where
is
rest of the paper, to simplify the discussion,
De Facto Power
Limits on
4.1
policies or institutions
may
=
Yli^E
^t-'^'^
Pf
^^^
de facto power
{D)
=
(34)
<j>^Zt,
other words, each unit of the final good spent by the
is
less effective in
democracy than
a reasonable assumption, since democratic institutions
may
in
prevent the
nondemocracy.
elite
from using
repression or paramilitaries or from buying politicians as effectively as in a nondemocratic
regime.
Therefore, in this model democracy has two simultaneous functions;
distribution of de jure political powers towards the citizens
facto power by the
^'^We
political
now
differs
it
shifts the
limits the exercise of
de
elite.
Pf (D) as opposed to Pf^ as
between the two pohtical regimes.
write this as
power
and
it
28
in (13), since
the technology of generating de facto
We now
have the probability of the ehte controlhng the pohtical agenda in democracy as
p{0\0{D)\D) = F{(Po{{M-l)9iD) + d')-v),
and the value function
Definition
unchanged and
is
interior solutions, then the
{D)
<Pj,f {cf>j,Me
We
democracy
applies to this modified model,
1 still
Assuming
in
-
v)
and
symmetric
[^ +pV[N\e{N),e
is still
specifies
MPE
{D))
is
(35)
given by (19).
characterized by (21) and
-^V{D\e{N),9 {D))^ =
is
(36)
l.
interior:
3'
mm |<^/ (0) -j^,<pDf i-v) -j^j
Now
that
symmetric MPEs.
can impose a variant of Assumption 3 to ensure that the equilibrium
Assumption
It is clear
piN)=p [6 (N) ,e{N)\ N)
recalling that
and
>
1-
p{D)=p [9 (N) ,9{N)\D),
compari-
son of (21) and (36) immediately implies that
p{N)>p{D).
To
more
see this
in the
explicitly,
F {(f>Md {N)).
(j)j;)M9
{D)
Note that p (N)
is
-
r]
<
is
—
implies / {cpj^AdO {D)
(pj^
decreasing and
<j)M9
{N) and p [D)
rj)
> f {4)M9
[N)). Since
F {) is strictly increasing every= F {cPdMO (D) -r]) < p{N) =
the probability of nondemocracy persisting, while
1
— p {D)
the probability of democracy persisting. This implies that labor repressive institutions are
less likely to arise in
cratic,
to
>
(p
neighborhood of equilibrium, /(•)
where, we must have
is
note that
(37)
democracy than
in
nondemocracy. Moreover, once the society
democracy from nondemocracy. Consequently,
and economic
Assumption
power
demo-
has a higher probability of remaining democratic than the probability of switching
it
political
is
to
3'
impose
in this
model there
is
persistence of both
institutions.
also implies
p (D) >
their favorite
0,^^ so
economic
even in democracy, the
institutions,
elite
and change the
have the potential
political
we have p{N) €
system back
to nondemocracy,
and moreover from Assumptions
Markov process
ergodic (irreducible and aperiodic). Next, dividing (21) by (36) yields:
is
2
and
3',
0o/ {4>dM9 (D) -v)^(Pf i^M9
"In
fact,
p{D) >
would not be
{N))
,
(0,1), so the
(38)
would follow from the weaker assumption that tj < —to, though this latter assumption
by the elite in democracy, which we use for comparative
sufficient to ensure positive contributions
statics.
29
which shows that the gap between
and
(f)
determine the gap between p {D) and p
will
cj)^
thus the extent of persistence of economic and political institutions
p[N)). This leads to the following
ified
2
model with
and
3' hold.
limits
on the
as (p^ —^
(f>,
p [D)
—
result:
De Facto Power and
Proposition 5 (Limits on
(e.g.,
(TV),
State Dependence) Consider the mod-
de facto power in democracy. Suppose that Assumptions
elite's
Then any symmetric
MPE leads to
a
Markov regime switching
1,
structure where
the society fluctuates between democracy with associated competitive economic institutions
[t
—
I)
and nondemocracy with associated labor repressive economic institutions
switching probabilities p (TV) e
The proof
and
(0, 1)
1
- p {D) E
where p (D)
(0, 1)
(r
=
with
0),
< p (TV).
1,
and
follows straightforwardly from the expressions in the text, in particular, equations, (21)
and
(36),
of this proposition
and Assumptions of
The most important
1,
2
and
omitted since
is
is
similar to that of Proposition
3'.
implication of this modified model
—which we
type of institutional persistence
of the system, s e {TV, D).
refer to as state
that there
is
While Proposition
1
irrespective of political institutions,
as likely to follow a democratic regime as
The
likely to
remain nondemocratic than
different
featured invariance in the sense that economic
same equilibrium process
regime.
now a
dependence, since the probability
—
not lead to persistence in political institutions; the fact that p {D)
democracy was
is
and economic institutions depends on the current state
distribution over equilibrium political
institutions followed the
it
results in Proposition 5 are different; once in
it is
to switch to
it
p
(TV)
it
did
imphed that
was to follow a non-democratic
nondemocracy, the society
is
more
nondemocracy from democracy. This
is
the essence of state dependence.
Also interesting
is
problem and gather
favorite
the fact that the
sufficient
economic institutions,
elite still
we need
to solve for
ability to solve their collective action
de facto power to dominate democratic politics and impose their
i.e.,
p (D) >
to also change the political sj'stem from
It is also possible to
have the
(though here this also corresponds to their ability
democracy to nondemocracy).
obtain additional comparative static results for this case. To do
V {N) = V {N
\
9
(N)
,9 (D))
and
V {D) = V {D
straightforward by imposing that equilibrium probabilities of the
the citizens are
p{D) and p{N), and combining
30
(16)
and
(19),
elite
\
9
{N)
this,
,9 (D)). This
is
having more power than
which gives
Combining
this equation
Assumption
with (21) and
(36),
we obtain
(again for an interior equilibrium since
imposed):
3' is
and
f^o-^ fU
Mfffn^
e{D)-eiN) + {piN)-p{D))ARL/M \
^AflL
.
(41)
p{D) =
where, clearly,
define
H^
We
is
and
F {cp^Me {D) -
rj)
= F{(pM9{N)).
and p{N)
These equations
also
iif^ for future reference.
The
can nov/ perform the comparative statics using these two conditions.
that the equilibrium
and 9 [D) might
is
no longer guaranteed to be unique, since multiple values of 6 {N)
satisfy these
situation in which there
difficulty
is
two conditions. Moreover, an equilibrium may correspond to a
"myopic instability"
in the sense that a small increase in
one of the
equilibrium variables, say 9 {N), will lead to a further increase in that variable. In these types
of situations, comparative static results are difficult to obtain. ^^
To make more
we
progress,
appeal to Samuelson's (1947) correspondence principle, which essentially restricts attention to
equilibria that satisfy a type of "myopic stability".
Under
this assumption,
it
is
possible to
obtain some comparative static results.
More
specifically,
with respect to
{9
myopic
stability
i.e.,
an increase
_
de{N)
dO(D)
\
\ ae{N)
dB{D)
/
I
in 6
[N) reduces
cross-partial terms are not large
dH^/d9{D).
>
dH^ /d9 {D) <
definiteness of
0,
for
an increase
if
in 9
only require the condition
9
{D) reduces
enough to dominate the product
and exactly the same arguments
satisfied.
dH^/d9{D) <
J
and
(i.e.,
may
the negative
0?°
this condition
not be negative definite
is
that
well known, when there are multiple equilibria, the comparative static results will typically
some intermediate equilibria; see, for example, Milgrom and Roberts (1994) for supermodular
is
games,
30
The
x
as in that proof also establish that
dH^ 189 (N) <
The reason why
H^
dH^ /d9 {N)
(N) and 9 [D) were solutions to an optimization problem,
would be automatically
"^In fact, as
,
so myopic stability and the correspondence principle
0,
J)
Notice that
be reversed
H^
In fact, the proof of Proposition 6 will show that
dH'-' 139 {N)
,
,
negative definite,
and the
[H^ H^)
{N) 9 {D)),
-r
is
would necessitate that the Jacobian matrix of
condition dH'^ /dO (N)
<
can always be ensured by making f {(j)M6 {N)) sufficiently decreasing.
31
the system of equations (40) and (41) arise from the equihbrium interaction of ehte agents
individually contributing to increasing the de facto political power of their group.
By
applying the correspondence principle, we can establish most of the comparative static
These are stated and proved
results of interest.
in the next proposition (again using *'s to
denote equilibrium values).
Proposition 6 (Comparative Statics for the Model with Limits on
Consider the modified model with limits on the
that Assumptions
matrix
J
is
1,
and
2
3'
elite's
De
Facto Power)
de facto power in democracy. Suppose
hold and following the correspondence principle, assume that the
negative definite. Then,
we have the
following comparative static results:
1.
89* (N)
09* (D)
„
dAR
dp*{N)
„
dAR
'
„
dp* (D)
,
dAR
dAR
Op* jN)
Op* jD)
'
„
2.
90* jN)
09* jD)
-or'^^''~w'^^~w~'''^^'''^^r''^09* jD)
09* jN)
We
Proof.
Let us
first
provide the proof for the
fact that
addition,
consider
it
first result.
The other two
are proved analogously.
apply the Imphcit Function Theorem and write
dSlN)
dd(D)
dH°
ohd
\ de(N)
The
Op*{D)
Op*iN)
J
d0(D)
\
{
I
\
99 (N)
^
/
09{D)
can be verified that QH'^ /Q9 (D)
OH^ /09 [D).
OH^ _
_ _l
I
\
\
m
9KS
dAR
<
and
i3(pf {(t)M9
jN))
(35)
OAR.
/
OH^/89
Differentiating (40) with respect to 9 [D)
and
0H^/09{D) <
{N)
>
0.
and using
To
0.
In
see this, first
(39),
we have-
-li{^-^^(-'-K
09{D) ~ l-(3ip{N)-piD))
and moreover, from
\
J
OH^ /09{N) <
negative definite imphes
is
I
^
we have
Op{D)
09 {D)
The
fact that
Now
M>
1
^^^Df{4>DM9{D)-r,).
combined with the
last
two equations and
(36) implies
OH^ /09 [D) <
with the same reasoning,
fA^
09{N)\ M
QH^ _/?<^^/(0^Me(Z?)-,)[
09 {N)
5p(A0
l-/3{piN)-p{D))
32
^^
^
^
^
'^
0.
and
(15) implies:
= M^/(^M^(iV)).
•
|||f|
Combining
this
with the appropriate first-order condition,
(17), gives
dH^ /d9 {N) >
0.
Therefore, we have the following sign pattern for J:
Moreover,
straightforward to verify that an increase in
it is
hand
sides of (40)
(e.g.,
Simon and Bloom,
and
(41),
dH'^/dAR >
i.e.,
1994,
Theorem
9.4),
AR strictly increases both the left
dH^/dAR >
and
0.
Then by Cramer's
rule
we have
dAR
det
J
Totally differentiating (38) implies that
^
(since /'
{(p^Me [D) -
and d9 {N)
/dAR >
??)
<
and
0. Finally,
/'
dAR
^
J
{4>Me (N))
dp* {D)
<
/dAR >
V '5A/?
0),
which establishes that 09 (D)
/dAR >
immediately follows from the monotonicity
ofF(.).
The same reasoning
establishes the comparative statics with respect to
/3
and
M.
This proposition therefore shows that most of the qualitative results from the basehne
model generalize
to the
Markov regime-switching model with
nondemocracy and labor repressive economic
institutions are
have greater rents from repressing labor
when
when they
are
more forward-looking
(smaller) group
(i.e.,
when
M
example, those concerning the
4.2
is
(i.e.,
low).
effect of
when
(3 is
AR
is
high),
more
high, for
likely
when
the landed
example because
and when they form a
A
n:iore
is
and
t],
are
low),
cohesive
However, some other results from Proposition
</>
elite
4, for
now ambiguous.
Sluggish Economic Institutions
Next we modify the above framework
democracy, the
ple,
(i.e.,
state dependence. In particular,
elite
in a different direction,
and assume that starting
cannot impose their favorite economic institutions immediately,
for
in
exam-
democratic politics has already taken some actions that cannot be reversed within the same
period. This implies that starting in democracy, economic institutions are "slow-changing" or
sluggish. This structure
cost
is
also formally equivalent to
when they change economic
institutions
one in which the
elite
incur a temporary
from competitive to labor repressive.
33
More
specifically,
we now allow
three different types of economic institutions:
=
corresponding to competitive markets, Tt
corresponding to
0,
=
Tt
1,
labor repression and
full
n=
1/2 corresponding to partial labor repression, in which case, wages are reduced to some
level
A < w < w [Tt =
=
l)
— Lf
f (L)
and thus returns to land owners with
(L),
partial labor
repression are equal to
RP
= R{t^
= Li^^hL^.
1/2)
(42)
Let us define A such that
RP-R'
^
with
AR as
The only
defined in (12).
difference
The
fact that
from the analysis
=
A<w<w
in Section 2
gain political power, they cannot impose r
(while starting in s
the analysis
is
=
A'',
^^R~'
=
any r G {0,1/2,1}
=
ensures A e
1)
that starting in s
=
(0, 1].
D, even
and the best they can do
0,
is
is
{t
allowed).
Given
this
is
if
the
to set r
elite
=
1/2
assumption, the rest of
similar to before, with the only difference taking place in the value function in
democracy, which now takes the form:
ViD\eiN),e(D))
= maxl-9'+p(9\9{D)\D)(^+pV{N\eiN),9{D))]
+ {l-p{9\9{D)\D))(^^+pV{D\9{N),e{D))^y
Once
(43)
again, focusing on interior solutions, this maximization problem implies the first-order
condition
cPf{cp{{M-l)9iN) + 9^)-n)(^^^^ + l3{V(N\0{N),9{D))-V{D\e{N),eiD)))^^l,
(44)
which only
differs
from (20) because the gain of capturing power
/' (0
((M — 1)9 (AQ
the policy correspondence be denoted by
T^
{N)
The
unchanged and
The corresponding second-order
let
is
condition
value function in nondemocracy
order condition for contributions
is
is
is
{9
,
now
-|-
0^)
XAR rather
—
t?)
