Digitized by the Internet Archive
in
2011 with funding from
Boston Library Consortium Member Libraries
http://www.archive.org/details/theoryofpoliticaOOacem
331
IHr )
1415
working paper
department
of
economics
A Theory of Political Transitions
Daron Acemoglu
James A. Robinson
massachusetts
institute of
technology
50 memorial drive
Cambridge, mass. 02139
WORKING PAPER
DEPARTMENT
OF ECONOMICS
A Theory of Political Transitions
Daron Acemoglu
James A. Robinson
No. 99-26
October 1999
MASSACHUSETTS
INSTITUTE OF
TECHNOLOGY
50 MEMORIAL DRIVE
CAMBRIDGE, MASS. 02142
A
Theory of
Political Transitions
James A. Robinson*
Daron Acemoglu^
July, 1999.
Abstract
We develop a theory of political transitions inspired in part by the experiences
Western Europe and Latin America. Nondemocratic societies are controlled by a
rich elite. The initially disenfranchised poor can contest power by threatening social
unrest or revolution, and this may force the elite to democratize. Democracy may
not consolidate because it is more redistributive than a nondemocratic regime, and
this gives the elite an incentive to mount a coup. Because inequality makes democof
racy more costly for the
elite,
democracy and may end up
highly unequal societies are less likely to consolidate
between regimes or in a nondemocratic repressive regime. An unequal society is likely to experience fiscal volatility, but the
relationship between inequality and redistribution is nonmonotonic; societies with
intermediate levels of inequality consolidate democracy and redistribute more than
oscillating
both very equal and very unequal countries. We also show that asset redistribution,
such as educational and land reform, may be used to consolidate both democratic
and nondemocratic regimes.
Keywords: democracy, dictatorship, inequality, political instability, redistribution.
JEL
Classification: D72, D74, 015, P16.
thank Abhijit Banerjee, Francois Bourguignon, Ruth Collier, Steve Durlauf, Jeff
Dani Rodrik, Ken Sokoloff, Mariano Tommasi, Jaume Ventura and seminar
participants at MIT, Wisconsin, NYU, Western Ontario, Toulouse, NBER Summer Institute, Berkeley
and Yale Political Science Departments, the conference on "Asset Inequality and Poverty" at the Ministry
of Land Reform in Brasilia, and LACEA 98 at the Universidad di Telia, for useful suggestions.
t
02319;
Massachusetts Institute of Technology, Department of Economics, E52-371, Cambridge,
e-mail: daronOmit edu
^University of California, Berkeley, Department of Political Science and Hoover Institute, Stanford
University; e-mail: jarobins@usc.edu.
*We would
like to
Frieden, Michael Kremer,
MA
.
1
"MASSACHUSETTS INSTITUTE
OF TECHNOLOGY
Introduction
1
Most Northern European countries extended the
franchise during the late 19th
and early
20th centuries, and succeeded in consolidating mass democracy. For example, in Britain,
following the
first
tentative reforms of 1832, voting rights were significantly extended in
1867 and in 1884. They were further expanded in 1919, when universal male suffrage was
introduced, and in 1928,
when
all
this process of democratization.
women were
allowed to vote. There were no reversals in
Although many
less
developed countries, most notably
those in Latin America, also became democratic during the late 19th and early 20th
1
century, they quickly reverted to nondemocratic regimes.
Latin American countries
is
therefore marred
by
In Argentina, for example, universal male suffrage
soon overthrown by a coup in 1930.
coup
1983.
The
oscillations in
became
Democracy was
why has
it
has mass democracy been durable in
been so hard to consolidate
and out
But
re-instated in 1946, but
and
many
of democracy.
effective in 1912.
in 1955, re-created again in 1973, subverted again in 1976,
Why
recent history of
it
fell
was
to a
finally re-installed in
many Northern European
countries,
this set of political institutions in less
and
developed
countries such as those in Latin America?
This paper provides a framework for analyzing this question.
in
We
emphasize that
democratic societies the poor impose higher taxes on the rich than in nondemocratic
societies.
This makes the poor pro-democratic while simultaneously giving the rich an
incentive to oppose democracy.
political
2
In nondemocratic societies, the poor are excluded from
power, but shocks, for example wars or depressions,
threat (or at least costly social unrest).
3
The
may
create a revolutionary
rich (elite) will try to prevent revolution
before the mass democratization of the 19th century, Britain had
elections with a very restricted
nondemocratic regimes have often been military dictatorships. We do not
distinguish between these different types of non-democratic regimes. We also define any significant move
towards mass democracy as "democratization".
2
For example, Rodrik (1999) shows that democracies tend to have higher wages and higher labor share.
In the context of Latin America, there are many examples of military coups specifically aimed at reducing
redistribution (see Skidmore, 1967, Smith, 1978, Stepan 1978, and Wallerstein, 1980). For example, in
Argentina, the labor share, which increased rapidly during the democratic regime of 1946-49, fell during
There were similar increases in
the following dictatorship (e.g., Di Telia and Dornbusch, 1989, p. 6).
inequality in Brazil after 1964 when the left-wing President Goulart was deposed, and in Chile, the coup
of 1973 was aimed in part at stopping and reversing the increasingly redistributive policies of Allende.
3
Haggard and Kaufman (1995) document the importance of shocks and crises in regime changes.
In Acemoglu and Robinson (1997), we provide evidence on the importance of wars and depressions in
triggering social unrest from 19th century Europe.
franchise, while in Argentina,
by making concessions to the poor,
for
is
guarantee future redistribution.
If this
credible
will
it
commitment
rights achieves
form of income redistribution.
in the
often only transitory, current redistribution does not
Because the threat of revolution
a revolution, as
example
temporary redistribution
be in a very unequal
This
to future income redistribution.
by changing the identity of the future median
The poor would
mount a coup.
prevent
like to
commit
They
to retake power.
relatively costly for
are
more
them. Taxes
likely to
will
a highly unequal society
political instability is
is
more
voter.
elite
may have an
opportunity
may
credible, the elite
be high in turn when inequality
likely to fluctuate in
likely
what extending voting
prefer
do so when, due to high taxes, democracy
democracy by making coups more attractive
destabilizes
is
be forced to make a
to low levels of future taxation to
But since such commitments may not be
this.
insufficient to prevent
society, the elite will
Democracies are not necessarily permanent because the
to
is
is
is
high. So inequality
for the rich elite.
and out of democracy.
4
As a
More
result,
generally,
when the stakes, economic gains from controlling political
power, are greater, a view that can be traced to James Madison (1788).
In consolidated democracies, such as the
OECD
economies, the threat of coups
is
not important, so taxes are determined by the usual trade-off for the median voter be-
tween transfers and deadweight
redistribution.
losses.
There
is little
or no variability in the
In contrast, in highly unequal economies,
fiscal
policy
cause as a society fluctuates between different regimes, the amount of
changes.
5
Even
if
often limit taxes.
4
more
more
More
of
volatile, be-
fiscal redistribution
an unequal society prevents regime changes, the threat of coups
generally, while greater inequality in a consolidated
increases redistribution (e.g. Meltzer
to be in the
is
amount
and Richards, 1981), an unequal society
will
democracy
is less likely
redistributive fully-consolidated democratic regime. 6
may
account for why many Latin Amerhad difficulty in consolidating democracy, while the more equal European countries and Costa Rica have had more stable democracies. It is
also in line with the correlation between inequality and political instability documented by Muller and
Seligson (1987) and Alesina and Perotti (1996).
°This accords with some of the patterns observed in the data. For example, Gavin and Perotti (1997)
show that fiscal policy in Latin America is much more variable than in Europe. They report that the
standard deviation of the growth of transfers in Latin America between 1970 and 1995 was 22.4% as
compared to 4.4% for industrialized countries.
6
This may explain the lack of a well-defined monotonic relationship between inequality and redistribution in the data (e.g. Perotti, 1996).
Our model is also consistent with the consensus view that
"populist" policies are a result of high levels of inequality, and suggests an explanation for why these policies are often reversed radically by regime changes (e.g., Sachs, 1990): populist policies arise as a method
of radical redistribution when the poor majority realize that democracy is not consolidated. Countries
that consolidate democracy, such as Costa Rica and Colombia, are therefore less likely to follow populist
policies. For example, Kaufman and Stallings (1991, p. 27) write "established democracies (Venezuela,
The connection between
inequality and political transitions
ican countries, such as Argentina, Brazil, Peru, and Uruguay, have
Our
basic analysis suggests that high levels of inequality induce frequent regime
When,
changes.
in addition to
this conclusion needs to
racy costly for the
elite,
making concessions, the
be modified. Very high
elite
can use a repression strategy,
levels of inequality,
may encourage them to use repression to
which make democ-
prevent democratization.