<
0.
than AR.
Once
again,
9 [D)).
is
given by (16), and the
first-
given by (17), with the policy correspondence given by
T^{9{N),9{D)).
To
define an equilibrium formally, let us also recall that
having political power.
set
in
r
=
if
they want
Now
to,
let
and
tt
let
=
n
stand for the
—
elite
tt
having
—
full
1
stands for the citizens
power, so that they can
1/2 denote the ehte capturing political power starting
democracy. Thus we have:
34
Definition 4
A symmetric AdPE of the model
with sluggish economic institutions consists of a
and
pair of contribution levels for elite agents 9 {N)
and
9 {D)
T^
S
such that T (^
and
s'
{n
Given
same
=
I)
[9
{N) 9 {D)]. In addition, economic and
0)
—
D, and
=
=
Af,
factor prices
ai'e
0, s' (tt
this definition of a
i.e.,
=
-
{<i^M9 {D)
Comparison
0)
symmetric
-
r (^
1/2)
=
1/2,
s'
given by (8)-(ll) and
MPE,
(^
=
iu
and
1/2)
=
(42)
T^
[9
(tt)
and
A^,
{N) 9 (D)]
,
r (^
=
when r =
the equilibrium condition for 9 (N)
is
are
s' (tt)
1)
=
1
1/2.
again the
equation (22), while with the same steps as in the previous section, the
equilibrium condition for 9 {D)
<t>f
political decisions r
,
=
as before,
(D), such that 9 (N) G
r?)
is
given by:
(^^ + f3{ViN\9{N),9
{D))
-ViD\9iN),9 {D)))^ =
of this condition to (22) immediately establishes that as long as A
long as democracy does put restrictions on economic institutions that the
elite
<
(45)
1.
i.e.,
1,
as
can impose, we
havep(Z)) <p{N).
As
before,
we impose an assumption to ensure an
interior equilibrium:
Assumption 3"
mm|<^/(O)^^,0/(-r?)-^^|>l.
Proposition 7 (Sluggish Economic Institutions and State Dependence) Consider the
modified model with sluggish economic institutions. Suppose that Assumptions
Then any symmetric MPE
hold.
fluctuates between
and
1
- p (D) G
Proof.
establishes
(0, 1)
where p{D)
3"
society
democracy and nondemocracy, with switching probabihties p {N) G
(0, 1)
<p (N).
which
(45),
p (D) < p [N). m
V {N)
full
and
solution to this model,
V [D)
we again need
to solve (16) together with (43).
to simplify notation, this implies
^^'^^~^^''^ =
Now
and
2
Markov regime switching structure where the
This result follows immediately from the comparison of (22) with
To obtain a
Using
leads to a
1,
9{D)-9{N)+p jN) ARL/M - p jD) XARL/M
.
(^^)
•
l-^ip{N)-piD))
combining this with the equilibrium conditions, (22) with
(45),
we have conditions
char-
acterizing the equilibrium similar to those in the previous subsection:
V
^^^
i- P[P{N) -p[D))
J
(47)
35
/^
nD-^fUAmtm
^
AAEL
e{D) - e{N) +piN)
AR/M -p{D) XAR/M \
(48)
To make more
progress,
we once again impose the correspondence
matrix J" as before and assmne that
it is
principle,
and define the
negative definite.
Proposition 8 (Comparative Statics for the Model with Sluggish Economic Institutions) Consider the modified model with sluggish economic institutions.
Assumptions
1,
2
3" hold
and
have the following comparative
and assume that the matrix
J
is
Suppose that
negative definite. Then,
we
static results:
1.
do* JN)
89* (D)
dp* jN)
dp* jD)
99* jN)
80* jD)
dp* {N)
dp* jD)
90* (N)
89* jD)
8p* (N)
8p* jD)
3.
^
4-
90* (N)
^\ '
dx
Proof. The proofs of the
are omitted.
To obtain the
first
(
1
mm
dHR_
\ d0{N)
An argument
„
dp*iN)
-~^
<
dx
,
and
„
0.
three results are identical to those in Proposition 5 and
last one,
totally differentiate equations (47)
<
we can again apply the Implicit Function Theorem and
and
(48) to write
Wd) \ ( 99 (iV) \
mR_ \\d9[D))~
y
\
ae{D)
J
^
'
/
\
^
m^
\
d\
\
\^^/
similar to that in the proof of Proposition 6 estabhshes that
dH^/d9 (N) >
0.
Next
Then applying Cramer's
it is
rule
also straightforward to see that
and using the
fact that det
dX
det
J>
dH^/dX >
dH^/dX <
J
F ().
36
and
and
j'ields
while the same steps imply that d9* {D) /dX has ambiguous sign.
follows immediately from the monotonicity of
dH^ /d9 {D) <
Finally, dp*
{N) /dX <
0.
This proposition shows that in the environment with sluggish economic institutions the
main comparative
static results
from Proposition 4 regarding the
effect of
economic
rents, the
Also interestingly, a
discount factor and the cohesion of the landed ehte continue to apply.
decrease in A, meaning more sluggish economic institutions in democracy, increases p* (N).
This
is
intuitive; a lower
take time for
them
to
A means that democracy
more
costly for the elite, because
it
will
impose their favorite economic institutions even when they take control
in democracy. Consequently, the value of
is
is
nondemocracy
higher, so in nondemocracy, the elite invest
more
institutions (and the associated labor repressive
a decrease in A on the
elite's
relative to
in order to preserve this set of political
economic
de facto power in democracy
immediate benefits of taking control
in
V {N) — V (D),
democracy,
democracy, but
The
institutions).
is
also,
implications of
ambiguous, since
by increasing
it
reduces the
V {N) — V (D),
it
increases the long-run benefits.
Durable
5
The assumption
Political Institutions
so far has been that
when the
and Captured Democracy
elite
have more political power than the
citizens,
they can change both economic institutions and the political system (though in the previous
subsection, they could only change the economic institutions slowly).
below
tured form,
very
least,
whereby the
It is
As discussed
institutions are
political
in detail in
more
historical
examples
endure, but in a cap-
are able to impose their favorite economic institutions (or at the
effect
on the choice of economic institutions)
and economic
an enduring
differences
institutions.
Acemoglu and Robinson
difficult to
in
model so that there can be
therefore important to generalize the
between equilibrium
ify
elite
have a disproportionate
democracy.
may emerge and
a different salient pattern: democracy
illustrate
The
change, and
may have
(2006a), in
many
situations, political
additional "durability"
.
We now mod-
the baseline model to incorporate this feature and assume that overthrowing a democratic
regime
is
more
difficult
than influencing economic institutions.
More
require greater political power to force a switch from democracy to
ply influencing economic institutions in democracy.
that
when they
economic
To
specifically, the elite
nondemocracy than sim-
simplify the discussion,
we assume
influence economic institutions in democracy, they can choose their favorite
institutions, labor repression.
Finally, for reasons that will
become apparent below, we now assume that the time
37
t
preferences of citizens,
i.e.,
those for
all
&
i
by
C, are given
oo
^/3^(cj+, + K5t+,)),
with
i^
{S
= N) =
and u {S
= D) >
the citizens from democracy (which
form
y {S
for other benefits
= D)
is
large
We
model
Pf, where
system.
If,
i^
may
Therefore, these preferences allow a direct utility for
0,
and vote
citizens always prefer
democracy even when
for
this
costs for them.
institutional change as follows.
>
be because of ideological reasons or a reduced-
in turn
provided to the citizens by democracy). Moreover, we will assume that
enough that
may have economic
0.
(49)
We assume
that
when
s
=D
and
Pf + > Pf >
£,
the elite can choose economic institutions but cannot change the political
on the other hand,
Pf > Pf +^,
the future political system. Symmetrically
the
elite
when
s
can choose both economic institutions and
—
N
Pf + ^ > Pf >
and
Pf, the
citizens
can choose economic institutions, but cannot change the political system. This formulation
builds in the assumption that changing political institutions
economic institutions
Throughout
in the
this section,
rate in this subsection,
is
we again
focus on symmetric
we assume that when the
Finally, to further simplify the discussion
2'
elite
MPE.
have more political power
F is defined
over
(cj,
oo) for
some w <
everywhere), and moreover
its
more complicated, but have
=
0,
we have
0, is
in
democracy,
as in the baseline model.
2:
everywhere
strictly increasing
and
density / and the derivative of the density,
/' (cj)
Given these assumptions, the structure of the model
are
than influencing
Also, to keep the issues sepa-
we strengthen Assumption
twice continuously differentiable (so that
/', exist
difficult
most straightforward way.
they can impose their most preferred economic institutions, r
Assumption
more
is
<
for all ui
and
3.
/
The value
similar to before.
similar intuition to those in Section
limtj_>oo
(t^)
—
0.
functions
In particular, in addition
to (18), let
p{e\9iD)
\
D) ^
F {ct>{{M -l)9iD) + 0')
38
-71
-()
,
(50)
so that
we
have:"*^
V{D\e[N),e[D))
= maxl-e'+p{9\e{D)\D)^+
{l-p{9\9{D)\D))^+p{9\9{D)\D)(3ViN\8{N),9{D))
+ {l-p{e\9{D)\D))l3ViD\9{N),9{D))},
where we have already imposed that when the
(51)
have sufScient power they
citizens
will
choose
democracy.
With
similar arguments to before, the maximization in (51) implies the following first-order
condition
A
or
cPf{<j>{M-l)9{D)+9^-v)^^
+P4>f {4>{M
which
is
now
-
1)9 {D)
sufficient since
The main
7]
-
Assumption
^)
2'
{V {N
\
elite
gain the economic rent
which they secure a change
(52).
As
9
{N) ,9 (D)) -
V {D
from the one before
ARL/M
is
is
which leads to a modification of the value function
= max
\
satisfied.
that the probability
T^
[9
(N) ,9 {D)).
define
p {9\9 [N) \N)^F{cP ((M -
V{N\9{N),9{D))
1,
from the probability with
different
before, denote the resulting policy correspondence as
we
is
=
For this reason, two different densities
in the political system.
Similarly for nondemocracy,
(N) ,9 (D)))
9
\
ensures that the second-order condition
difference of this first-order condition
with which the
appear in
+ 9' -
(52)
-9'
1) 9
for
{N)
+
+ C)
9')
nondemocracy
(53)
,
as
+p(9\9{N)\N)^+
{l-p{9\9{N)\N))^+p{9\9{N)\N)f3V{N\9{N)^9{D))
+
which
(1
-p
{9\ 9 {N)
also has a similar structure to the value function in
sequently, the first-order (necessary
''^An alternative
the
let
\N))pV{D\9 (TV)
elite
way
of writing (51)
and
sufficient given
would be as
follows:
,
9{D))}
democracy
Assumption
define p(d^ ,6 {D)
\
(54)
,
in this case, (54).
2')
D)
Con-
condition for optimal
to be the probability that
are able to impose their preferred economic institutions but not change political institutions, and
p[6',d{D)
I
D) be the
p{9\6(D)
probability that they are able to change the political institutions as well.
Then
D), they only receive ARL/A4, whereas with probability p(9\9{D) D), they
receive ARL/M +
{V (N) — V {D)). This way of writing the recursive formulation is equivalent to (51) with
v{e\B{D) D) =f,{B\e{D) D) ^ndp{e',e{D) D) =p{e',e(D) d) -p{e\e{D) d).
with probability
I
\
\
I
\
I
39
\
contribution by an
agent
elite
is
also similar:
or
A
^f[^^M-i)e{N) + e')^^
+I3<t>f
(M -1)0 {N) +
(</-
(55)
+
9'
[V [N \6 {N)
which again defines the policy correspondence
To
now introduce
define an equilibrium,
T^
{0
,9 {D))
{N)
-V [D
,
s
=
or
A''
Pf > Pf + ^ when
the additional notation such that
=
s
=
D); n
tt
=
^
> pf >
=
TT
pf)',
gain any power in democracy
and
finally,
n
—
means the
(1,1)
power
elite loses
Pf > Pf +
(i.e.,
when
^
s
=
in
(i.e.,
Pf >
(i.e.,
nondemocracy
TV or
nondemocracy or
Pf > Pf when
power
(1,0) corresponds to the citizens maintaining de jure
losing control over economic institutions
denotes
(0, 0)
(0,1) corresponding to the elite keeping
control of de jure power but losing control of economic institutions in
pf +
1
9 {D)).
,
the elite keeping total power in nondemocracy or gaining total power in democracy
Pf when
=
\e {N) 9 (D)))
Pf +^ > Pf >
in
(i.e.,
fails
—
s
to
D);
democracy but
Pf)- Imposing that
citizens
always prefer democracy to nondemocracy (from preferences in (49)), we have:
A
Definition 5
symmetric
MPE of the
model with durable
pair of contribution levels for elite agents 9 (N)
and 9 [D) e T^
such that r
r
(tt
=
(tt
(0, 1))
[9
=
=
and 9 (D), such that 9 {N) G
{N) 9 {D)]. In addition, economic and
,
(0, 0))
=
1, s' (tt
=
r
(tt
=
(1, 1))
=
(1, 0))
=
s' (tt
0, s' (tt
=
political institutions consists of a
(1,0))
=
political decisions r
=
(0, 0))
= D, and
s'
(^
=
(0, 1))
=
T^
[9
(tt)
and
TV,
r (^
factor prices are given
(N)
9 (D)]
,
are
s' (tt)
=
by
(1, 1))
=
(8)-(ll)
as a function of r G {0, 1}.
Given
this definition, a
MPE with
symmetric
9
{N)
>
and 9 {D)
>
is
a solution to the
following two equations
<Pf [ct^MO
{D)
-
v)
ict>M9 (TV))
{4>M9
-^ + m {4>M9
A
</./
A RT
-^+W
r>
{D)-r,-
{V{N\9 {N)
,
9 {D))
-V{D\9{N),9 (D))) =
(56)
7"
(TV)
+ ^)
{V {N
\
9 (TV)
,
9 (D))
-V{D\9{N),9 {D))) ^
1.
(57)
It
be
can be verified that Assumption 3 above
is
sufficient to
ensure that zero contributions cannot
equilibria.
The
interesting result in this case
remain so potentially
for
a long time
control the economic institutions
is
(i.e.,
(i.e.,
that once the society becomes democratic,
p {D) can be
p{D) could be
in the next proposition.
40
small), but the elite
quite large). This
is
wiU
still
it
will
be able to
stated and proved
1.
Proposition 9 (Captured Democracy) Consider the modified model with durable pohtical
institutions.
Suppose that Assumptions
1,
and
2'
switching process with state dependence and
is
economic institutions
in
> p (A^ > p (D) >
1
< p{N) < p{D) <
captured in the sense that
hue with the
Then we have a Markov regime-
3 hold.
i.e.,
1,
elite's interests
Moreover, democracy
0.
democracy
will survive
but choose
with even a higher probability than does
nondemocracy.
Proof. The probabihty of labor repressive economic institutions under democracy
is
p{D)=p{e{D),9iD)\D) = F{<j>M9{D)~rj),
while
it is
p{N)^p {9 [N) ,9{N)\N) = F {(j)Me (N))
in
nondemocracy. Suppose, to obtain a contradiction, that p{D)
< p{N).
This
is
equivalent
to
{D)-T]< 4>Me [N)
(t>M9
Since from Assumption
2'
/
(58)
decreasing everywhere, this implies
is
f{ct>M9{D)-ri)>}{(j)M9{N)).
This equation combined with (56) and (57) implies that
/ {4>M9 (D)
Since from Assumption
2'
/
is
-
7,
-
<
decreasing, this
is
0M0 {D)-r]-^>
which, given
.^
>
0,
/
(0M0 (N) +
.
equivalent to
4>M9 {N)
contradicts (58), establishing that
+ ^,
p{D) > p{N),
i.e.,
that democracy
is
captured.
p{D)>p {N) implies / {(pM9 (D) - - ^) > f {(j)M9 (N) + C),
< cpAdd {N)+£,. Since F is strictly monotonic, this implies p (A^) > p{D),
But, by the same reasoning,
thus
(?!)M6'
(
LI )-?7-^
establishing the
The
-q
Markov regime-switching
equilibrium in this proposition
have encountered so
fluctuations between
far.
is
The equilibrium
structure.
the richest and perhaps the most interesting one
still
takes a
Markov regime-switching
democracy and nondemocracy; but
in
democracy, there
that economic institutions will be those favored by the citizens.
41
While
we
structure with
is
no guarantee
in all the previous
models we studied, the
were able to impose both their
elite
we have an equilibrium pattern whereby democracy
the same time, here
are able to impose their favorite economic institutions.
(given Assumption
the
2')
elite will
intuition for this
in
(somewhat paradoxical)
democracy to nondemocracy. Consequently, the
in
democracy that they are able
of investing in their de facto political
nondemocracy.
result
is
that in democracy there
elite invest in their
is
an addi-
to induce a switch from
is
de facto power sufficiently
economic institutions with a
power
in
democracy
to
is
significantly higher.
democracy that they are
In fact,
of democracy, u {S
willing to invest
— D) >
What about
citizens? If there
were no additional benefit
then citizens would actually be worse
0,
off in
democracy than
nondemocracy, because they only care about economic institutions and economic
tions are
v {S
more
= D)
it
de facto political power in democracy and obtain the labor repressive economic
in their
institutions with a high probability.
in
which
to obtain their favorite
nondemocracy
precisely because they prefer
more
elite
Nevertheless, the elite are happier in nondemocracy, because the cost
greater probability.^^
is
but the
In fact, the proposition shows that
tional benefit for the elite to invest in de facto political power,
more
persists,
at
be able to impose labor repressive economic institutions
with a higher probabihty under democracy than
The
and economic wishes
political
=^ 0,
likely to
citizens
be labor repressive
in
democracy than
institu-
nondemocracy. Thus when
in
would never choose democracy, and would be happy to remain
democracy (given the limited
ability that they
in
non-
have to solve the collective action problem).
Therefore, the ideological or other benefits of democracy encapsulated
inu {S
= D) >
create
the possibility of the captured democracy equilibrium, whereby the citizens are willing to vote
and defend democracy, but democracy
at the
end caters to the wishes of the
elite.
Therefore, this model features both state dependence in political and economic institutions,
and
(i.e.,
also coexistence of
change in
political institutions
with persistence in economic institutions
the presence of labor repression in democracy). In fact,
the larger
is
(^,
the more likely
is
it is
the configuration with stable democracy choosing economic in-
stitutions in line with the interests of the elite. Consequently, this
institutions
straightforward to see that
and captured democracy provides both the
model with durable
political
richest set of predictions regarding the
persistence of economic and political institutions, and a potential explanation for the patterns
discussed in Section
^^This result
is
7,
which
illustrate
how many
societies
become and remain democratic,
not independent of functional form assumptions. For example,
democracy where
if
\vc relax
Assumption
2', it
a
Markov-switching structure with state dependence, but the probability of labor repressive economic institutions
in democracy is positive but no higher than in nondemocracy, i.e., p(N) > p(D) > 0. Nevertheless, relaxing
Assumption 2' does not guarantee that such an equilibrium will exist; it only makes it possible.
is
possible to obtain an equilibrium with a semi-captured
42
political institutions still follow
but continue to pursue policies that favor the traditional landed ehte.
Effective
6
In this section,
reforms.
Reform
we
briefly discuss
how
institutional persistence can be broken
Our framework shows how the equilibrium path may
by
feature invariance
effective
(i.e.,
labor
repressive and generally dysfunctional economic institutions remaining in place despite shocks
that change the political organization of society) or state dependence (where dictatorship
more
likely to follow dictatorship
than
it is
to follow democracy).
is
Are there any major reforms
that could break these various types of persistence?
The comparative
static results
ticular, the results so far
democracy
kets
if
is
show that a change
be effective
likely to
above suggest potential answers to this question. In par-
(in
in political institutions
from nondemocracy to
terms of leading to equilibrium competitive labor mar-
and persistent democracy) under two alternative (but complementary)
democracy creates a substantial advantage
then as shown by Corollary
1
for the citizens in the
scenarios.
First,
form of a large value of
77,
end the cycle of institutional persistence and make the
this will
permanent consolidation of democracy an equilibrium.
Second, one of the following reforms
racy,
then the economy
nomic
institutions:
(1)
is
less likely to
a reduction in
is
undertaken simultaneously with the switch to democ-
switch back to nondemocracy and labor repressive eco(p£,
in
terms of the model of subsection
4.1, so
that the
traditional elites are less able to control politics in a democratic society (for example, prevent-
ing local threats of violence or the capture of political parties by the traditional elites would
achieve such an outcome).
(2)
a reduction
in
AR,
for
A, which will reduce the potential rents that the landed
them from
example, by means of an increase in
can obtain and
elites
even
discourage
investing in de facto political power.
This discussion therefore illustrates that while politics as business-as-usual
elite
will
in
democracy, undertaking simultaneous and significant reforms
may
favor the
may change
the
character of the political equilibrium, making democracy and competitive labor markets more
likely.
An
attractive
example of simultaneous reform leading to a
distribution of political power in society
is
significant
change in the
the 1688 Glorious Revolution in England, which
not only changed de jure power by dethroning the Stuart monarchy, but also by substantially
increasing the role of the Parliament and the allocation of economic resources in society,
reversibly altered the distribution of de facto political
power
Weingast, 1989, Acemoglu, Johnson and Robinson, 2005a).
43
(see,
for
ir-
example. North and
Another interesting example of
simultaneous reforms arises in the next section where we describe
economic institutions of the U.S. South, having survived the
Civil
how
War,
central aspects of the
finally
changed
in the
1960s.
7
Historical Perspective
We now examine
some case
studies of the persistence of power, elites
U.S. South, Latin America and Africa (Liberia).
Our
objective
hensive historical account nor to test our model. Instead,
is
and institutions from the
neither to provide a compre-
we wish
to illustrate salient cases
of coexistence of change in political institutions with persistence in the underlying economic
system, which has been the motivating empirical pattern for our theoretical analysis. In
cases,
we
will interpret the historical events
though we
among
will also
historians
all
through the lenses of the models presented above,
attempt to highlight potential disagreements or alternative interpretations
when they
exist.
Persistence of the Elites in U.S. South
7.1
"De landlord
ever'body
an
An
official
is
is
landlord, dc politician
landlord, en
we
is
landlord, de judge
ain' got nothin."
is
landlord, de shurf
illustrates
landlord,
Testimony of a Mississippi sharecropper to
of the Agricultural Adjustment Administration in 1936,
important example which
is
Schulman
(1994, p. 16).
our thesis in the continuation of the economic
system based on labor repression, plantation and low- wage uneducated labor
in the U.S.
South
before and after the significant changes in political institutions brought about by the Civil
War.33
Before the Civil War, the South was significantly poorer than the U.S. average income at
about 70% of
GDP
per-capita
(e.g.,
Barro and Sala-i-Martin, 1992, Easterlin, 1960). ^-^ The
South lacked industry (Bateman and Weiss, 1981, Wright, 1986, Table
and
in
1860
the total manufacturing output of the South was less than that of either Pennsylvania,
New
2.4, p. 27)
'' Although many features of the Southern economic system, such as the plantation system and the reliance on
low-wage labor, persisted, as Wright (1986) emphasizes, there were also important discontinuities, for example,
the fact that the planter class transformed itself from "laborlords," mostly relying on their slave property, to
"landlords," more reminiscent to the large-scale plantation owners of South America.
^""The relative poverty of the South has been hotly debated by scholars. Although Fogel and Engerman (1974)
pointed out that if the South had been an independent country in 1860 it would have been amongst the 10
richest in the world, the consensus view is that the South was relatively backward with poor institutions in 1860
(a view which in its modern form goes back at least to Genovese, 1965). To the extent that it was relatively
prosperous compared to other parts of the world at the time, this was because it had recentlj' benefited from
a huge boom in cotton prices (Wright, 1978) and also because it weis embedded in a society whose institutions
were formed in the 17th century, before the development of the plantation economy.
44
York or Massachusetts (Cobb, 1984,
(around
9%
as
opposed to 35%
p.
The South had very low
6).
and
in the Northeast)
relatively httle investment in infrastruc-
ture. For example, the density of railroads (miles of track divided
The
higher in the North than in Southern states.
by land area) was three times
situation with respect to canal mileage was
more importantly,
similar (Wright, Table 2.1, p. 21). Perhaps
rates of urbanization
especially in the context of the
potential for future economic growth and industrialization, the South was not even innovative
for the sectors in
which
Although there
in various
ways
is
it
specialized.
no consensus about why the South was backward,
economy and
to the planation
slaves were a mobile asset, there
in public
"'''*
Wright (1986) argues that because
slavery.
was no incentive
scholars relate this
all
for planter interests to
support investment
goods such as infrastructure, and so manufacturing could not develop. Bateman and
Weiss (1981) show that Southern planters did not invest in industry, even though the rate of
return was superior to that in agriculture.
is
A
plausible explanation for the lack of innovation
Slaves were forbidden to
that slavery limited the possibilities for productive investment.
own property
them
or to
become educated
But
easier to control.
low-skilled labor forces
in
most Southern
states,
presumably because
this pattern of labor repression also
and possibly removed the incentives
made
this
condemned plantations
of planters to innovate. It
is
to
also
possible that the lack of urbanization and industry was a consequence of concerns about the
control of slaves and collective action, though Goldin (1976) disputes this.
In the aftermath of the Civil War, the income per-capita of the South
of the U.S. average.
South had been
If
the organization of the slave
relatively
backward
slavery in 1865 would have
historical interpretations
the Southern economy. '^^
removed
in 1865,
this
economy based on labor
blockage to Southern prosperity.
of slavery
plairters initially tried
repression.
Cut
off
The evidence and
had a surprisingly small
and
effect
failed to reintroduce the
War emerged a
on
gang
low wage labor
from the rest of the United States, income
per-capita remained at about half the average until the 1940s
converge.''^ Just as before the Civil
50%
economy had been the reason why the
labor system with the freed slaves, out of the ashes of the Civil
intensive
to about
one might have imagined that the abolition of
show that the abolition
Though
fell
when
it
finally
began slowly to
War, there was systematic underinvestment
in education
^^For example, during the period 1837 to 1859 while the average numbers of patents issued per year relating
to corn and to wheat were respectively 12 and 10,
^^
it
was just
1 for
cotton (see Schmookler, 1972).
why Southern incomes fell after the
War. The interpretation by Fogel and Engerman (1974) is based on the idea that the slavery system was
relatively efficient, though this has been challenged, see, for example, Ransom and Sutch (2001).
'^Wright (1986, p. 70) notes "the isolation of the southern unskilled labor market was a basic background
condition for virtually the whole epoch between the Civil War and World War II."
Despite this consensus, there
is
a debate on the related matter of
Civil
45
(Margo,
1990)."^^
Wright, 1986,
The main
79). In
p.
incentive for this seems to have been to
1900
all
impede migration
(see
but two of the non-Southern states had enacted compulsory
schooling laws, while none had such laws in the South except Kentucky (Woodward, 1951, p.