Therefore, low levels of inequality are necessary for a democracy to consolidate, and high
inequality
to lead to
is likely
social unrest suppressed
some form
of political instability (either regime changes or
by repression).
Since the incentives to engage in or avoid
by underlying asset
fiscal redistribution,
which are generated
inequality, are a key factor in shaping regime dynamics, there
important role for asset redistribution,
e.g.
educational and land reforms.
is
an
Educational
reforms that increase the earnings capacity of the poor and land reforms that achieve
a more egalitarian distribution of assets
difficult to reverse,
less costly for
however, as their anticipation
Brazil in 1964,
and Chile
Although the reasons
may make
the
elite.
7
fiscal redistribution
There
is
by the median
a danger in radical reforms,
a coup more likely as in Guatemala in 1954,
in 1973.
for the
ferent social groups appears to
(1997),
consolidate democracy because they are
reduce the preferred amount of
and make democracy
voter,
may
changes in regimes are numerous, conflict between
be important
we presented evidence suggesting that
in practice.
In
dif-
Acemoglu and Robinson
in Britain, France,
Germany and Sweden
democratization was in large part a response to the threat of revolution and social unrest.
In Latin America,
same
factors.
many
instances of democratization appear to have been driven by the
8
Our paper
is
related to the analyses of the political
economy
of redistribution
(e.g.,
Meltzer and Richards, 1981, Persson and Tabellini, 1994, and Alesina and Rodrik, 1994,
Benabou, 1999) and to models of
Tornell and Velasco, 1992,
There
is
social conflict (e.g.,
Roemer, 1985, Grossman, 1991,
Ades and Verdier, 1995, and Benhabib and Rustichini, 1995).
a large political science literature on democratization, starting with the work of
our study) were also associated with orthodox macro
policies... It was the
and Brazil) that followed populist policies."
7
This may help to explain why democracy was consolidated in Costa Rica after 1948 and Venezuela
after 1958, two countries that instituted land and educational reforms following democratization.
8
Collier (1998) has recently argued that social pressure from the masses was the driving force behind
the most recent democratizations in Peru, Uruguay, and Brazil. Similarly, in Argentina both the 1912
reforms, which institutionalized universal male suffrage, and the withdrawal of the military in 1973, which
reinstated Peron, appear to have been a response to increasing social unrest (see Rock, 1987, Ch. 8 and
circa p. 188). In Venezuela, the democratizations in 1945 and 1958 both happened in the context of a
widespread uprising (see, e.g., Levine, 1989, p. 256, and Kolb, 1974, p. 175). The pattern also appears
to be quite similar in Guatemala and El Salvador (see, for example, Paige, 1997).
Colombia and Costa Rica
in
transitional democracies (Peru, Argentina
Moore
Lipset (1959) and
(1966), which emphasizes the structural determinants of
racy (such as income level and class composition).
strategic interaction
economic factors
between regimes and
More
recent
work has focused on the
and on
their opponents,
democ-
political rather
than
Rustow, 1970, O'Donnell and Schmitter, 1986, Przeworski, 1991,
(e.g.,
Therborn (1977) and Rueschemeyer, Stephens and Stephens
Linz and Stepan, 1996).
(1992) share our emphasis on the actions of the disenfranchised poor as the key to under-
we
standing democratization, and in our previous work, Acemoglu and Robinson (1997),
also
emphasized democratization as a commitment to future redistribution. The
ture on coups
is
much
less
how
developed and focuses mostly on
litera-
purely political factors
explain the persistence or collapse of democratic politics (for example, Dahl, 1971, and
This contrasts with our focus on social conflict and redistribution (though
Linz, 1978).
O'Donnell, 1973, also discusses the importance of social conflict in the context of Latin
American development
The paper proceeds
strategies).
as follows. In Section 2,
we
present our basic model of political
development and study the determinants of transitions between regimes. In Section
we
consider a simple extension of the model to discuss
to avoid democratization.
redistribution.
In Section
In Section
5,
we
discuss
specific increases political instability.
2
4,
how
we extend the
how
the elite
basic
may
model
3,
use repression
to consider asset
the possibility of making transfers group-
Section 6 concludes.
The Basic Model
There are two groups of agents: the poor and the
The
rich (the elite).
political state
be democratic or nondemocratic. In a democracy, the median voter sets the tax
because the poor are more numerous, the median voter
regime, taxes are set by the rich.
When
can attempt a revolution, and the
the system
is
democratic, the rich can
and revolutions are
also enables us to
will
the political system
model the
be determined
is
nondemocratic, the poor
decides whether to establish democracy.
mount a coup. The
may be more
and
a poor agent. In a nondemocratic
stochastic, capturing the notion that
following wars or depressions,
which
elite
is
rate,
can
costs
some
When
and benefits of coups
periods, such as those
conducive to social and political unrest. This
fact that those in
power cannot commit to future tax
in future political equilibria.
rates,
The Environment
2.1
We
economy with a continuum
consider an infinite horizon
of these agents are "poor"
,
while the remaining
— A form
1
A
of agents.
1
a rich
"elite"
.
proportion A
Throughout the
paper superscript p denotes poor agent and r denotes rich agent (or member of the
We
poor agents as
will treat all
power
Initially, political
there
is full
There
is
is
democracy, the median voter
a unique consumption good
at time
=
t
all
members
of the elite are also identical.
concentrated in the hands of the
can be thought of as physical or
economy
and
identical,
human
elite,
and a unique
We
capital or land).
we
to higher inequality. Without loss of any generality,
=
is y\
hl
for
i
is
/5
The
good
final
is
produced from
>
is
group
i
the discount factor and
will
and
from the
is
,
=
(7(0)
We
5.
a deadweight cost of
=
C'(t)
0,
also
>
is
specific,
is
>
r
for all
C'{r)
>
is
The A poor agents
1
>
1.
of the poor take part,
it
a proportion
agents
it
1
— [i >
who have
among
an agent
cl + j,
=
(1
—
r t )h
=
i
p,r,
z
+T
l
t
,
all
where
transfer that an agent of
>
C
0,
=
T =r
t
oo,
t
=T
t
.
Group-specific transfers
costly to raise taxes: at tax rate
it is
twice continuously differentiable with
is
>
and C"
i.e.
If
0.
some
r
=
of the comparative statics
To avoid keeping track
1.
which ensure an
[Xhp
+
(1
—
A)/i
there were no costs of
r
]
— C (r = rt — C(r
f)
of
The
interior tax rate.
t
We assume that if a revolution is
t
).
always succeeds.
of the capital of the
attempted and a fraction
However, in the course of the revolution,
economy
is
destroyed. After a revolution, poor
participated in the revolution take control of the output,
themselves.
for
are initially excluded from the political process but can attempt a
revolution in any period
<
J2'jLoP
lump-sum
hence T\
at the corner,
and C"(l)
government budget constraint implies
£
total output of
t+J
t
given by, y\
the
assume that
C(r t ), where
would not apply when the tax rate
we assume
E
by
taxation, our general results would not be altered, but
this case,
and
capital,
For now, we simplify the analysis by assuming that taxes
be person
transfers cannot
C"(0)
0,
a low level of 6 corresponds
the expectations operator conditional on
is
t
Post-tax income
t.
state.
be discussed in Section
Tt there
E
the tax rate on income, and T\
receives
are linear
let
= p,r.
information available at time
Tt
begin our analysis of the
also normalize the aggregate stock
All agents have identical preferences represented
where
if
h (which
parametrize inequality,
.
1.
| so that
where each poor agent has capital hp and each member of the
r
of capital, h, to
>
asset with total stock,
> hp These capital stocks are exogenous. To
h r = (1 — 9)h/{\ — A) and hp = 9h/\ where A > 6 > 0, so that
has h
elite
but A
a poor agent.
is
y,
elite).
Therefore,
when
all
fi,
and share
poor agents take part in the revolution, each
A
receives fi/X in every period, while the rich receive nothing.
that a revolution
h
and
values:
We
assume that q
(j.
fj,
implies
We assume that // is stochastic and changes between
h
= q, independent of the previous realization.