Though
399).
industrial development did begin
more systematically
after 1865,
Cobb
(1984,
p. 17) notes:
"The industries that grew most rapidly
of an underdeveloped
materials
...
economy
in that
in the post-Reconstruction
decades were typical
they utilized both cheap labor and abundant raw
such industries hardly promised to elevate the region to economic parity with
the rest of the nation"
So why did the economic system of the South change so
little
especially given the significant changes in political institutions?
economic institutions appears
at
At
odds with the significant changes
following the Civil War,
first,
this persistence of
in the distribution of de
jure power that took place after the Civil War, for example, with the enfranchisement of the
freed slaves,
political
We
and the repeal of the Missouri compromise, which had previously cemented the
power of the South
believe the answer
in the federal government.
related to the forces emphasized in our model, in particular, to
is
the exercise of de facto political power by the Southern landed elites to compensate for the
loss of their
de jure political power. ^^ Consistent with our approach, there was considerable
persistence in the identity and power of the political
elites.
counties of the black belt of western Alabama. Tracking
the persistence of the planter
elite in 5
families from the U.S. census
and considering those with
that
(p.
9) "of
the 236
members
For example, Wiener (1978) studied
at least $10,000 of real estate, he
of the planter elite in 1850, 101
Interestingly, this rate of persistence
remained
found
in the elite in 1870."
was very similar to that experienced
in the
antebellum
period; "of the 236 wealthiest planters families of 1850, only 110 remained in the elite a decade
later"
(p.
9).
Nevertheless,
"of the 25 planters with the largest landholdings in 1870, 18
(72%) had been in the ehte families in 1860; 16 had been
in the
1850
elite
group." ^° Table
is to spoil a good field hand."
Another possibility would be that the continuation of large plantation agriculture was because of its relative
economic efficiency. Yot this seems inconsistent with the available evidence. For example, Wright (1986, p.
84) argues "The plantations survived the Civil War, and their survival had little to do with their efficiency as
producing units ... cotton and tobacco could be grown just as efficiently on family-sized farms. No, the key to
the survival of the plantation was the ability of the former slave owners to hold on to their land in the midst of
intense legal and political struggles after 1865. In national politics, the planters successfully blocked proposals
for land confiscation and redistribution to the freedmen."
'"' Death during the Civil War appears to have had little impact on the persistence of the planter elites because
the law exempted one slaveholder from military service for every 20 slaves held (Wiener, 1978, p. 18).
''^Kousser (1974, p. 17) records a post-bellum adage "To educate a 'nigger'
^^
46
1
reproduces these data from Wiener (1978) and shows the high degree of persistence in the
identity of Southern landed ehtes."^^
After the end of the Civil War, more or less the same group of planter elites controlled the
land and used various instruments to re-exert their control over the labor force.
economic institution of slavery did not
specific
persist, the
Though
the
evidence shows a clear line of per-
sistence in the economic system of the South based on plantation- type agriculture with cheap
This economic system was maintained through a variety of channels, including both
labor.
control of local pohtics
words of W.E.B.
in the
and exercise of potentially violent de facto power. As a consequence,
Du
Bois (1903, p. 88), the South became "simply an armed
camp
for
intimidating black folk."
The
planter elite successfully staffed or co-opted the
whose remit was to supervise the
members
of the Freedmen's Bureau,
freed slaves. In 1865 the state legislature of
Alabama passed
the Black Code, an important landmark towards the repression of black labor. Wiener (1978,
describes this
p. 58)
as:
"The Black Code
the planters a reliable supply of labor
laborers".
of
Alabama included two key laws intended
to assure
—a vagrancy law, and a law against the 'enticement' of
These laws were designed to impede labor mobihty and reduce competition
in the
labor market. ^^
In addition to moulding the legal system in their favor, "Planters used Klan terror to keep
blacks from leaving the plantation regions, to get
cotton
field"
World War
(Wiener, 1978,
Key
II,
(1949, p. 9)
South both before and
small minority
A
p.
62).
them
sums up the pattern
after the Civil
to work,
In his seminal study
''^
War
of.
and keep them
at work, in the
the pohtics of the South after
of persistence of the institutions of the
as the "extraordinary achievement of a relatively
— the whites of the areas of heavy Negro population."
key to the persistence of the antebellum system after the Civil
War was
the continued
control over land. For example, in the debate over the redistribution of 40 acres of land to the
freedmen (vetoed by President Andrew Johnson
Juhan argued (quoted
in
Wiener, 1978,
in 1865),
Congressman George Washington
p. 6):
""Other studies find similar persistence in other parts of the South. See, for example, Ransom and Sutch
78-80) on landonwership in Dallas County Alabama, Huffman (1974) on Clarke County, Georgia,
(2001, pp.
on North Carolina. This research to some extent contradicted earlier studies which had seen
the identity of Southern elites after the Civil War (e.g., Woodward, 1951)
''"The extent to which southern planters were able to exert classical monopsony power is controversial, see
Alston and Kauffman (2001). Nevertheless, economic historians do agree on the use of coercion, intimidation,
violence and segregationist legislation to repress and control free black labor.
''''Relatedly, Alston and Ferrie (1989) show how planters controlled their labor force by offering them "security"
from violence and lynching.
and
far
Billings (1979)
more change
in
47
"Of what
avail
would be an act of congress totally abolishing slavery
agricultural basis of aristocratic
A
power
shall
if
...
the old
remain?"
third strategy, again consistent with the emphasis on the de facto political power of the
elite in
our theoretical analysis, was control of the local political system. Following the Civil
War, the period
called 'Reconstruction' lasted until 1877 (see
Stampp, 1965, and Foner, 1989).
In this period Republican politicians contested power in the South and, with the help of the
Union Army, engineered some
social changes. Nevertheless, this
in the guise of support for the
Democratic Party and the so-called 'Redeemers.' In 1877,
induced a systematic backlash
in the
context of a log-roll between President Rutherford Hayes and Southern national politicians,
Union
after
soldiers
were withdrawn from the South and the region
left
1877 then marked the real recrudescence of the antebellum
its
own
elite.
The
to
devices.
The period
'redemption' of the
South involved the systematic disenfranchisement of the black (and poor white) population
through the use of poll taxes and literacy tests (Key, 1949, Kousser, 1974) and the creation of
the one-party Democratic regime. ^^
Key
(1949, pp.
309-10), in his analysis of the primary elections of the Democratic party,
noted the hegemony of southern society's "upper brackets" and the
of
its
"lower brackets."
garchy over
politics,
He
discusses in detail the control of
noting that
(p.
been achieved through the elevation
"The
211):
political marginalization
North Carolina's economic
oli-
effectiveness of the oligarchy's control has
to office of persons fundamentally in
harmony with
its
viewpoint."
This picture
is
also confirmed
by the analysis of Wright (1986,
in the 1930s, southern representatives in
push
for
new
Washington did not use
federal projects, hospitals, public
as the foundations of the low-wage regional
Jim Crow laws
postbellum South into an
lives.
As
in
gamut
—was enacted (Woodward, 1955,
78),
who
writes "Even
their powerful positions to
works and so on. They didn't, that
economy
In addition to disenfranchisement a whole
p.
is,
as long
persisted."
of segregationist legislation
for the classic analysis).
effective "apartheid" society
—the so-called
These laws turned the
where blacks and whites
lived different
South Africa, these laws were aimed at controlling the black population and
its
labor supply.
example, notes: "Two-party competition would have been fata! to the status of
would have meant in the 'nineties an appeal to the Negro vote and it would have meant
...
Negro rule in some black-belt counties. From another standpoint, two-party competition would have meant
the destruction of southern solidarity in national politics ... Unity on the national scene was essential in order
that the largest possible bloc could be mobilized to resist any national move towards interference with southern
authority to deal with the race question as was locally desired."
8-9), for
''''Key (1949, pp.
black-belt whites.
It
48
Consequently, the South entered the 20th century as a primarily rural society.
an agrarian society with a backward technology that
virtually unassisted
"It
remained
employed hand labor and mule power
still
by mechanical implements" Ransom and Sutch (2001 pp. 175-176).^^ In
1900, the South's urbanization rate was 13.5%, as
compared
to
60%
in the Northeast
(Cobb,
1984, p. 25).
Ransom and
Sutch's (2001, p.
186) assessment of the implications of this economic
pohtical system in the South for economic progress
is
and
representative of the consensus view:
"Southerners erected an economic system that failed to reward individual initiative on the
part of blacks and was therefore ill-suited to their economic advancement.
inequities originally inherited
all
whites used threats of violence to keep blacks from gaining an education, prac-
the three routes most
advancement.
commonly
Witla over
agricultural laborers,
it is
lialf
by other Americans
travelled
from following
in their quest for self-
the population held in ignorance and forced to work as
no wonder that the South was poor, underdeveloped, and with-
out signs of economic progress"
all,
this effort
economic growth"
ticing a trade, or purchasing land, they systematically prevented blacks
All in
result, the
from slavery persisted. But there was a by-product of
at racial repression, the system tended to cripple
"When
As a
Ransom and Sutch
(2001, p. 177).
the Southern equilibrium, based on the exercise of de facto power by the landed
elite,
plantation agriculture and low-wage, uneducated labor, persisted well into the 20th cen-
tury,
and only started to crumble
World War
after
II.
Interestingly,
of this Southern equilibrium, that the South started
its
was only
it
after the
demise
process of rapid convergence to the
North.
What
Section
6,
caused the collapse of the Southern equilibrium? Consistent with the emphasis in
it
of factors.
seems that
Initially,
this collapse
was brought about by the juxtaposition of a variety
the depth of the Great Depression weakened the resolve of Southern
politicians to block federal policies that
might integrate the Southern economy into the
rest
of the nation (Schulman, 1994). Other important factors include the great outmigration and
the impact of the simultaneous mechanization of cotton picking on the labor requirements of
agriculture.
As blacks became harder
became
to trap in the South, they also
less
necessary to
generate rents. ''^
""^See
Whatley (1985)
for
a hypothesis explaining the lack of innovation
in the
post-bellum South
in
terms of
the poor incentives created by sharecropping contracts.
'"'An interesting question, which is not central to our interpretation, is whether mechanization of cotton
reduced the cost of black migration to the Southern landed elite, who then tolerated such outmigration, or
49
Wright's (1999, Table
1, p.
281) data on net migration from the South shows the large
increase in black outmigration starting in the 1940s and Heinicke (1994, Table
506) shows
1, p.
the concurrent rapid spread of mechanized cotton picking. First introduced in 1949, by the end
of the 1950s mechanical cotton picking
was the
half of the crop in the key cotton states
was being picked mechanically.
rule rather
than the exception
in the
South;
In addition to these changes, the Southern equilibrium was disrupted by the collective action
of the Civil Rights
as
movement and a sequence
Brown versus Board
Act of 1965. These
elite to exercise
of
Supreme Court and government
of Education in 1954, the Civil Rights
significant events
number
Act of 1964 and the Voting Rights
can be interpreted as a decline in the ability of the landed
McAdam's
de facto political power in democracy.
739) data on the
decisions, such
(1983, Figures
of civil rights collective actions (bus boycotts
and
and
1
sit-ins)
sharp increase starting in 1960 (see also Chong, 1991), while Wright (1999, Figure
2, p.
shows a
2, p.
273)
records the dramatic increase in the registration of black voters in Southern states over this
period.
It
thus became increasingly more difficult for the Southern landed elites to control
the political process (see also the detailed analysis of the impact of the Voting Rights Act on
economic growth
in the
South
Persson and Sturm, 2005).
in Besley,
Persistence of Elites and Economic Institutions in Latin America
7.2
Despite the significant changes in political institutions that have taken place since the colonial
era in Latin America, the capture of political and economic institutions by elites
theme
of
much
of the historical literature
politics
by traditional
continuation of labor repressive policies in agriculture in
Chile,
Some notable examples
many
cheap labor
is
therefore
how and why
in agriculture failed to
by traditional
in
Central and Latin American
(or lack thereof) in Central
A
change following the
South, the answer
elites in
many
and the options open to the traditional
in the
and Latin America and
political reforms that
have gradually
in these societies.
related to the exercise of de facto political
the context of democratic politics.
Colombia, for example, took
whether the changes
is
central question for
the economic system based on labor repression and
brought more democratic means of collective decision-making
in the U. S.
and the resulting
include Zeitlin (1984) and Zeitlin and Ratcliff (1988) on
understanding the economic development
As
elites
Smith (1979) on Mexico, and Paige (1997) on Central America.
the Caribbean
an enduring
on the sub-continent. There are numerous studies
documenting both the capture of democratic
countries.
is
different forms,
elites,
The capture
of democratic pohtics
depending on the
with the most
common
power
specific conditions
forms being the actual
supply of cheap labor caused by black outmigration encouraged mechanization.
50
domination of
and use of
political parties
elites, electoral
fraud, blocking of entry of
new
parties,
violence.
The most obvious form
offices
by the
of political capture
by members of the landed
America and has been
carefully
elite.
Figures
and
1
the actual running of political parties and
This has been particularly the case in
documented by Stone (1975,
shows the extraordinary extent to which
American countries from the
is
elites
much
of Central
1990). His genealogical research
have persisted in Costa Rica and other Central
colonial times.
2 (from Stone, 1990)
show the
striking
number
of presidents since indepen-
dence in various Central American countries that are related to two conquistadors, Cristobal
de Alfaro and Juan Vazquez de Coronado, who both arrived in Central America in the 16th
Figure
century.
1
shows that no
25 of them in Costa Rica.
Nicaragua
for
less
that 48 presidents were direct descendents of Alfaro,
Alfaro's descendents also include the
most of the 20th century
until the Sandinista
of 1850.^^
in
Somoza
clan and
many
of the
of Costa Rica related to Alfaro.
As a case study we now examine Colombia
from Spain
ruled
Revolution in 1979. Figure 2 shows
the 29 presidents descended from Vazquez, including again the
same presidents
Somoza dynasty who
in
more
detail.
1819 and the Liberal and Conservative parties
These two parties have maintained
their hold
on
early as 1856 the Liberals introduced universal male suffrage
Colombia became independent
first
competed
political
in the election
power ever
and as many
as
since.
40%
As
of adult
males voted (Bushnell, 1971). After the 1863 Constitution, voting rights were determined at
the state level, but several of
them maintained
universal suffrage (see Delpar, 1981).
19th century, the parties contested elections, but they also fought for power.
fighting
sometimes took the form of
civil
In the
Although the
wars, the parties also frequently reached collusive
agreements to make sure that other political parties, especially those representing peasants
and workers, were unable to enter the
political system. For
example, after the
civil
war
1899 to 1902, the two parties introduced a power-sharing agreement, which divided
seats in the legislature between the two parties by
and Robinson, 2004). This system lasted
know
as
La
of the 'incomplete vote'
until 1929. After a civil
war
all
the
(Mazzuca
in the 1950s (a period
Violencia), the parties resurrected this collusive agreement in 1958, splitting the
legislatm-e 50-50
this
means
fi-om
and
also agreeing to take turns at the presidency for 16 years.
power sharing agreement endured
until the
In practice
middle of the 1980s.
'"'The best studies of the origins of the parties are Colmenares (1968), SafFord (1972)
and Delpar (1981).
Some trace the alignments to factions that formed in the late 1820s over whether or not Simon Bolivar should
be made Emperor, though there have also typically been significant policy differences between these two parties.
51
The
Liberals and Conservatives
managed
to successfully block entry of other parties to
the political system through a variety of strategies. Even after the reintroduction of universal
suffrage in 1936, fraud, violence,
and engineering the
were widespread.'*^ Electoral fraud
many
forms
Fraud was
in fact, a recurrent
theme
in
Colombian
politics
and took
example, Bushnell, 1993, and Chaves, Fergusson and Robinson, 2005).
(see, for
also used
is,
electoral rules to create barriers to entry
by the traditional parties to directly derail populist challenges, such as
during the attempt by Rojas Pinilla to win the 1970 presidential election.^^
Violence has also been endemic in Colombian politics. The potential "Peron" of Colom-
Jorge Eliecer Gaitan,
bia,
murdered
in 1948 just
who was
when
it
likely to
seemed
likely
pursue populist and redistributive pohcies, was
he would become the leader of the Liberal party.
Luis Carlos Galan was murdered in 1990 just as he was emerging as the leader of the Liberal
party.^° In the 1980s the first alternative political party to seriously
of the National Front in 1958
in the 1980s
was the Union Patriotica. This party was
by the systematic murder of
its
candidates
(see, for
Finally, as in Central America, the political parties in
family
ties.
emerge since the beginning
The Conservative Ospina
effectively obliterated
example, Kline, 1999).^^
Colombia have featured strong
family has provided three Presidents, Mariano Ospina
Rodriguez (1857-1861), Pedro Nel Ospina (1922-1926) and Mariano Ospina Perez (1946-1950).
The
Liberal Lopez family provided not just Alfonso between 1934 and 1938, and again between
1942 and 1945, but also Alfonso Lopez Michelsen between 1974 and 1978. The Liberal Lleras
""^This
is
related to the absence of a socialist party in Colombia.
Some
scholars have suggested that the
relatively egalitarian frontier expansion that took place in the second half of the 19th century has contributed
which checked radicalism in Colombian politics. This
be very convincing in view of the fact that Colombia has one of the most unequal
distributions of land ownership and income in the world (on this Antioqueno expansion see Christie, 1978).
""^
Another factor that appears to have contributed to the continuation of the two-party system is the way
the electoral system created incentives for dissidents to stay within the traditional parties. For example, the
electoral system ensured that candidates could be elected with far fewer votes than the electoral quotient, thus
potentially guaranteeing that elite factions would get representation. At the same time, the traditional parties
allowed "dissident lists" to run, providing a platform to potential dissidents. In line with these ideas, Dix (1967,
p. 250) notes "proportional representation may have helped to a degree to reinforce the two-party system by
to the creation of a middle class of property owners,
thesis does not appear to
allowing dissident factions to gain representation according to their strength in the electorate while
forcing
them from the
party. Retaining the party label, or
some version
of
it,
still
not
they have usually been reabsorbed
two elections."
Gaitan and subsequently Alfonso Lopez Michelsen and Galan, considered forming a third party, but then always returned to the fold because the electoral system made success by third
parties difficult. Indeed, it is quite likely that the assassination of Gaitan was precisely because he had finally
emerged as the leader of the Liberal party, not because political rivals anticipated the creation of a third party.
''in the 1986 elections, Uni6n Patridtica obtained 5 seats in the Senate and 9 in the Chamber of Representatives at the national level, and 14 deputies, 351 councilmen and 23 municipal mayors at the local level. By 1988,
the party was wiped out, however. The Amnesty International reported that more than 500 of its members,
into the officialist ranks after one or
^"Political entrepreneurs, like
including
its
previous presidential candidate, Pardo Leal, and 4 congressmen, had been assassinated, most likely
with the involvement of members of the Colombian military and government.
52
family provided presidents in 1945-1946 and 1958-1962 (Alberto Lleras Camargo) and again
The Conservative Pastrana family provided Misael
in 1966-1970 (Carlos Lleras Restrepo).
Pastrana between 1970 and 1974 and Andres Pastrana between 1998 and 2002. The son of
Conservative President Laureano
in the late 1980s
Gomez
(1950-1953), Alvaro was the Conservative challenger
and early 1990s and the father of Liberal President Julio Turbay (1978-1982)
had previously been the most powerful
figure in the Liberal party.
Overall, through a variety of means, including elite domination of politics, electoral fraud,
blocking of the creation of
new
and violence, traditional
parties
elites
to have a disproportionate effect in the democratic politics of Central
We
seem
to have
been able
America and Colombia.
conjecture that this pattern of elite dominance has been an important element in the
continuation of the economic system based on cheap labor and quasi-labor repression in the
countryside of
many
of these countries.
Persistence of Elites and Economic Institutions in Africa
7.3
The
issues
we study here
are of importance outside of the Americas, and here
we
illustrate
them
with the example of Liberia. Modern Liberia was started in 1820 by the American Colonization
Society (ACS) as a
from
local chiefs
home
and repatriated African
for freed
around the
site of
the
modern
capital of
slaves. ^^
The ACS bought land
Monrovia (named
after the then
President of the United States James Monroe), and in 1847 Liberia became independent of the
ACS. The year 1877 saw the emergence
politics until the
coup of Samuel Doe
of the True
century. This group set themselves
as
if it
were a colony.
By
up
TWP
comprised of the descendents of the
who dominated
the party and the country for over a
in 1980.
repatriated slaves, the Americo-Liberians,
as
Whig Party (TWP), which would dominate
an
The
elite
over the 'tribal peoples' and ran the country
the 1960s Americo-Liberians were about 3-5% of the population.
Despite the early emergence of the Americo-Liberian
elite in Liberia,
the political system
has relied on elections since the 19th century, and became fully democratic starting in 1944.
Nevertheless, like their Central American or Southern U.S. counterparts, the Americo-Liberian
elite
have been able to dominate democratic pohtics through a variety of means, including
extreme electoral fraud and violence, and their domination of
politics
has been a major cause
of the continued economic backwardness of Liberia.
The
first
study of the economic backwardness of Liberia was undertaken by a team of U.S.
based economists from Northwestern University hired by
USAID
in the early 1960s (Clower et
This discussion draws on the general history of Liberia by Sawj'er (1992).
53
al.,
1966). Dalton (1965, p. 581)
by the
elite of
TWP
the
sums up
their findings
and Americo-Liberians and that
economic backwardness of Liberia
"...the
by concluding that Liberia was ruled
domination by foreign financial or
is
not attributable to the lack of resources or to
political interests.
that the traditional Americo-Libcrian rulers,
who
The underlying
difficulty
is
rather
fear losing political control to the tribal
people, have not allowed those changes to take place which are necessary to develop the
national society and economy.
...
Like the Portuguese in Angola or the Afrikaners in South
Africa, the rulers of Liberia are the descendents of an alien minority of colonial settlers.
Americo-Liberian families."
The
TWP
used indirect rule adapted from the British to control the hinterland and ma-
As Glower
nipulate chiefships.
laws, access to education
Dalton (1965,
p.
et al.
and the
record, Liberia
legal
became a two
class society
system governed Americo-Liberians and
584) points out that "Ironically,
it is
and
different
tribal peoples.
the ethic of Mississippi that most nearly
characterizes their outlook: to retain power in traditional fashion and keep the natives in their
place."
Before 1944, the hinterland had no political representation at
the
all,
and even
after
TWP were able to cling to power through their complete control of the economy,
significantly,
The
it
did so,
and more
through intimidation, corruption and coercion. ^'^
resulting
economic institutions in Liberia certainly approximated the labor repressive
institutions in the
model discussed above.
In the 1960s tribal peoples were
still
subject to
forced labor on public works and rubber plantations (1/4 of the labor force in early 1960s)
and the economy was
either completely controlled
by the
TWP
or by foreigners.
The
apparatus became a huge patronage machine dedicated to promoting the interests of the
state
TWP
and the Americo-Liberians.
A
key feature of Liberian
politics, facilitating the exercise of
power by the
kinship networks of the Americo-Liberians (see Fraenkel, 1964). Dalton
understand Liberian
politics,
knowledge of kinship connections
of the Liberian constitution." Figure 3 shows
the kinship network of President
how
Tubman, who
is
(p.
more
TWP, was
the
589) argues that "to
useful than knowledge
higher political offices were monopolized by
ruled between 1944 and 1971 and was from
one of the established families of the Americo-Liberians.
^''indeed Liberia has the sad reputation as having
King's election
in 1927.
The Guinness Book
had the most corrupt election ever held for Charles B.
most fraudulent ever
of Records (1982) qualified the elections as the
reported in world history. Suffrage was constitutionall}' limited to some 15,000 citizens,
but according to the
official election results
some 240,000 votes were
54
all
Americo-Liberians,
cast for Charles B. King.
After his defeat in the 1927 presidential elections,
elect,
Thomas Faulkner accused
the President-
Charles D.B. King, of allowing slavery to exist in the Republic and that certain highly
placed government
in the forced shipping of laborers to the
were engaged
officials
wave
island of Fernando Po. Faulkner's accusations spurred a
Committee of the League
of Nations, established to
method
of recruiting carried compulsion with
were shown to have
illegally
misused their
it.
officials
in Liberia, Firestone. ^^ In essence the
associated with slavery because the
office in recruiting
official
positions
labor with the aid of the Liberian
TWP). Not
benefited from forced labour, but so
and a
allegations, indeed concluded
Moreover, persons holding
Frontier Force (the national military controlled by the
government
of international reactions,
examine the
Po and Gabon was
that shipment' of laborers to Fernando
Spanish
only had Americo-Liberian
had the most major
TWP had offered Harvey Firestone
foreign
company
a guaranteed cheap
labor supply through coercion.
There can be few better
the
TWP.
Initially,
historical
examples of an
than the Americo-Liberians and
they were able to restrict de jure power only to themselves, completely
disenfranchising the indigenous Africans.
facto
elite
They were
power because they had the support
of the United States
time the political system opened. After 1944 the
of peoples in the hinterland.
also able to
TWP
They were nonetheless
dominate
in their use of
de
and superior weapons. Over
had to deal with the pohtical inclusion
able to use their de facto
power
to maintain
the economic institutions they favored.
Therefore, the domination of politics by the Americo-Liberian elite and their choices of
(highly inefficient) economic institutions to redistribute resources to themselves gives another
example of the successful exercise
of de facto
power by an
elite
both under nondemocratic and
democratic political institutions.
8
Conclusions
Almost
all
theoretical
and empirical research
in political
economy
starts with the
presumption
that institutions, once in place, persist and shape the political-economic interactions of different
groups and agents. Nevertheless,
institutions, such as the
many
societies experience frequent
changes in their political
end of the colonial era and the creation of the republics throughout
the 19th and 20th centuries, or multiple switches between dictatorship and democracy during
''''Vast
rubber plantations were started by Firestone in 1926. Firestone obtained a one million acre concession
corresponding to 10% of what was considered arable land in Liberia), was granted
for a 99-year period (roughly
the exclusive rights upon the lands selected, and became
— with small, exceptions — exempted of
future taxes.
55
all
present and
the 20th century. Certain economic institutions also change, with slavery, forced labor, and the
encomienda coming to an end
and privatization of industry
America and Caribbean, and waves of nationalization
in Latin
many
in
less
developed countries. In the face of this picture of
frequently changing (political and economic) institutions, do institutions really persist?
This paper has argued that the answer
which are essential
stitutions,
institutions, or even specific
is
yes, in the sense that the
broad economic
economic outcomes, can persist even when
for
economic institutions, change. In
specific political
with this perspective,
line
in-
it
has
proposed a simple model of the coexistence of change and persistence in institutions.
The
basic idea
more
sess
is
that economic institutions are decided by groups or individuals that pos-
political power. Pohtical power, in turn, consists of de jure
political institutions
and de facto
political power,
power regulated by formal
which comes from the
ability of various dif-
ferent social groups to solve their collective action problems, lobby or bribe politicians, capture
and control
political parties, or use paramilitaries or other
means
Changes
of repression.
example a change from nondemocracy to democracy,
specific political institutions, for
the distribution of de jure political power.
This, however,
may
in
affects
induce cohesive (and small)
groups, such as (landed) elites to increase their investments in de facto political power to offset
their loss of de jure political power.
To
illustrate this
mechanism
in its starkest form, our baseline
technology of generating de facto political power for the
elite is
model assumed that the
the
nondemocracy, and demonstrated the possibility of invariance, that
equilibrium distribution of economic institutions
mocratic or not
—despite the
fact that
jure power towards the citizens.
enough to
is,
in
democracy and
a pattern where the
independent of whether the society
democracy creates a
The reason
fully offset their loss of
is
same
is
de-
real shift in the distribution of de
for this stark result
is
that the landed elite invest
de jure power when society switches from nondemocracy to
democracy.