Pr(/x = \i
relatively costly.
is
two
low value of
=
\i
<
with
)
t
1/2 so that the threat of revolution
is
relatively rare.
In a democracy, the elite have no special voting power (one-person-one- vote), but they
can attempt a coup. After a coup, the
quo with the
the
<p
elite,
G
h
1
h
{ip ,<p }, <p
>
l
(p
>
and
threat, like the threat of revolution,
to
mount a coup,
If
a coup
Similarly,
if
is
we
so
let (p
h
—
\x
is
=
l
<p
)
=
for
s.
initial
ignore the free rider problem
member
each
Assume
relatively rare.
that s
Only
of the elite
<
is <p,
status
among
where
1/2 so that the coup
1
in state
if
will the rich
want
+oo.
>
mounted, then
coup
cost of a
Px(<p t
We
power.
elite controlling political
and assume that the
back to the
political situation reverts
=
1
fj,
at first so that there
is
no revolution immediately.
democratization occurs, then democracy starts with the coup cost at
h
<p
,
implying that a democracy has at least some window of opportunity before a coup can
Finally, in each
occur.
extend the franchise.
median
voter, a
The timing
nondemocratic period the
If it is
rate.
of events within a period can be
1.
the state
2.
the poor set the tax rate, r,
3.
in a
or
<p is
if
we
and the
are in a democracy,
coup takes
place, the party that
new tax
mount a coup.
rich set r, otherwise.
If
they extend the franchise
comes to power decides whether to keep the
rate,
the poor decide whether or not to initiate a revolution.
share the remaining output of the economy.
If
there
decided at 2 or 3 gets implemented.
5.
as follows.
nondemocratic regime, the rich decide whether or not to extend the franchise.
tax t set at stage 2 or set a
4.
summarized
revealed.
In a democracy, they decide whether to
or a
have to decide whether or not to
extended, then the economy becomes a democracy, and the
poor agent, sets the tax
//
elite
incomes are realized and consumption takes place.
is
If
there
is
a revolution, they
no revolution, the tax rate
Analysis
2.2
The tax
rate,
r m which maximizes the indirect
,
a coup threat,
satisfies
r
where the term
m = argmax
((1 -
in square brackets
the
is
lump-sum
)
decreasing in
when
=
6
0.
When
A, so that
h
r
=
hp
fact that
inequality
=
9/X.
is
hp we have r m
,
=
same
participate in a revolution, there
no
maximum
is
m
uniquely defined and
is
tax rate
is
also higher,
"free-rider
it
comes to whether
problem" because
if
or not to
an agent does not
take part in the revolution, he can be excluded from the resulting redistribution.
we can
treat all poor agents as one player.
preferences,
and we can
also treat
them
Also
members
all
We will characterize the pure strategy Markov Perfect
h
((p
,D),
l
(/j,
,E), or
h
(fi
D denotes democracy.
7
=
1
in state
Clearly,
is
if
7
=
,E), where
The
E
denotes
=
the state remains at (-,E), and
{.,D), the elite choose £
coup, the elite also
=
The
(i.e.
if
7
=
when
elite
who move
decision.
when the
state
The
is
Markov
is
in
one of
which
(tp
l
,D),
1, it
switches to
(,
=
0,
is
(.,
not extended).
D).
If
the state
no coup. Following a
actions of the poor consist of a decision
(.,E),
and a tax rate r d when the state
is
on the current actions of
before the poor according to the timing of events above, while in state
(.,D), the actions of the elite are conditioned
strategy
state
franchise
(.,D). In state {-,E), the actions of the poor are conditioned
the
game
power (nondemocratic regime) and
they mount a coup, or
1, i.e.
make a taxation
to initiate a revolution, p,
and the poor.
actions of the elite consist of a decision to extend the franchise
E, and a tax rate r e when 7
0,
elite in
elite
game. This economy
Equilibria of this
depend on the current state of the world.
So,
of the elite have identical
as one player in a repeated
can therefore be represented as a repeated game between the
strategies only
and
no redistribution.
and when
preferences,
is
that r
Hence, in the case of complete equality,
0.
the median voter sets a zero tax rate and there
All poor agents have the
first-order condition
(1)
(1) implies
higher, the
The
transfer, T.
^
C\r m =
where we used the
+ [t-C (t)]>
r)hP
problem gives
of this
a poor agent in the absence of
utility of
Perfect equilibrium
is
on the actions of the poor. Then, a pure
r
a strategy combination denoted {ap a }, such that
9
,
There could be a coordination problem whereby all poor agents expect others not to take part in
However, since taking part in a revolution imposes no
additional costs irrespective of whether it succeeds or not, it is a weakly dominant strategy, and we
a revolution, so do not take part themselves.
therefore ignore this coordination problem.
ap and a r
the
are best-responses to each other for
be the value of an agent of type
mounting a coup
is (p
the state
h
is (ip
choose the tax rate r
h
=
i
p,r,
when the tax
budget
is
while
will characterize
rate
l
(9) is
rm
.
= —^ <
'
J}
{
+
(ip
l
be the value of agent
)
h
l
((p
,D)
democracy and when the cost of
is
when the
i
+P[(1- s)VV, D) + sVV)]
8\e)
cost
is ip
l
(2)
.
the net amount of taxation that a person of type
= T m — rm h
l
Hence, 8
{6)
T m = T m -C(T m
balanced,
V
%
V
Let
1/2).
returns to poor and rich agents are:
.
is
p,r when there
>
,D), there are no constraints on the median voter, so he will
m The
and 6
a poor agent (since A
is
a switch to a nondemocratic regime, as a result of a coup).
VV, D) = h
where
=
i
Similarly, let
.
may be
which case there
When
We
possible states.
Perfect Equilibria by writing the appropriate Bellman equations.
Markov
In a democracy, the median voter
(in
all
,
that, 6 (9)
by the Envelope Theorem.
i
receives
and from the assumption that the
r
Note
).
%
<
<
6P (6), and
^™ = ^ >
Intuitively, higher inequality raises
the tax rate on the rich, while simultaneously increasing net transfers to the poor.
V(</?')
is
the value to agent
below r m
tax rate
i
in state
say to a level r
,
V(^, D, r d =ti + A
rd
i
< Tm
= p,r,
.
A
and
l
(9)
,
l
)
where
d
{9)
in
(ip
l
,D). In this state, the poor
s)V
l
(ip
Clearly, A*(0)
h
<
,
D)
+ sV^ip
8 P {6),
and
1
,
,
= max
[VV, E) - if +
1
(C
}
(1
-
C)
D, r
d
A r (8) >
After observing the tax rate r d the elite decide whether to
VV)
reduce the
an attempt to prevent a coup, yielding values
+p[(l-
=Td - rd h\
may
(3)
.
)]
r
8 (9) because
mount a coup, so
VV, D, r
d
(4)
))
ytf) = CV(^£) + (i-C)V(^,AA
where
recall that ^
=
where the
elite are in
Hence, for
%
=p
or
1
implies a coup.
power and there
we
r,
coup
in state
(ip
l
,
is
coup
will take the society to the state
no threat of revolution, thus they
D), (
=
(3 [(1
1, is
-
q)
TV, E) + SfVV, E)\
therefore optimal for the elite
V r (p ,E)~V r (^,D,r d )><p
l
This
is
set
1
(p,
r
,
=
E)
0.
have:
VV, E) = ti +
A
A
the coup constraint: a coup occurs
power and reducing taxation
is
if
if
l
(6)
the gain to the rich of capturing political
greater than the cost of the coup,
8
(5)
l
<p
.
We
can
long as
>
ip
determine a
first
= rm
to tax at the rate r
V
r
E)-V
1
{fj,
,
a coup
(p(9,q,s),
r
D, r m
l
(<p
,
),
never beneficial for the
is
in state
l
(<p
10
we obtain
where V{(ft, D, r m )
^ ^
Q
l
dated.
<p(6, q, s),
The tax
0)
Imposing
.
Vr
rate,
S)=
h
(fj,
= V r (tp h ,D)
,E)
in
-ps r (e)(i-q)
(7)
•
l-P(l-q)
the coup threat does not play a role, and democracy
rm
which implies that a
This
is
unequal society
less
is
fully consoli-
always determined by the usual trade-off for the median voter,
is
,
balancing transfers against the deadweight losses of taxation.
democracy.
=
given by (p(9,q,s)
the return to the rich of always
is
"+-fl
= rm
the poor continue
if
critical value is
h
=
even
rich,
this critical value as:
...
When <p >
This
,D).
remaining in democracy with a tax rate r
(5),
the cost of coup, (p(6,q,s), such that as
critical value of
more
is
Observe that
<
dg
a fully consolidated
likely to achieve
makes democracy
intuitive since a greater level of inequality
0,
less
attractive for the rich.
We
</?(#, q,
if ip
>
the poor can stop a coup by setting a low enough tax rate in the state (D,
<p )
s),
when
(or conversely,
r
l
(ti
ip
,E)-V r (<p
Combining
l
\Zb
—
rich agent
V
r
l
{<p
)
Tp(9,q,s), even a policy of setting r
=
,D,T d
and
(2)
of always remaining in
—+
<
l
Since the lowest tax rate that the poor can set
coup).
V
l
can next determine the value of the cost of coup, 7p(6,q,s), such that
as the
=V
l
{fi
,E)
and setting
(3),
democracy
maximum
-
we
<p\
s.
If
q
Similarly,
)
,
Prom
this,
),
we
define
value the median voter can credibly
(2)
and
(5)
-(38
r
r
given by
is
with
V
h
r
(fi
calculate the value
V r (ip
l
,D,T d
commit
,E)
=V
=
0)
=
to give to a
r
(cp
h
,D) and
(6)(l-s-q)
<
and
1
—
s
—
high, then elite control following a
is
will reoccur quickly.
s is high,
q
>
coup
0,
so 7p(6,q,s)
will
is
be short lived
This reduces the expected benefits
the coup constraint binds regularly, and because in
pay relatively low taxes, democracy
this state the rich
<
if
we can
l
l
Recall that S (6)
because a revolutionary threat
from a coup.
lp(8,q,s)
0,
obtain:
(8) is intuitive.
decreasing in q and
Tp(9,q,s)
V r (ip = Vr (<p D,r d
for the rich.
.
Expression
=
does not stop a
0)=lp(6,q,s).
under democracy. Now, solving
r
r
is
=
d
l
is less
costly to them.
Also
clearly,
(p(d,q,s).
E) = V r {f h D) applies along the equilibrium path. Since the economy starts with
the elite in power, we can only be in a democracy because the franchise was extended. Therefore, if a
coup takes place, along the equilibrium path, democracy will occur again in the future when fi = (i h
10
Note that
V r ((i h
,
)
.
More important
paper
for the focus of the
increases the threshold Tp(Q,
more under democracy.
>
democracy
l
lp(6,q,s), then
below r m
reducing the tax rate
ip
l
= V r ((f
is still
l
,
is
in state
D) and
(</?',
setting r
=
fp(0, q, s),
is
d
is
increasing in
6:
is
given by the equation
like to
=
prevent such an outcome, and
d
dr jdQ >
Now
rm
would stay
by the poor
,
(fx
=
0,
will
,E).
If
ip
=
The
l
<p
.
they could, they would promise lower
set
a coup, even
=
r
would receive
so the poor
0,
fi
=
1
y!
they would obtain
,
1: //*
>
at the time of the revolution matters,
constant over time (and this also implies that in the state
is
will
h
fj,
never occur).
We now
attempt to prevent
it
assume
6,
//
=
1
(J
,
the revolution threat
In case of a revolution, the rich lose everything,
at all costs.
can extend the franchise,
7=1,
They can do
is
i.e.
this in
binding.
V r (R) =
two
0.
They
will therefore
different ways.
they
First,
giving the poor their return under democracy,
Vp (cp
l
,
D).
Notice that this equation can be written as
(1-0(1-8- q))
is
when
the poor did not attempt a revolution in this
Notice that only the value of
which ensures that when
11
not prevent a coup.
the per-period return from revolution for the infinite future discounted
a revolution
Assumption
ti.
so higher inequality
0,
elites, realizing this, prefer
power forever and
in
hence the per-period return
fJ
if
In contrast, with a revolution in state
equal to j^s.
to the present.
=
d
and once the threat of coup disappears, the
Forward-looking
.
h
consider the state
VP (R) = x mq\
H
(9)
a costly outcome for the society.
is
state, the elite
utility
back to
rise
=Td -r
l
However, such promises are not credible because future tax rates
are determined in future political equilibria,
tax rate will
A
where
0,
11
required to prevent a coup.
even a strategy of setting r
tax rates in the future.
this
—
E)
,
d
In this case, the society will revert back to a nondemocratic regime
though
,
than r m which the poor
is less
This tax rate r
absence of this threat.
set in the
reduces the tax rate that
poor would
l
(fi
D, r d ). Although the society always remains democratic, the threat of a coup
Implicit differentiation shows that r
<
Vr
r d such that
l
l
because in an unequal
likely
(l-p(l-s-q))A r {6) + (3{l-s-q)8 r {6) + <p {l-p(l-q)) =
If ip
higher inequality
0:
semi-consolidated: the poor can avoid a coup by
important and influences taxes: the tax rate r d
would have
<
qq
and makes a coup more
q, s)
society the rich lose
If ip
that
is
the
maximum
[^^ - C(r
d
tax rate defined by
we obtain dr d /d9 >
)\
+
[^f - ^1] 0(1 - s - q) +
(1). Differentiating
0.
10
l
<p
(l
-
(3(1
and noting that d6p (9)/d9
-
<
q))
= 0,
where r™
and dr m /d8
<
0,
=
Second, they can choose to maintain political power, 7
taxation. In this case, the rich impose a tax rate r e
but redistribute through
0,
and give the poor return
Vp
h
,
(fj,
E, r e )
where
^(/A£,T e =
ti
)
+ rf{9) +0
\qV\ii
h
,E^) +
(1
- g)VV,£)]
(9)
= T e — r e h So the poor receive net income (1 — r e )hp from their own earnings
and transfer T e — r e — C (r e ), giving them income of hp + rf(9). If next period, we are
k
then redistribution continues. But, if in the next period the economy
still in state fi = fx
= fJ, redistribution stops. This captures the notion that the elite cannot
switches to
l
and
rf(6)
.
,
jjl
commit to future
e
Also note that r
rm
is
since this
,
r m that
<
,
the rate that maximizes redistribution to a poor agent.
is
Thus, given the actions 7 and r
h
state (n ,E)
e
e
G
by the
either democratization or redistribution
revolution.
effective revolution threat.
the elite will not tax themselves at a rate higher than
is,
not sufficient to stop a revolution, then no tax rate r
With
an
redistribution, unless the future also poses
of the
[0, 1]
elite,
will
do
the poor
If this
tax rate
so.
may
still
prefer a
the value to the poor in the
elite,
is
Vp
Combining
h
j,
(f
(5)
,E)=m^(v p (R)^Vp
and
we
(9),
(ip
calculate the
h
,D)
+ (l- 1 )Vp
maximum
i
h
(f
,E,T e )).
utility that
can be given to the
poor without extending the franchise:
vy^-O.^ -^
-«»*(«>
1
1
where we used the
r = rm
hence
,
value of
fact that to give
=
rf (9)
l
6 (9).
maximum
The poor compare
(10 )
\j
utility to the poor,
(10) to
V
P {R).
the
elite
would
This defines a
set
critical
h
fji
,
Ji(e,q)
such that
= \(l-(3(l-q))6p (6) +
V p (]J(6,q),E,T - r m = V P (R).
)
attractive for the poor in state
the rich cannot stop
Notice also that
it.
^^ =
h
fj,
that even the
Democratization
1
+
A(l
For
-
(3(1
-
is
1
>
h
fi
(11)
>
Jl(9,q), a revolution is so
maximum amount
of redistribution
therefore the only option
q))^1 >
0,
left
to the
by
elite.
so high inequality reduces the
revolution threshold because the poor are worse off in a nondemocratic regime.
For
/l
(/i
,
h
fi
<
~j2(9,
q),
democratization can be avoided by redistributing to the poor in state
E). In this case, the tax rate that the elite have to set in order to avoid revolution
11
is
r
=
Vp
r e such that
,
h
((j,
,E,T e )
= VP (R).
increasing in the level of inequality).