The
result that investment in the de facto
de jure power
the form of a
is
special,
and the analysis showed how, more
Markov regime-switching process with
switching structure emerges
for
when some
example, when democracy
the elite
(e.g.,
on
costs of changing
The
richest
tutions are
power by the
state dependence.
This Markov regime-
of the assumptions of the baseline
model are relaxed,
when
there are
in the short run.
presented in Section
difficult to
in
allowed to place limits on the exercise of de facto power by
economic institutions
is
changes
generally, the equilibrium takes
their capture of political parties or use of paramilitaries), or
model
more
is
elite entirely offsets
5,
and allows
for the feature that political insti-
change than are economic institutions. Under
56
this
assumption,
a pattern of "captured democracy"
to have a disproportionate
eflPect
arises,
whereby democracy endures, but the
on equilibrium economic
elite
are able
institutions. In fact, in this model,
paradoxically, the probabihty of labor repressive economic institutions in agriculture
higher in democracy than in nondemocracy. This model
patterns discussed in Section
7.
useful for interpreting the historical
is
These patterns show the coexistence of change and
tence in institutions in the U.S. South, Latin America and Africa, and
may be
In
which
their
how
persis-
traditional elites
able to control democratic politics and maintain their favorite economic institutions.
the cases
all
may be
may
we
discussed, changes in
some important dimensions
have been potentially quite costly
de facto power,
for
for the elites,
of political institutions,
were neutralized by the exercise of
example, control of the party system or key
offices, or local violence.
These examples suggest that the mechanisms highlighted by our model may be important
understanding how frequent changes
in specific political institutions
in
go hand-in-hand with the
persistence of economic systems, with their broad distributional and efficiency characteristics
unchanged.
The mechanism proposed
economic systems
persist,
is
appears salient in practice,
is
one
in
in this paper,
where the identity of the
elites as well as
the
only one facet of institutional persistence. Another pattern, which
for
example, in cases such as Bolivia, Zimbabwe, Kenya and Ghana,
which the identity of the
elites
changes, but
the worst practices of their predecessors. This
is
new
elites
adopt policies
reminiscent of Michels' (1911) Iron
Oligarchy, as well as of the emphasis by sociologists such as Mills (1956),
Pareto (1968) on the persistence and creation of
in line
elites.
This pattern seems
but important area of study. Most ambitiously, future research
may
model that can explain the composition of elites, when existing elites
strive
persist,
Mosca
like
with
Law
of
(1939) and
a challenging
towards a unified
when elites change
but institutions persist, and when institutions truly change. Alternatively, a model isolating
the iron law of oligarchy
may be
the political arena, a direction
why
dysfunctional incentives persist in
are currentlj' pursuing in
Acemoglu and Robinson (2006b).
useful in understanding
we
57
Appendix: Proofs of Propositions 2 and 3
9
Proof of Proposition
Ay {6-' {N)
From
Let us
2:
0-'
,
(£>))
first
define
=V'{N\
9'' (N)
0"' (D))
,
the recursive formulations in (29) and (30), for
all
€
i
£"
- V' {D
\
0-' {N)
we have the
e^' [D])
,
first-order necessary condi-
tions:
't'lUlj:
'Pf
U( E
S'
e^
(^)
(D)
+
A (^ + 0^V'
(N) e-' [D)))
<
1
and
6'
>
0,
(59)
/3Ay' {e-' (N) e-' (D))^
<
l
and
e'
>
0,
(60)
and
9^
{D)
]
+9'-vj]
(^ +
{0-'
,
,
both holding with complementary slackness.
The proof proceeds
all
i"
in several steps. First,
Assumption
G f cannot be an equilibrium. Therefore there must exist some
i
e £
such that 9'"
(D)
>
not.
Then
respectively, (59)
1
it
and
must be the ca^e that
for
(60) holding as equalities
= 4>fU(
>
=
e £ such that
i'
9^
(AQ
=
>
for
and
0.
Second, we claim that there must exist some
Suppose
[N)
3 implies that 9^
j:
0'
^
{N)
<PfU(^p /{D) +
9^'
if
and
i'
and
9^]
+
6
i
for
whom
both (59) and (60) hold as
i" defined in the previous
(60)
and
(^^ +
equalities.
paragraph, we have,
(59) are slack. This implies
/3A1/'' (9-^'
{N)
9-^'
,
(D)))
-7A](^^+pAV^' {9-Un),9-Ud))^
or
Similarly for
f\4>\
z".
E
\je£,]Tti"
0^N)+9^
=
) I
/U
<
E^^AO
\3e£
\
f('p(EO'{D)-v]]=f\<P\
V^sc
V
J J
\
E
9^{D)
+
0'^v
\je£,j^i
yielding a contradiction with the previous inequality.
Third, the fact that there exists some
i
e £
for
whom
both (59) and (60) hold as equalities implies
that
f\4>\
Z/iN)\\ =
f
58
U iZ/ (D) - V
This
is
clearly only possible
if
ie£
ie£
which implies p{D)
= p{N) >
The
0.
fact that
p(D) = p{N) <
1
again follows from Assumption
2,
completing the proof.
argument establishes that
Finally, for use in the proof of Corollary 2, also note that this
z
e f
for
whom
Proof of Proposition
can be supported as an
We
3:
SPE and
eUte agents, there exists a
equality.
[0,1),
such that this distribution
p{D) = p{N) G
(0, 1).
The
special
part of the proposition, and the fact that as
first
/?
—
>
1
true for any distribution establishes the second part.
First
With
G
/3
involves equilibrium probabihties
case of equal contributions will establish the
is
any
prove this proposition by showing that for any Pareto optimal
will
among
distribution of contributions
this
both of these equations must hold as
(59) or (60) holds as equality,
for
the
suppose that a single individual controls
same arguments
all
problem of
as above, the
contributions by elite agents to de facto power.
can be written recursively
this indi\'idual
as:
V{N)=ma^l^-e + F{ct>e)(^^+0V{N)^+{l-F(4>e))(^^+(3V{D)
V{D) = m^xl^-e + Fi4>e-v)(^^+l3V{N)^+il-F{<pe-7i))(^^+pV{D)
where now 6
=
X^jg^
0''
the total contribution by elite agents, and these expressions only differ from
is
and (19) because the entire cost of contributions and the
(16)
taken into account.
Denoting optimal choices
in this
entire benefit in terms of land rents are
program by
9'
{N) and
6* (D), the first-order
conditions are
cPf
(0r
4>f {<P9*
(TV))
-V
[^ +
(^ + V{N)-V
P
{N)
(l>
[D) -r,)P
{D))^ <
1
and
{D)) <
1
and 6* {D) >
9' (TV)
>
0,
holding with complementaury slackness. Assumption 3 ensures that 9* [N) and 9* {D) have to be positive,
so the solution to this problem involves
4>9*
{N)
and thus p (D)
=
an equilibrium
satisfying the above conditions
elite
p [N) >
0.
That p (D)
p {N)
4,9*
<
1
(D)-T,
(61)
again follows from Assumption
2.
It is
made
Next we show that there
[N) and
9*
(D) across
elite
better off without
exists
e
[0,1)
clear that
would be Pareto optimal from the ex ante viewpoint of
agents (for some distribution of the total contributions 9" {N) and 9* (D) across the
since no elite agent can be
9*
=
=
some other
such that
for
elite
>
P,
agent being
elite agents),
made worse
off.
any distribution of contributions
agents can be supported as an SPE. To simplify the notation, consider a
stationary distribution (though, with more notation, the argument easily generalizes to any time- varying
59
distribution):
{6" (A^)}^^^ and
{^'*
(D)}
Consider a candidate
.
SPE
with this distribution and the
feature that following a deviation, the equilibrium reverts back to an 1VIPE.^° Recall from Proposition
any
2 that in
MPE
we have r
(tt
=
=
0)
and
= E,e£
{e'^iN)}^^^, {^"'(-D)},,^ such that e''{N)
Given
this
>
9^ [N) and 9* [D)
{r
played
Now
9^ {D). Let
=
/i*
denote the history
/?/
(iV)},^^, {9'' [D)}^^^, r (^
i
G
and
£:
and some contributions denoted by
all st
With
-
S^ (D)
=
=
0)
and
s'
which
in
nl4'
= E,e£^'^
any agent
for
=
{n
0)
{9" [D)}.^^, r
€ {N, D}, r
=
(tt
0)
some date
=
and
t'
s' (tt
<
=
t
from
v/4>.
to contribute
9*
we have
2,
-
{N)
>
have
= N.
=
(tt
i is
W
in all past periods, all agents
consider the following strategy profile to support the SPE:
there has been a deviation at
if
W=
^'^
the above single-agent maximization problem to Corollary
specifies that elite agents play {6*" (A^)}^^^,
(i.e.,
= N,
0)
punishment structure following a deviation, the best deviation
By comparing
nothing.
=
s' (tt
0)
h*
if
=
=
and
h^,
s' (tt
=
=
0)
6^ {t,st)
then
this play),
then the equilibrium
=
TV. If
/
/i*
A*
9^^ (sj) for all
= N.
0)
nondemocracy,
this strategy profile, following a deviation, say starting in
agent
elite
i
will
obtain:
^
V,''iN)=FL
e^*{N)\(^+pvaN))+(l-F(4'
E
(^+^^'(^))'
^'*(^)|
I
(62)
where i^Yliss
and
ijH
=
(N)
-0'P
(^)
MPE
{D) are the
Vi
V,
^^*
(TV)
'®
^^^ total contributions to the de facto power after the deviation, and Vi (N)
values in the punishment phase following the deviation,
+ F i^pe"
(TV))
+
(^^
PV, (TV))
+
F {cP9^
-
(1
{N)))
{?^ + PV, {D)\
(63)
,
and
V,.
{D)
=
-9'^ [D)
+ F [4>0^
{D) -
77)
[^ + m
(TV))
+
(1
- F
-
(00" (D))
r?)
(^ +
PV. {D)
(64)
The
value of deviation in state
9P (TV)
If,
=
9^ {D)
-
ri/(j),
we have
on the other hand,
TV and
9" {D) when
V,^ (TV)
= -9"
(TV)
s
= D, Vf
F {(j)9*
(D)
{N))
is
+ F (0r
s
=
Z?,
(TV))
defined similarly to (62). Since 6* (TV)
= F {4>9*
this agent follows the
=
when
s
{D)
-
77)
SPE strategy of
> F {4,9^
(TV))
"cooperating"
,
=F
i.e.,
=
{(f>9''
9* {D)-t]/4>
(D) -
>
77).
contributing
0"
(TV)
he will obtain
(^^
+ PV,^
(TV))
+
(1
- F (<^r
(AO))
[j^ +
/?V;'=
(D))
.
(65)
Similarly,
V[ [D) = -9'* [D) +
F (,^r
{D) -
77)
f^ +
/3K^ (TV))
+
(1
- F (^r
(i?))
-
77)
[j^ + /?y/ (Z?))
,
(66)
^°
Clearly, there
may
exist other
punishment
given discount factor, for example by using
strategies, potentially supporting a larger set of
minmax punishments
SPEs
for
any
for sufficiently high discount factors (e.g.,
Abreu, 1988). Nevertheless, for our purposes, it is sufficient to characterize the Pareto optimal SPEs with the
MPE punishment, since as ^ —» 1, these will coincide with the entire set of Pareto optimal SPEs.
60
and also
Qip
Qip
_
(TV)
Vf [D) and
for
MPE
Such an
for 5
(p)
V,.
e
MPE for
(D). Choose an
{£>, A^} for all
i
6
<f '
C
the punishment phase such that 9'*
=
such that S'
£•
{i
6'*
e 8
can always be constructed in view of the fact that 6" (N)
=
>
(N)
:
9*
{N)-e'* [D) >
>
or O'* [D)
-
[D]
>
t]/(J)
0}.
9^ {N)
=
i
6^ [D)
—
r]/4>.
Now
SPE
each agent would be happy to follow the
strategy as long as
Vf (D) > Vf (D) and V^ (N) > Vf {N)
These inequalities are naturally
because he
i
is
satisfied for all
e £\S' (since any such agent has no incentive to deviate
i
making zero contributions along the equilibrium path),
G C. Next combining (65) and
and using the
(66)
fact that,
we have
V^
and
(TV)
=
-0'* (TV)
+ F (0r
(N)) f
=
-e'^ (N)
+ F {4>9P
[N))
^M
By
F {(j)9*
virtue of the fact that
9" [D) - e" {N) >
e'P
(D) -
{N) and
P than
same observation
implies that for any
{9'^ (N),9'P (D)}.^^),
thus there exists
Let P
=
G
/3j
Vi
+P
(61),
F {cpe"
[9" [D]
to check
= F {(pd*
(N))
- 9" {N)]\
them
for
-
rj),
{D)
(68)
,
[0,
1)
Vj {D),
5'
C
i^
is
{4)9*
(D) -
9^*
,
>
all
Pi, player
p G
finite,
[0,1),
9'^{D) when
s
satisfied for all
Pareto optimal
i
p (N) G
SPE
involves
SPEs
and p{D)
(0, 1).
£'
,
p (U) = p [N) G
= p{N)
levels 9 (TV)
9* (D) are feeisible as
(TV)
and
9
-
9'^
{N))\
(69)
.
-
rj)
and
~P
and we have that
—
/?
with'
[0, 1)
<
proving the
(0, 1),
>
1,
and lim^^i {Vf (N) /Vf (N)) >
proof,
and
we only have
To obtain a
to
9" {N) >
such that for
1
9'^
(N) and 9" (D) >
> ^ G
all /3
>
such that for P
first
any distribution of
for
[0, 1),
(67)
P, the symmetric
part of the proposition.
any combination of {0"
(A'^)
,0'*
(D)}^ ^
show that there cannot
exist
will
from
9* (TV)
shown above, and an SPE with
and 9* {D). However, as P -*
{9^* (A'^)} ^c
and
{0'*
(D)}
(D) can be constructed, showing that no other Pareto optimal
p-^i.m
61
=
any Pareto optimal
contradiction, suppose that there exists another
9 [D) different
1,
does not wish to deviate from the SPE.