(i.e.
we
Finally,
Vp (ip
is
h
V
p
absorbing state, the value function
The value
unconsolidated democracy
V p (n
h
,E)
= Vp (ip
Assumption
h
D)
,
greater than
takes into account the future possibility of
Vp (ip = V p {fx
l
,E),
)
democracy
_
_
~
(3 {1
s
fi
we obtain
this sufficient condition as:
<
p
q))9
q ))8 (9)
amount
h
fj,
is
higher level of
and
jl(9, q)
The
racy.
and
l
<p
>
h
/j,
=
//
h
fi
,
r
—
h
fi
>
and the
society switches to
and
~p>{9,q)
</?'
<
=
=
tp
ip
h
fj,
,
then we are in a
democracy the
(p(9,q,s),
The
T
=
rd
< rm
A
may
,
))
)
Assumption 2
will
first
but
l
<p
>
time
thereafter.
such that
Vr
l
(/i
,E)
/tx
=
h
//
by r = r m
,
,E,r
e
)
0,
=
democ-
and remains
.
we
!p(9,q,s), then
are in a semi-
democracy the
When <p = <p r =
- ip = V T {up D r d
h
first
r"\
,
1
l
,
so de
of C),
-
(x-e)-xc(T'-)-
we have
,
)
.
time
When
A
lower
fiscal
fiscal
hold when democracy
is
%<
since T
<
L
<
i-/3(i-</)'
sufficiently redistributive.
too limited, democratization does not prevent revolution.
12
ana y A
°)/ A
—
0.
This leads to an interest-
a highly redistributive democracy leads to political instability, but
is
=
t
increase or decrease the overall level of taxation.
- (\-e)(i-0(i- q + \-e
m
> C'(r e ) (by the convexity
C'(j
T
redistribution
Vp (n
h
l
/i
lower level of 9 (higher inequality) increases the variability of
policy but
ing trade-off:
=
reduces the amount of redistribution and increases the variability of
l
policy.
13
where
fj,
fully consolidated
society switches to
and remains democratic
tp\
r
e
When
variability of fiscal policy.
(p(9, q, s),
consolidated democracy.
[i
nondemo-
starts in a
or lower 9 (higher inequality) increase both the
democratic thereafter, and taxes are always given
If
"'
/
text):
and the society
2 hold
When
no redistribution.
of redistribution
>
1
h
""
"
then the society remains nondemocratic.
Ji(6, q),
VP (R). A
3.
13
Then:
and there
If
higher than that of
Solving (2) and (5) simultaneously with
+ A (l - 0(1 1-/3(1---?)
_
Suppose Assumptions
1:
h
If
is
ensure that the value to the poor of an
V r (R).
l
not necessarily an
is
2:
cratic regime.
2.
((p
Since democracy
can establish the following result (proof in the
Proposition
1.
is
,D) and
(1
Now we
h
suffices to
it
in 9
parameter space where democratization
to the poor of a semi-consolidated
a democracy subject to coups, so
and decreasing
,
fj,
12
> VP (R).
,D)
h
increasing in
is
restrict attention to the area of the
prevents a revolution, that
coups.
which
if
the potential for
If
4.
h
fi
The
>
ip
<
l
Tp(9, q, s),
then we are in an unconsolidated democracy.
l
the elite set r
,
democracy, when
=
ip
=
and when
0,
h
<p
other three cases, and there
more than
in case
amount
that given the
tributing. Therefore, in state
by setting the tax rate r
there
,
=
both the
ip
l
lower
and the
greater than in the
is
and
redistribution than in cases 2
is less
A
a coup.
is
level of redistribution
h
which
elite set
just
is
a revolution
Ji(9,q),
and the value of
the
/i',
<
fi
r
—
u
q,
3,
and
sufficiently costly
is
the elite can avoid
1
0,
while in state jJ
,
is
it
by
redis-
they redistribute
enough to stop a revolution. In
never democratization and the amount of redistribution
is
there
,
1.
of inequality
e
ip
variability of fiscal policy
type of equilibrium where
first
they extend voting rights. In a
,
,
The
variability of fiscal policy.
h
fj,
= r m and when
r
,
—
fj,
level of 9 (higher inequality) increases
In the
when
society continuously switches regimes. In a nondemocratic regime,
=n
fi
and
/!(#, q)
this equilibrium,
More
relatively limited.
inequality nonetheless increases the level of redistribution in this regime because the rich
are forced to choose higher taxes to prevent a revolution in the state (n
Now
h
/j,
>
1
consider the case with y!
When
]l(6,q).
h
,
E).
the economy transits into state
the rich can no longer maintain their political power via redistribution, and must
,
extend the franchise. There are three types of equilibria depending on the value of
cp
>
l
from
democracy, once created,
<p(9, q, s),
fJ to
t
= rm
is
very
n
h
the
,
little
or
no
We
countries.
It is
The second
democracy
l
(<p
,
likely to arise
possibility
=
coup.
at its highest level, there
when
l
that
(ip
l
,
redistribution
<p
<
if
9
is
high, that
With
,
once the society
most
<p(9,q,s),
but
cp
l
>
the society
is
fairly equal.
In this case,
lp(9,q,s).
the poor were to set a tax rate
OECD
rm
in the state
However, the poor can avoid a coup by setting a lower
D), which
is
just sufficient to dissuade the elite
it is
in
from mounting a
some sense "under the
poor realize that they will only receive
think of this as a non-credible promise
Unconvinced by this promise, the masses would attempt a revolution.
q low, the revolution threat
and then
of future redistribution by the
is
role
as the threat of a coup keeps overall redistribution below the level
transfers for a short while
revolution
is
when
is
Although the society always remains democratic,
of a coup"
moves
amount of
not fully consolidated;
T d in state
shadow
The
is
first
the poor always set
interpret this case as similar to the situation in
D), a coup would occur.
tax t
14
is
more
the state
If
this,
and the threat of a coup plays no
fiscal volatility,
becomes democratic.
When
fully consolidated.
.
extend the franchise. After
elite are forced to
In this type of society, the
.
is
l
<p
elite.
is
transitory, thus the
transfers will cease.
only prevented by franchise extension.
13
We might
15
In fact,
of a fully consolidated democracy.
we show below that higher
may
inequality
reduce redistribution in a semi-consolidated democracy.
The
final
democracy
elite,
is
type of equilibrium involves n h
unstable:
state
and cannot be halted by reducing
fluctuate between
the
when the
elite in
democracy and
power and they
set
=
r
0.
ip
l
a coup
,
More
is
But
state
,
policy
is
in this
3,
but more than in case
regime because
maximum
eventually
come
the economy starts with
h
/j.
they extend the
,
=
0.
The
h
(<p
D)
,
to
variability of fiscal
is less
Higher inequality increases redistribution
1.
when
never any redistribution during dictatorships.
are in power, they set the
will stochastically
and the amount of redistribution
increases the tax rate
it
relatively attractive for the
as soon as the state goes from
therefore highest in this equilibrium,
than in cases 2 and
In this case,
Tp(8,q,s).
moves to
D), they mount a coup, regain political power, and set r
l
<
specifically,
Whenever the
.
l
ip
As a result, the economy
taxes.
= rm
and
~p(8,q)
moves to
elite control.
franchise after which the poor set r
(ip
>
there
is
democracy, while there
is
when the poor
Notice that in this case,
tax rate, fully anticipating that redistribution will
to an end as a result of a coup.
This result
may
help to explain the
existence of highly redistributive, but relatively short-lived, populist regimes of Latin
America
(see, for
Kaufman and
example,
Stallings, 1991,
There are four conclusions to be drawn from
to regime changes.
A
decrease in 8 reduces
Therefore, societies with
more
racy and dictatorship, and
results are in line
and
political instability (e.g.
see this, fix the cost of
l
=
p(8
L
,q,s).
so inequality
is
are
have a
0.
coup
ip
l
and define
,
Moreover, suppose that
semi-consolidated,
15
Tp(8, q, s)
likely to switch
fully consolidated
h
fj,
>
8
>
~p(8
low that democracy
sufficiently
However, as 8
reduce taxes
more
pertains to the link between inequality
and
inequality
<p(8, q, s).
between democ-
democracy.
falls
1
i.e.,
from T m
if
to r
Wantchekon (1998) argues
€
d
is
8
L
and
So our
such that
L
,q).
When
8
redistribution.
fully consolidated.
(e.g.
H democracy
below 8
,
(lp(8,q,s),<p(d,q,s)).
To
= <p(9 H ,q,s) and
<p
> 8 H <p > <p(8,q,s),
l
l
,
Now
consider an
This will increase redistribution at
the standard models of voting over redistribution
<
and increases
first links
Muller and Seligson, 1987, and Alesina and Perotti, 1996).
increase in inequality (a reduction in 8).
A^p
The
with the empirical findings of a positive association between inequality
The second conclusion
tp
less likely to
this analysis.
Jl(8, q),
initial inequality
and Sachs, 1990, on populism).
first
as in
Meltzer and Richards, 1981), since
is
no longer
fully consolidated,
but
In this case, the poor are forced to
in the state (</?',£)), so overall redistribution falls.
In fact,
been the case in El Salvador, and Weyland (1996) argues that
restrained by the threat of military intervention. This result is
also related to Ellman and Wantchekon (1998) and Wantchekon (1998) who analyze how the threat of
conflict initiated by the loser of a democratic election affects the voting outcomes.
this has
the redistribution in democratic Brazil
is
14
in a semi-consolidated
The average tax
ambiguous.
while r d
is
democracy, the relationship between inequality and taxation
rate
r
is
a
=
(1
— s)T m +
decreasing. If the cost of taxation C(t)
dominates and the average tax rate
is
sr
rm
d
.
highly convex, then the second effect
10
that taxation increases
when
a coup more attractive
for the elite, so to prevent the coup, the
inequality increases.
tax rate substantially in the state
As
inequality increases further,
increasing in inequality
is
as inequality rises, though
falls
is
it is still
more
Intuitively, higher inequality
likely
makes
poor have to reduce the
leading to lower redistributive taxation on average.
</?',
we have
<
9
6
L
so
l
tp
<
and democracy
Tp{9, q, s),
is
now
unconsolidated with lower overall redistribution than both in fully and semi-consolidated
democracies.
Therefore, there
is
a nonmonotonic relationship between inequality and
redistribution, with societies at intermediate levels of inequality redistributing
both very equal and very unequal
The
societies.
third implication of our analysis
between
fiscal volatility
and inequality
related to fiscal volatility.
is
is
likely to
higher inequality leads to more variability.
which has the highest amount of
fiscal
policy has been
much more
more than
17
be increasing.
The
relationship
Within each regime,
Moreover, higher inequality makes Case
fiscal variability,
more
likely.
America than
volatile in Latin
This
in the
may
explain
OECD
4,
why
(Gavin and
Perotti, 1997).
The
final implication of
impact on the equilibrium
less
convex, so that
C(T m )
our analysis
is
that the costs of redistribution will also have an
Suppose that the cost of taxation becomes
political system.
is
unchanged, but C'(T m ) decreases.
from taxation are now lower, the median voter
However, as r m increases, so
and hence
less likely to
will
—Sr (9),
be consolidated.
creates less economic distortions, for
GDP
1G
A
sufficient condition for the
° \T
17
I
^
>
elite,
This implies that in societies where taxation
example
average tax rate to
(l-«-/?(l-«-g))[(l-A)C'(T")-(0-A)1
As[(l-j3(l-s- 9 ))T«'+/3(l-«- 9 )ri']
choose a higher level of taxation.
democracy becomes more costly to the
in societies
generated from natural resources, democracies
is
rn ( ^m\
so
will
Since deadweight losses
fall
where a large fraction of the
may be
harder to consolidate.
as inequality rises
is
'
may not generate high levels of redistribution
have emphasized different ideas. First, the rich may control more votes (Benabou, 1999, Rodriguez, 1998).
Second, the poor may view themselves as upwardly mobile and not vote for high redistribution (Wright,
1996, Benabou and Ok, 1998). Third, if redistribution fulfills an insurance function and if insurance
is a normal good, increasing inequality may reduce the demand for insurance and hence redistribution
(Moene and Wallerstein, 1998). Finally, there may multiple dimensions on which citizens vote, splitting
Previous explanations of
why high
levels of inequality
the constituency for redistribution (Roemer, 1998).
15
Repression
3
So
far
we have assumed
not allowed them to use repression
Such repression
social unrest.
some
when the
society
In this section,
example, in Indonesia in
cases, for
we
The main
of our results.
a democracy, but we have
in
nondemocratic in order to prevent
is
many
observed in
is
1965 and in El Salvador in 1932.
repression will qualify
mount a coup
that the rich can
show that the
will
results of our analysis so far are:
1.
democracies are more likely to consolidate in more equal societies.
2.
frequent regime changes are more likely in
In this section,
we show
possibility of
more unequal
societies.
that in very unequal societies, the elite
may have
much
so
to lose from democratization as to prefer a repression strategy to suppress revolution
and prevent democratization. As a
changes
be more
likely in
To analyze
power, and
it
will
nonmonotonic as
potentially
is
a more equal society.
we have
h
>
fi
at per period cost
M
(Rep)
use repression
but democratic consolidation continues to
well,
18
=
l
ip
for
ft
-
[Zff
ip(9,q,s), so that
19
of the
It is clear
,
D, r m )
=
—*_i
Assume
'
.
elite,
and that
also
this strategy completely
follows immediately that the rich will find
it
beneficial to
if
This condition will be satisfied
if
qM _
R
(-M).
Next, define 9
revolution with redistribution
no democratization: a
necessary will also
if
9
inequality
such that
< 9R
.
r
is
h
fi
(12)
(9)
sufficiently high,
=
R
~[i{9
,
9
military repression desirable.
)
will
,q); so
This implies that
level of inequality that is large
make
a society with 9 G (8
the
i.e.,
elite
if
<
9
9
M
where
cannot prevent a
when 6 R < 6 M
,
there will be
enough to make democratization
R
M
interestingly, if 9
>
More
9
,
democratize because social unrest cannot be prevented
by redistribution and military repression
18
l
that with this strategy the rich will have a
M < -8
p
V r ((p
the society democratizes,
if
purpose of suppressing revolutionary threats,
sole
member
each
>
rich will obtain
army with the
avoids the threat of revolution.
V
and
Jl(8,q)
and the
forever,
that the rich can hire an
return
between inequality and regime
these issues in the simplest possible way, suppose that the elite are in
remain so
r
result, the relationship
is
too costly.
In contrast,
if
9
<
9
M
,
then
is more likely in more unequal societies also, as long as social unrest suppressed
counted as "political instability".
19
For example, the army may be financed by taxation, in which case
is the tax paid by each member
Political instability
by repression
is
M
of the elite for this purpose.
16
inequality
is
so high that the elite are willing to
prevent democratization.
either, this
It is still
20
time because the
true that
among
are less likely to consolidate
and dictatorship.
result that
low
elite
if
>
9
9
pay
for military repression in order to
then there will be no democratization
,
can prevent social unrest by redistribution.
the countries that democratize, those with greater inequality
democracy and
only those with 9 G (9
It is
The second
consolidate democracy.
main
Finally,
R
between democracy
will therefore oscillate
M
9
,
R
)
and
ip
>
l
that transit to and
7p(9, q, s)
condition requires inequality to be low. Therefore, our
conducive to the consolidation of democracy
levels of inequality are
continues to hold in this model with repression.
of frequent regime changes or repression,
is
also
Political instability, either in the
more
when
likely
inequality
is
form
high.
Asset Redistribution
4
We now
extend the model of Section 2 to allow
redistribution.
To
simplify the analysis,
on the configurations of
interest rather
we
for asset redistribution as well as fiscal
ignore the possibility of repression, and focus
than giving a complete characterization of
possible cases. Let us define 9q to be the initial distribution of assets
and use 9
all
the
for the
distribution after redistribution has been undertaken.
Unlike
fiscal redistribution, if asset
be reversed, but
it
also has
inequality
permanent
costs.
is
reduced,
We
it is
model the
permanent 21 and cannot
costs
asset redistribution reduces the total stock of assets in the economy.
stock, then the post-redistribution stock
differentiable decreasing function,
is
<
H'(.)
i.e.
reduces total resources. Also obviously h
4.1
H(9), where if
= H(9
=
If
h
is
the
0,
so asset redistribution
1.
Asset Redistribution under Democracy
Recall that 9
L
is
defined by
l
(p
=
L
where 8
!p(9 ,q,s)
so that without any asset redistribution the
L
>
9
,
and assume
economy would
oscillate
l
<p
<
7p(6 ,q,s)
between regimes.
Hence, for democracy to be (semi-) consolidated, inequality needs to be reduced,
needs to be raised to 9
20
initial
a concave twice continuously
is
and H"(.) <
)
by assuming that
L
.