MPE
G
1
MPE,
G (0,1), thus we have SPEs that are Pareto optimal with p{D)
that do not have this feature.
dominates 9
some
part, first note that as
To complete the
with contribution
there exist
i
G 8. This establishes that there exists P
To prove the second
satisfy (67)
6
i
{D)
F {4>9^ (N)) = F (06*^ {D)
>
rj)
(D) /Vf (D)) >
{V,''
{9" {N),9'* (D)}.^^ and any punishment
for all
{9'P
(D)] (and associated appropriate punishment
contributions
= D
/3
and by implication, than Vf {D) and Vf [N). Moreover, the
[9" [N)
such that for
(^^ +
comparison of (68) and (69) shows that Vf {D) and Vf (N)
we have that hm/j^i
and since
ina,Xi^£' Pi
= F
{N))
e'P {N), the
increase faster in
and
from
we only need
so
also
V, (AT)
is
(67)
1,
SPE
9* (TV)
„
that Pareto
SPE
can exist as
References
10
Abreu, Dilip (1988) "On
rica, 56,
Games with
the Theory of Repeated
Discounting" Economet-
383-396.
Acemoglu, Daron, Simon Johnson and James A. Robinson (2001) "The
nial Origins of
Comparative Development:
Review, December, 91,
5,
An
Colo-
Empirical Investigation," American Economic
1369-1401.
Acemoglu, Daron, Simon Johnson and James A. Robinson (2002)
Fortune: Geography and Institutions in the Making of the
"Reversal of
Modern World Income Distribution,"
Quarterly Journal of Economics, 118, 1231-1294.
Acemoglu, Daron, Simon Johnson and James A. Robinson (2005a) "The
Rise of
Europe: Atlantic Trade, Institutional Change and Growth, American Economic Review, 95,
546-579.
Acemoglu, Deiron, Simon Johnson and James A. Robinson (2005b)
as a
Fundamental Cause
of Development," in
"Institutions
Handbook of Economic Growth, edited by Philippe
Aghion and Steven Durlauf, Amsterdam; North-Holland.
Acemoglu, Daron and James A. Robinson (2000) "Why
Franchise? Democracy, Inequality and
Growth
West Extend the
did the
in Historical Perspective,"
Quarterly Journal
of Economics, 115, 1167-1199.
Acemoglu, Daron and James A. Robinson (2001) "A Theory
tions,"
American Economic Review,
91, 938-963.
Acemoglu, Daron and James A. Robinson (2006a) Economic
and Democracy,
New
of Political Transi-
Origins of Dictatorship
York; Cambridge University Press.
Acemoglu, Daron and James A. Robinson (2006b) "A Model
of the Iron
Law
of
Oligarchy," in preparation.
Alston, Lee
American
and Joseph
J.
welfare state:
P. Ferrie (1999) Southern paternalism and the rise of the
economics,
politics,
and
institutions in the south,
1S65-1965,
New
York; Cambridge University Press.
Alston, Lee
J.
and Kyle Kauffman (2001) "Were Postbellum
Markets Competitive?" Explorations in Economic History,
Arthur,
W.
Brian (1989) "Competing
Historical Events,"
Economic Journal,
Austen-Smith, David (1987)
bilistic
Agricultural Labor
38, 181-194.
Technologies, Increasing Returns and Lock-In by
99, 106-131.
"Interest Groups,
Voting," Public Choice, 54, 123-139.
62
Campaign Contributions and Proba-
P. (1994) "Electoral Competition with Informed and Uninformed Voters,"
Baron, David
American
Political Science Review, 88, 33-47.
Barro, Robert (1997) Determinants
Study,
MIT
University Press, Cambridge
Barro, Robert
Economy,
litical
J. aind
of
Economic Growth: A Cross-country Empirical
MA.
Xavier Sala-i-Martin (1992) "Convergence," Journal
of
Po-
100, 223-251.
Bateman, Fred and Thomas Weiss (1981) A
trialization in the slave
economy, Chapel
deplorable scarcity: the failure of indus-
University of North Carolina Press.
Hill;
Bednar, Jenna and Scott Page (2006)
"Culture, Institutional Performance
and Path
Dependence" University of Michigan, mimeo.
Benabou, Roland (2000) "Unequal
tract,"
American Economic Review,
Benabou, Roland (2005)
Societies:
Income Distribution and the Social Con-
90, 96-129.
"Inequahty, Technology and the Social Contract" in the Hand-
book of Economic Growth, edited by Phihppe Aghion and Steven Durlauf, Amsterdam; North-
Holland.
Besley, Timothy, Torsten Persson
and Daniel Sturm (2005)
"Political
Competition
and Economic Performance: Theory and Evidence from the United States," London School of
Economics, mimeo.
Billings,
Dwight B. (1979)
and development
in
Planters and the making of a "new South":
class, politics,
North Carolina, 1865-1900, University of North Carolina Press.
Bushnell, David (1971) "Voter Participation
in the
Colombian Election of 1856," His-
panic American Historical Review, 51, 237-249.
Bushnell, David (1993) The Making of Modern Colombia:
A
Nation in Spite of
Itself,
Berkeley; University of California Press.
Chaves,
ing:
IsEa'as,
Leopoldo Fergusson and James A. Robinson (2005)
"Ballot Stuff-
Theory and Evidence from the 1922 Colombian Presidential Election," Unpublished,
Department
of
Government, Harvard University.
Chilcote, Ronald H. (1990) Power and the Ruling Classes in Northeast Brazil,
New
York; Cambridge University Press.
Chong, Dennis (1991)
Collective action
and
the civil rights
movement, Chicago; Uni-
versity of Chicago Press.
Christie, Keith H. (1978) "Antioqeho Colonization in Western Colombia:
praisal," Hispanic
American Historical Review,
A
Reap-
58, 260-283.
Glower, Robert W., George H. Dalton, Mitchell Hcirwitz and Alan A. Walters
63
(1966) Growth without development; an economic survey of
Liberia, Evanston;
Northwestern
University Press.
Coate, Stephen and Stephen E. Morris (1999) "Policy
nomic Review,
Persistence,"
American Eco-
89, 1327-1336.
Coatsworth, John H. (1974)
Porfiriato," Hispanic
"Railroads, Landholding and Agrarian Protest in the Early
American Historical Review,
Cobb, James C. (1984)
Industrialization
54, 48-71.
and southern
society, 1877-1984, Lexington;
University Press of Kentucky.
Colmenares, German (1968) Partidos
Universidad de
los
politicos y clases sociales, Bogota;
Ediciones
Andes.
Dalton, George H. (1965)
"History, Politics
and Economic Development
in Liberia,"
Journal of Economic History, 25, 569-59L
David, Paul (1985)
ciation Papers
"Clio
and the Economics of
and Proceedings, volume
American Economic Asso-
75, 332-337.
Delpar, Helen (1981) Red against
1899, Baton Rouge; University of
QWERTY"
blue:
Alabama
Dix, Robert H. (1967) Colombia:
the Liberal Party in
Colombian
politics,
1863-
Press.
the political
dimensions of change.
New
Haven; Yale
University Press.
Dixit,
litical
Avinash (1989a) "Entry and
Economy,
Dixit,
Exit Decisions under Uncertainty" Journal of Po-
97, 620-638.
Avinash (1989a)
Import Penetration and Exchange-Rate Pass-Through"
"Hj'steresis,
Quarterly Journal of Economics, 104, 205-228.
Du Bois, W.E.B.
(1903) The Souls of Black
Folk,
New York;
Easterlin, Richard A. (1960) "Interregional Differences
lation,
and Total Income, 1840-1950,"
in
in
Trends in the American
A.C. McClurg
&
Company.
Per Capita Income, Popu-
Economy
in the Nineteenth
Century, Studies in Income and Wealth, vol. 24, Princeton University Press, 73-140.
Esteban, Joan M. and Debraj
Ray
(2001) "Collective Action and the Group Size
Paradox," American Political Science Review 95, 663-672.
Engerman, Stanley
tutions,
How
and
and Kenneth
L. Sokoloff (1997) "Factor Endowments, Insti-
Growth Paths among New World Economies,"
Latin Am.erica Fell Behind, Stanford University Press, Stanford
Fogel,
of
Differential
L.
Robert
American Negro
W. and
Stanley L.
Engerman
slavery, Boston; Little,
Brown.
64
in
Stephen Haber
ed.
CA.
(1974) Time on the
cross; the
economics
Foner, Eric (1989) Reconstruction: America's unfinished
revolution,
New
1S63-1S77,
York; Perennial Library.
Merran (1964)
Fraenkel,
Tribe
and Class in Monrovia, London; Oxford University
Genovese, Eugene D. (1965) The
&
society of the slave South,
New
political
economy of
Press.
economy
slavery: studies in the
York; Pantheon Books.
(1948) The constitutions of Colombia, Durham; Duke University
Gibson, William M.
Press.
Goldin, Claudia D. (1976) Urban slavery
tative history, Chicago; University of
in the
American South, 1820-1860:
Chicago Press.
Grossman, Gene M. and Elhanan Helpman (1996)
Special Interest Politics," Review of
Hagopian, Frances (1996)
Cambridge University
a quanti-
Economic
"Electoral Competition and
Studies, 63, 265-286.
Traditional Politics
and Regime Change
in Brazil,
New York;
Press.
Haxtlyn, Jonathan (1988) The
politics of coalition rule in
Colombia,
New
York;
Cam-
bridge University Press.
Hassler John, Sevi Mora, Kjandetil Storlesseten and Fabrizio Zilibotti (2003)
"Survival of the Welfare State" American
Economic Review,
93, 87-112.
Heinicke, Craig (1994) "African-American Migration and Mechanized Cotton Harvesting, 1950-1960," Explorations in
Huffman, Frank
Economic History,
(1974) Old South,
J.
New
31, 501-520.
South: continuity and change in a Georgia
county, 1850-1880, Unpublished Ph.D. Thesis, Yale University.
Jack, William and
of Economics,
Roger Lagunoff (2006) "Dynamic
Enfranchisement," Department
Georgetown University, http://www.georgetown.edu/faculty/lagunoff/franchlO.pdf
Key, V.O.
Jr.
(1949) Southern
Politics: In State
and Nation,
New
York; Vintage Books.
Klein, Herbert S. (1992) Bolivia: The Evolution of a Multi-Ethnic Society,
New
York;
Oxford University Press.
Kline,
Harvey
F. (1999) State building and conflict resolution in Colombia, 1986-1994,
Tuscaloosa; University of
Kousser,
J.
Alabama
Morgan
Press.
(1974) The shaping of Southern
the establishment of the one-party South, 1880-1910,
Krugman, Paul
New
politics:
suffrage restriction
and
Haven; Yale University Press.
(1991) "History versus Expectations," Quarterly Journal of Economics,
106, 651-667.
Lagunoff, Roger (2006) "Dynamic
town
University,
Stability
mimeo.
65
and reform
of Political Institutions,"
George-
Margo, Robert A. (1990) Race and
history, Chicago; University of
schooling in the South, 1880-1950:
an economic
Chicago Press.
Matsuyama, Kiminori (1991)
"Increasing Returns, Industriahzation, and Indetermi-
nacy of EquiHbrium," Quarterly Journal of Economics, 106, 617-650.
Mazzuca, Sebastian and James A. Robinson (2004)
sharing in the Origins of
"Political Conflict
and Power-
Modern Colombia," Unpublished.
McAdam, Douglas
(1983) "Tactical Innovation and the Pace of Insurgency," American
Sociological Review, 48, 735-754.
McCreery, David
cial
(1986) "An Odious Feudalism: Mandamiento Labor and Commer-
J.
Agriculture in Guatemala, 1858-1920," Latin American Perspectives, 13, 99-117.
Michels, Robert.
Political Parties.
McCreery, David
J.
New York NY; The
Free Press.
(1994) Rural Guatemala, 1760-1940, Stanford; Stanford University
Press.
McGreevey, William
Cambridge University
P. (1971)
An
economic history of Colombia 1845-1930,
Press.
Milgrom, Paul and John Roberts (1994) "Comparing
nomic Review,
New York;
Equilibria,"
American Eco-
84, 441-459.
Mills, C. Wright. (1956) The power
elite,
New
York; Oxford University Press.
Mulligan, Casey, Richard Gil and Xavier Sala-i-Martin (2004) "Do Democracies
Have
Different Public Policies than Nondemocracies" Journal of
Economic
Perspectives, 18,
51-74.
Mulligan, Ccisey and Kevin Tsui (2005) "Causes and Consequences
petitiveness" University of Chicago,
Mosca, Gaetano (1939) The
North, Douglass C. (1990)
New
of Political
Com-
mimeo.
ruling class.
New
York; McGraw-Hill.
Institutions, institutional change
and economic performance,
York; Cambridge University Press.
North, Douglass C. and Barry R. Weingast (1989) "Constitutions and Commitment:
Evolution of Institutions Governing Public Choice
in
Seventeenth Century England, Journal
of Economic History, 49, 803-832.
Northrup, David (1995) Indentured
labor in the age of imperialism, 1834-1922,
New
York; Cambridge University Press.
Olson,
Mancur
C. (1965) The Logic of
Collective Action; public goods
and
the theory of
groups, Cambridge; Harvard University Press.
Page, Scott (2006) "Path Dependence" Quarterly Journal
66
of Political Science,
1,
87-115.
Paige, JefFery
M.