Also assume that 9
R > 9 where
h
jj,
=
R
~p(9
,
q).
We
first
i.e.
9
assume
Notice that the formulation with a fixed cost of military repression is not essential for this result.
if the degree of repression is variable, it will be the societies with greater inequality, where the elite
Even
have more to lose from democratization, that use more repression.
21
For our general results, asset redistribution does not need to be permanent, it only needs to be
harder to reverse than fiscal redistribution. In practice, it may be easier to reverse asset distributions
than democracy, which
is
ultimately an empirical question.
17
that 9
R
is
Vp ((p h
Let
,
Vp (ip h
where 8P {9)
W
9
<
9
L
,
p
is
is <p
D,
,
h
=
9)
then the poor
if
W$(6)
=
if</>'
9
+
8 P {9)
=
9H{9)/\
>
9
L
,
[(1
is
given by
- s)Vp {ip h D, 9) + sWp (9)]
l
ip
=
9H(9)/X. The continuation
depends on 9 relative to 9
,
9.
(13)
,
+ p[(l-
q)W?{0)
coups can be stopped,
9H{9)/\
)
L
.
If
+ qVp (<p\ D, 9)}
so:
+ A p (9) +p[(l- s)Vp {ip h
,
D,
9)
+ sW$(9)\
defined as in the previous section (and
is
)
(14)
(15)
Ap (0) =
8 P {9)
><p(9,q,s)).
'
which
is
function
>
9
9
L
,
we
when coups
XS
-
=
>
Wf(9)), we obtain that
(16)
(1-/3K1-W-,-,))
(i.e.
(15)
W
p
{9)
=
path.
W$(0)) to obtain the value
22
i^m . vm + x > vm
)
asset redistribution, let
.
=
6>o
argmaxFp (^),
9'
(17)
1
bearing in mind
where no asset redistribution
is
chosen.
aigmaxV£(9). Then:
9
L
,
the masses will redistribute assets up to 9".
"Observe that \6P (6)
+
is
illustrated in Figure
(1
-
r
X)8 (e)
T
gi
1
~
5P e
(
)}
Intuitively, in this case, the
poor prefer ignoring the coup constraint also prevents
= -C(r m
)
1.
and AA?(0)
these with the equation that determines r d in footnote 11,
[~
=
DsS'(e)
at a corner solution with 6'
coups. This case
))
(9)
]
level of redistribution that the
%T-7- q
p
^ [-2^ + .^k^^g^pM^ —>
we might be
If 9"
(i
W
when coups occur along the equilibrium
y
are avoided as:
To determine equilibrium
Also, let 9"
and
solve (13)
vm . «wa ±
where
imposing
(i.e.
ejm/x + g(g)
TTs
^=
the value function that applies
When
{1
a democracy when
be unable to stop a coup by redistributing, and therefore:
Ap (9) = r d (l - h p - C{r d
V
1.
+
coup cost transits to state
Solving (13) and (14) simultaneously
that
in
democracy, and the distribution of assets
is
9H(9)/X
will
the other hand,
where
in this subsection.
it
defined as in the previous section and hp
W?(6)
On
there
,
after the
(9),
ignore
D,9) be the present discounted value of a poor agent
the cost of a coup
value,
we
very high, so
Substituting in (17),
18
+
- \)A r (6) = -C(t"). Combining
A p (8) = ^' ^i-p\l-J- ))^ ~ ^p- +
(1
we obtain
we obtain V%(6)
q
=
V?{6)
+ X.
2.
9"
If
will
3.
<
9
L
,
and V[{9') > Vf(0 L ), then the poor
probably the most interesting one
may
for
L
.
This case, illustrated in Figure
our purposes as
it
illustrates that the
2,
poor
choose a high level of redistribution in order to prevent coups.
e
L
e"
Figure
An
9\ and coups
occur along equilibrium path.
Otherwise, the level of redistribution will be 6
is
will redistribute to
increase in
making coups
l
(p
increases the
less attractive for
1:
gap between Vf(9) and V${9)
the
elite.
(i.e.
raises
X
in (17)),
This reduces the concessions that the poor
have to make to defend democracy. Therefore,
it
makes them more
willing to choose a
higher level of asset redistribution to consolidate democracy. Also, an obvious comparL
ative static pertains to the level of inequality: if 9$ > 9
there will not be a motive to
,
redistribute assets in order to consolidate democracy. So
tion to emerge as a
method
of consolidating
we may expect
democracy especially
asset redistribu-
in relatively
unequal
democratic regimes that are expected to be threatened in the future.
Overall, therefore, the
main implication
help to consolidate democracy.
of this analysis
Whenever the choice
is
that asset redistribution can
of the poor
is
9
L
or greater, coups
no longer occur along the equilibrium path because asset redistribution has changed the
level of inequality
permanently, and makes coups
19
less attractive for
the
elite.
9'
9"
9L
Figure
2:
number
of
we argued that educational expansion
in
In practice, asset redistribution appears to have played such a role in a
instances.
Acemoglu and Robinson
In
(1997),
19th century Britain and France was in part a result of democratization, and Engerman,
Mariscal and Sokoloff (1998) argue the same for Latin America. In Britain and France,
these and other policies reduced inequality and there were no significant reversals in the
process of democratization. In Costa Rica, the educational and land reforms that reduced
both earnings and land inequality
after the democratization in 1948
appear to have helped
with the consolidation of democracy (see Yashar, 1997, for this argument, and Vilas, 1995,
for
some numbers). 23 The
situation in Venezuela after the return to
democracy
in 1958,
which led to a land reform redistributing 19.3 percent of agricultural land, 24 also provides
some support
4.2
We
to this view (see Powell, 1971).
Anticipated Asset Redistribution and Political Instability
have so
far illustrated the role of asset redistribution in consolidating
Since asset redistribution
is
permanent and costly to the
23
elite,
democracy.
the anticipation of such
Chalker (1995) writes that "the most remarkable egalitarian measure in Costa Rica occurred in the
when the concentration in income distribution was reduced. Interestingly this was an
outcome, rather than a cause of, democratic politics."
1960's and 1970's
24
See Table 10.2 in Cardoso and Helwege (1992).
20
may
redistribution
We
l
ip
>
now analyze
will
~ip(9o,q,s)
so that democracy
)
(semi-)consolidated without asset
is
period of democracy, in the state with
first
However, we assume that there
and the implementation of
h
<p
The poor
.
then the rich
a one period delay between the
is
h
(ip
may mount
a coup to avoid asset redistribution before
it is
we assume that during the
be no
will
implemented.
(i.e.
in
fiscal redistribution.
this case, first write the value to the rich after asset redistribution
inequality to 9 as
,D),
((p
process of asset redistribution
both the period of legislation and implementation), there
To analyze
l
,D) to
during this period, the state switches from
simplify the analysis,
legisla-
For example, land reforms involve
asset redistribution.
If,
administrative delays.
To
9
Suppose that
to redistribute assets, this time not to consolidate democracy but to increase
their incomes.
tion
>
9
(i.e.,
L
attractive.
a simple extension of our model.
this using
Consider the
redistribution.
may want
make a coup more
also
changes
V T (D),
~
1
'
(1
=
is
9)H{9)/{\
-
A)
+ P(l -
r
+ j3sAr(9)
s)6 (9)
1-/3
l
(whether the state
-
(<p
h
,D) or
((p
,D) does not matter as there
is
no current
fiscal
redistribution).
Suppose now that asset redistribution
to the rich
is
h
legislated in the state
((p
,
D), then the value
is
~^ Hx
(1
VV.D) =
{do)
+p
- s)W(D) +
[(1
This value function takes into account that
l
the state switches to
<p
,
the rich
if,
may mount
sm^{V r (D),Vr (f/,E)
-
1
<p
}]
during the period of administrative delay,
a coup. The value to the rich in the state
(/*',£) is
V V, E) = (l- 9 Q )H(9
)/(1
-
A)
Solving these equations, the critical value,
.
V =
where kt (6)
=
(1
-
-(1
+ (3 [(1 - q)V
l
tp
,
at
which
6)H(6)/(1
-
A)
-
(1
-
following asset redistribution legislation,
s)8p (9)
=
+ PsA p (9)
0H(d)/\ - 9 H(9 )/\
is
l
(fi
,E)
+ qV r & h D)
,
V r (D) = V r
l
(fi
,E)
—
l
<p
,
is
+ 0q) Ur (6) + P(l - s)6 r {6) + psA r (6)
~
l-P(l-q)
)H(6
a result of asset redistribution. In the case where
Define kp (6)
T
if
)/(1
l
<p
<
-
A)
l
ip
,
<
is
the loss for the rich as
the elite would
mount a coup
the cost of the coup were to change to
similarly,
21
l
.