(1997) Coffee and power: revolution and the
rise of
democracy in
Central America, Cambridge; Harvard University Press.
and Ruth Monge Guillergua (2003) Las
Palacios, Dario
1823-1993, Lima; Editorial
"FECAT".
Pareto, Vilfredo (1968) The
Pierson, Paul (2004)
constituciones del Peril,
rise
Politics in
and
fall of the elites,
New
York; Arno Press.
Time: History, Institutions and Social Analysis, Prince-
ton; Princeton University Press.
Ransom, Roger
and Richard Sutch (2001) One Kind
L.
consequences of emancipation, 2nd Edition,
Safford,
America:
Frank
New
New
of Freedom: the economic
York; Cambridge University Press.
P. (1972) "Social Aspects of Politics in Nineteenth-Century Spanish
Granada, 1825-1850," Journal of Social History,
Samuelson, Paul (1947) The Foundations
of
5.
Economic Analysis, Cambridge, MA; Har-
vard University Press.
Sawyer,
Amos
(1992) The emergence of autocracy in Liberia: tragedy and challenge,
San Francisco; ICS Press.
Schmookler, Jacob (1972) "Time
Series of Patents Classified
1837-1957," in Zvi Griliches and Leonid Hurwicz eds.
by Industry, United States
Patents, Inventions and
Economic
Change: Data and Selected Essays, Cambridge; Harvard University Press.
Schulman, Bruce
Development and
J.
(1994) From Cotton Belt
Federal Policy, Economic
to Sunbelt:
the Transformation of the South, 1938-1980,
Durham; Duke University
Press.
Simon, Carl and Lawrence Blume (1994) Mathematics for Economists, W.W. Norton
Co.,
New
York.
Smith, Peter H. (1979) Labyrinths of Power:
Political
Recruitment in Twentieth-
Century Mexico, Princeton; Princeton University Press.
Solberg, Carl E. (1969) "A Discriminatory Frontier Land Policy: Chile, 1870-1914,"
The Americas,
26, 115-133.
Stampp, Kenneth M. (1965) The
era of reconstruction, 1865-1877,
New
York; Vintage
Books.
Steinmo, Sven, Kathleen A. Thelen, and Frank Longstreth eds.
turing politics: historical institutionalism in comparative analysis,
New
(1992) Struc-
York; Cambridge Uni-
versity Press.
Stone, Samuel Z. (1975) La dinastia de
los conquistadores:
la crisis del
Costa Rica contempordnea, Costa Rica; Editorial Universitaria Centroamericana.
67
poder en
la
stone, Samuel Z. (1990) The heritage of
America from
Germany, Britain,
the United States,
Nebraska Press.
institutions evolve: the political
and Japan,
New
economy of
skills in
York; Cambridge University Press.
Davide and Andrea Vindigni (2005) "Endogenous
Ticchi,
ruling classes in Central
the Conquest to the Sandinistas, Lincoln; University of
Thelen, Kathleen A. (2004) How
University,
the conquistadors:
Constitutions," Princeton
mimeo.
Hugh
Tinker,
1830-1920,
New
A new
(1974)
system of slavery; the export of Indian labour overseas,
York; Oxford University Press.
Trigo, Giro Felix (1958) Las constituciones de Bolivia, Madrid; Instituto de Estudios
Polfticos.
Wade, Robert (1988)
New
South India,
economic conditions for
Village republics:
collective action in
York; Cambridge University Press.
Whatley, Warren C. (1985) "A History
of
Mechanization
in the
Cotton South: The
Institutional Hypothesis," Quarterly Journal of Economics, 100, 1191-1215.
Wilde, Alexander (1978) "Colombia: Conversations amongst Gentlemen,"
down of democratic
regimes. Latin America, edited by
Juan
J.
in
The Break-
Linz and Alfred Stepan, Balti-
more; Johns Hopkins University Press.
Woodward,
C.
Vann
(1951) Origins of
the
new
South, 1877-1913, Baton Rouge; Louisiana
State University Press.
Woodward,
C.
Vann
(1955) The Strange Career of Jim Crow,
New
York; Oxford Uni-
versity Press.
Wright, Gavin (1978) The
political
and wealth in the nineteenth century,
Wright, Gavin (1986) Old
South,
Wright, Gavin (1999) "The
Economic History,
New
economy of
York;
New
the cotton South: households, markets,
W.W. Norton &
South,
New
Co..
York; Basic Books.
Civil Rights Revolution as
Economic
History," Journal of
59, 267-289.
Young, H. Peyton (1998)
Individual Strategy and Social Structure:
an Evolutionary
Theory of Institutions, Princeton; Princeton University Press.
Zeitlin,
Maurice (1984) The
civil
wars in
Chile, or,
The bourgeois revolutions that never
were, Princeton; Princeton University Press.
Zeitlin,
Maurice and Richard E. Ratcliff (1988) Landlords &
nant class of Chile, Princeton; Princeton University Press.
68
capitalists: the
domi-
i
o
d
:3
PQ
u
6fl
C
o
o o o
in o o o
00 o o. o
T^
o' irT CN
CO m (N
o
o o
o O
o
o,
^
o
o o
o
o o o
o
o
o" o
xf
o
^
v-T
o
o"
IT)
o
o
o
o
o. o
o
o O
O
o
o o
o
CO
T—
st-"
IT)
<N
Si
c
o
j2
3Cu m
^
ft.
o
o
o
o
o
o
o o
o o
o.
o. o o o o
o
irT o" o" oo"
T—
^ (N rO o
^ m
(N
o
o
00
^
o
O
o
O
o, O
^ O
§8
ro'
^-'
(N
Cn"
Csi"
o"
•o"
LT)
-^|-
<N
LO
1
IN
I^
c
U
CO
<N
oooooooo
oooooooo
o
o
o o
o O
o
o o
o
o
o o
o
o o
o o
o o
o_ o o o^ o_ o_ o, o_ o_ o_ o_ o_
o
o o
irT r4"
un Lo (N o
o" CnT o" o' o" o" o" o"
00 t^
r^ SO VD o vo OLOl-OLnLOLOLOLr) LO
__
o o
o
o o
o
o o
q_ o_ o.
o
cnT in IN
o
o o
o
u
^
^
r-'"
OOOOOOOO
oooooooo
o_^
o__
o^ o^
oo"
oo"
ltT
in
o__
o^
o__
oS
cnT
^^
m
IN
o__
^
I--
o
C
3
o
o
O O
bO txO tuO
C C C
U
U
W
o
o
o
o
CflCCCfiG
o
G
O
C
G
(L)
V-i
jj
V-(
rt
c3
rt
rt
v_(
G
c^
SS§:i:^S^^SoS3:o
c"3
3:
o
G
C^
U
c
K
«
O
JJ
xxx^xo^^
13
c/)
U
o
a-
13
o
00
c
o
u
o
H
S
I?
(U
u
a
C/D
u
V
o.
(U
j:i
H
(U
o
CQ
-
w
-a
:3
u
rt
z
J
cu
H
X
u
G
O
i2
G
u
5P G
G o
i)
e
TA
J3
OJ
O
lU
a!
O
u
C
3
Q
^
rt
jy
— U
_G
.
g'
</."
O
G
S^
~0
U
G
—
•
u
S
1)
S
G
G
o
^
u
u
=i
O
"a,
<
bS
-
c?
^
£:-
^
2
:^
S
^
.
^
=^.
^
u
y
^
^.
'^^
G
u
u
u
"-•
(1.
o
M
-i-i
V-i
u
_
V-l
O
<
01
O W
-cm
01
LU
•
CD
,'^
1—
cr
™ 2"
n .
,.
^
" >
n " m
£
O
U
0)
D)
3
0"
n
0)
3
2
-o
ra
c
D)
E
n
c "
re
Jua
iS
^
.
ID
~
Ana
obe
rez,
.
(T
!r,
:9J
.
0)
cas
in
rria
0)
>, >,
Gut
eba
-
n
arci
eba
D D G
Sa Sa
sa
oza
oza
«
TO
8 S
3
O
(0
D.
(0
-0
E E E
(0
o w mw w M w
.
J
N
TO
i^
tf)
CDU
^^
nj
E
o
V,
~i
^
10
d<
3
E
ra
c
r
^ a
b
i:
"1
a
"D
§2
,S
'^ 'd
D) O)
«
CD
1
'l,
CD
< < < < <
m
c
2
(31
1J
E
CO
-,
£1 W
_-
5 5
ro
o
-ro
O
?-
6?
N N
<
ID
O
W
(/)
N
C
'2
0)
TO
C 3
-TO
-"^
a
c
u-
(U
Q
(C
N-
E m
5
N >
^
Ql
S:
(!!
O)
c
.
:
0)
UJ LU UJ LL u.
(5
O
as"
c c
!i
TO
3
00 00
E E
_•
p
^
Q
£ N
m
-0
c m W
D
-D
CD
ra
^0
c
o
.y N 2
_ ^
2 £
5
0)
E
S25 2 00<
0)
10
6
0^
u.
J,:
-1
E
^^
P
-5
N N
'D
> > £
2 t
oo
(13
0)
^
to
c
.J)
,
ai
N
(0
P
•K
N
"0
o
o
I
u
.§2
•
oza
E
CD
CD
ni
0!
sa
0)03
0)
rzano
ra
0)
d:
'01
D
Q. •0
1.
-1
(y -)
c
D
3
.
(1!
15
3 I
V)
T3
(C
An Lu
n
3
•n
"C
0)
"i
S
0 K
<"
S
D)
<
0y
:
.9
I
Q. q:
CO
LO
-
--
-I
^i?
w
u
yi
Oh
p
-ii ii
?-
.=
ni
2
m
i5
-0)
oil
WH h D > >
X!
ID
N
CO
N
>-
0)
O
i-rocosoa)o^cM-3-iDroNT-nNiDO)ON!Min^ooococoi/)r)iDino^c7i'j-c\iir)iD
rgcNjcMninDn'j'3-^n'jcNn\ffCNv-cM^-ni-T-co-3-co>-rsj>-T-n'»CN(Mnc\i^t
-^3
Oh
-M
<
ON
ON
UJ
'
u
h
>
1
l/D
s
c
o
-M
V-l
fl.
d
u
(tJ
<
u.
J
<
u
O
CO
QJ
V)
I-
z
UJ
Q
«
5-1
-1-1
U
in
"—I
0.
o
a
o
o
v-i
U
<u
N
(U
N
>
o<
I
•^
I)
>.
ro
O
v^
-
O
o
>.
05
C/3
§
S
ro
2
ro
en
^
o
o
-Q
CD
CO
-I
H
CD
CO
TO
en
03
S
o
i^
CD
c"
m
j3
o
.
n
>,
m
m
N
SOU
a
o
O
U
CI)
L
0)
cn
n *
o
c
t:
=>
-9
CN
O
CD
o
->
'^
ID
..-
^
0)
CD
o
.y
™
2
o
CD
<
-
en
O -5
O
c
E °
.
>, >,
:Hi
m
U
lU
W W
CD
CO
m
(/)
m
01
(
1
CD
CD
03
CD
0)
(D
Q Q
o
£
03
03
03
ii
"o
t^
03
CD
01
ID
*-
.
«;
,_
O
o
03
CD
N N N
M
y O O O
:o E E E
000
www
c c
N 0)
0!
en
o
o
cj
"
en
CD
CD
ll
^
o
S
61
>
N
CD
1
>1
jr
cr
%C "o
CD
-ra
^
t c
CO
0)
o
LU LU LL u.
N
=! ir
03
03
—
%
E
N
03
-03
6 ^
N O
03
5
03
cn
O
03
c
o3
2
c
u.
E-03
S^^
Q ^
o
o
.b CO
03
0^
^
cn
•2
-i
£
<
R «
o o
.2,
03
03'
Qi
cn
CD
cn
a
C33
Ow H
O O
03
-2
-•
O
W 5
I O
03 5 c 5
CD
^
E N o o
''I
cn
03
0)
ra
o
'o
03
->
i 2 E
D O CO g
t
<
< < O O
w
O
T3
W
c
-CD
03
O
-3
.
o
o
cn
"D
•O CC
D- -03
Ql
lu"
^ W W
"D "D
(U
CD
HI
"D "P
cr
o c
(D
03
o
03
I-
S
o
cn
CD
cn
_a3
133
> >
m-^cor--->a-csicDoncnmcD^a30c\icMcO'a-T-s-cococn
COi--i-CNCOT-CNCOi-(Nv-v-C\i'^CNCMCOCNCMCO-^nCNCM
aoi-~cD-^oc33
T-T-
C^J
CVJ
CO
CM
'^
^^
-1
<J>
't
CM
,^l
UJ
m
CD
a>
CM
r^ 00
^
a.
'
"
Tco
CO
CM
col
V
CO
CO
*
CO
1
CM
u;
(U
u.
N
\r>
CD
CM
Tt
"'^
tu
CO
^
Z
p
(/I
C^
1-
(
•
T—
V^
<u
a
<
i—\
c^
V(
4-»
a
(U
U
<N
(U
v^
3
W)
\^
z
m
n
1^
CM
CO
CO
CM
CM
CO
-^r-
1
r
nr
(/)
UJ
Q.
OJ
in
CM
^
p
d
03
col
C/3
PU
in
CM
N
UJ
D
^
>
m
CO
o
a
^
a
Conf.
H
g
Gen.
of
Ambassador
c
.5J
Labor
President,
g
UK
to
rO
OJ
(U
fin
t/1
Oh
f^
o
r—
1-J
11
V-l
C/5
G
'o
Oh
Oh
C3
r-"
CJ
^
U
1/3
3
o
o
13
U
VO
CN
rt
•13
(U
;:=]
O
(U
rj
V-(
iS
^
O
^J
II
W-t
d
o
•^
'^
^
13
o
u
CO
en
<u
U
.
.
<u
Kn
u
o
OJ
u
^
O
V-l
OJ
Date Due
Lib-26-67
MIT LIBRARIES
3 9080 02617 9694