+ /5(1 hence k p (9) +
then we can notice that kP(9)
the discounted average per-period transfer to the poor,
ip
-
(3(1
s)S p (6)
K
+/3{1
(d)
<p
-
(3(1
- s)6 r (B)+PsA r (d) < K r (8
>
l
Hence,
> kp (9) +
(3sA p (9)
s)8p (9)
+
(3sAp (9) for any 9 since 9
Tp
l
as defined
asset redistribution,
by
(8). So,
democracy
is
)
+ f3(l - s)6 r (9 )+(3sA r (9
there exist values of
consolidated, but
if,
l
ip
If s,
the masses
(7p
which
may
may
is
T
<
-s-q).
,ip) such that
without
during the period of administrative
l
<p
,
the
the probability that the cost of the coup switches to
attempt
elite will
tp
l
is
,
low enough,
prefer to enact asset redistribution despite its potentially destabilizing
Therefore, the
effects.
chosen
(38 (9 )(1
)
l
€
delay after asset redistribution, the cost of a coup switches to
a coup.
is
Since taxation and asset redistribution are costly, this in turn implies that
optimally.
r
+
main conclusion
of this subsection
is
that asset redistribution,
generally in the interest of the poor and often useful in consolidating democracy,
create a temporary period of instability for a democracy.
A
number
by a
of coups in Latin America appear to have been motivated
prevent radical land reform.
desire to
For example, in Brazil, a central aim of the coup in 1964
was to prevent the attempt by the left-wing President Goulart to by-pass the veto of
the Congress and use other means to push through agrarian reform
(see, for
example,
Skidmore, 1967, and Wallerstein, 1980). Similarly, most scholars argue that the agrarian
reform after 1952 in Guatemala was the main motivation
1984, Trudeau, 1993),
cially
for the
coup of 1954 (Handy,
and that the increasing radicalization of Allende's
on land reform, precipitated the coup of 1973
in Chile.
25
policies, espe-
The same
is
true in
Venezuela, where the 1948 land reform law was immediately repealed by the incoming
military government (see Dorner, 1992, p. 47,
4.3
The
Asset Redistribution in Nondemocratic Regimes
rich
may
also wish to undertake asset redistribution in order to stop a revolution or
Therefore, our analysis in this subsection will reveal another strategy
democratization.
for the elite
repression.
when
We
that democracy,
faced with the threat a revolution, in addition to democratization and
continue to assume
if
n
h
>
~p,(9
26
,q),
but suppose that
created, will be fully consolidated.
away from the poor
2o
and Powell, 1971).
is
not possible,
See Valenzuela (1989).
Brown
i.e.
(1989, p.
9
We
also
ip
l
>
<p(9 ,q,s) so
assume that redistribution
>9q.
236) writes "a second generation of [agrarian] reform
was clearly gaining momentum in the latter part of Allende's administration - a fact that was not lost
on counterreform elements that ultimately supported Pinochet's coup d'etat." Of the land originally
expropriated by Allende's government, 43% was returned to previous owners or excluded from the reform
by other means, see Jarvis (1989, p. 249).
26
Hence,
>
0o, so that the level of inequality that
level.
22
prevents democratization
is
less
than the
initial
The
F V, E, 9) = (Iwhere h r
If #
ip
>
l
will
when the
value function for the rich
=
—
(1
< #K
9)H{9)/{\
9)H(e)/(l
—
and
A)
W
r
,
ensures that
>
l
<p
(p(9, q, s) for
Our above
always be fully consolidated.
is
=
(9)
(18),
the elite
may
on whether
1
l
{fi
,E,9)
In contrast
when
9
if
=
9
> 9R
,
=
Whether the
The
\
)
=
\
Hence, using
)
_
^%
+ (i-m- %-,))
(1
-
-
9)H{9)/{\
elite will
and
)
V{((j,
W
,
{9)
=
,,,, )=
+
rf{6)
+ {3
depending
(9)
what
we
follows,
do
this,
so
r
).
If
V{(fj!,E,9
)
implication
is
that the elite
reform, South Korea redistributed
E)
- q)V T ^\ E, 9) + qW[{9)]
[(1
is:
(19)
<
may
determined simply by comparing
R
V{(n ,E,9 ), then the elite prefer to
l
will
redistribution sufficient to prevent
choose not to redistribute, so 9
= 9q.
choose to redistribute assets in order to
cases appear to
50%
(20)
is
minimum
.
Two
,
M«1^
=
28
prevent democratization.
h
(fj,
Wffl), we obtain
9
is
costly to them. In
choose asset redistribution
E,9
1
r
A)
R Otherwise, they
9
democratization, that
The main
more
In this case,
r
continuation value of the elite in state
prevent democratization and will choose the
p. 55)
)
\
democratization can be avoided by redistributing to the poor
W
,E,6
so democracy, once created,
.
Solving (18) and (19) with
V{(fi
,
The assumption
E).
sufficiently costly that the elites will not
is
/A
27
the revolution threat arises.
WZ(9)
l
9
,
wish to undertake asset redistribution in order to reduce 6
assume that asset redistribution
axgmaxV
>
ft
(/x
when they cannot prevent democratization.
is
(18)
analysis then implies that the return to
1 - e H e )/0-- x +s
e
is
asset or fiscal redistribution
r
any 9
W[{9) =
we find that when 9 < 9 R ,V{(n ,E,9)
the value function of the elite
is:
return after the state switches to
l
W{{9) in
equal to 9
is
+ p [(1 - q)V r (fJ, E, 6) + qW r (9)\
(0) is their
the rich under a consolidated democracy
W
and inequality
is /r
then democratization would take place in the state
,
<p(9 Q q, s) also
r
A)
state
fit
this implication nicely.
of the agricultural land in Korea.
In a 1949
Haggard
(1990,
argues that the reforms were aimed at defusing rural insurrections and counteract-
ing the destabilizing spillovers from land reform in North Korea. Taiwanese land reforms
"Formally
t oi many,
-H(e ,)+(\-e a )H'(e n
(
(i_/3)(i_a)
)
+
ggQgoj
(i-/3)(i-/3(i-<?))
<
Q
- a
28
This result is related to previous analyses of land reform, such as Grossman (1993) and Horowitz
(1993), which argue land reform can prevent revolution, though these papers do not compare asset and
fiscal redistribution. Also in contrast to these papers, asset redistribution in our economy may prevent
not only a revolution but also democratization.
23
163) also appear to
of 1949-1953 that redistributed 24.6% of the land (Ho, 1978, p.
have been an attempt to defuse rural protest (see Amsden, 1985). In the words of Ch'en
Ch'eng, the governor of Taiwan at the time of the reforms, "...the situation on the Chinese
mainland was becoming
critical
and
It
of unrest
instability.
and the
villages
on the island were showing marked signs
was feared that the Communists might take advantage of
the rapidly deteriorating situation" (quoted in Haggard, 1990,
p. 82). Interestingly, until
very recently both South Korea and Taiwan remained relatively nondemocratic, especially
compared to other countries with similar per capita income
levels.
can also be observed that asset redistribution as a method of preventing democ-
It
ratization
more
is
emerge when 6
likely to
democracy, are larger. This result
may
r
(9),
the transfers away from the rich in a
help explain
why
asset redistribution
emerged
in
South Korea and Taiwan, where the threat of communism made social unrest very costly
to the
elite,
but not in the Philippines, where the threat was
Finally, although
among
we have modeled
whereby the
elite redistribute assets selectively to
portant for the threat of revolution and
This
transfers.
ample of Mexico
is
by the
assets redistributed
the poor, a more appealing interpretation
all
less serious.
may be
groups
equally divided
a strategy of co-opting,
among
who can be persuaded
elite as
the poor
are im-
to switch sides with such
may
equivalent to our formulation, but this interpretation
in the 1930s better,
who
fit
the ex-
where small groups of peasants that supported the
ruling party were given land.
Targeted Transfers and Political Instability
5
Our
analysis so far has
This assumption
is
assumed that
transfers are not targeted to a particular group.
clearly unrealistic in
means-tested and taxes are non-linear.
many
Perhaps more importantly, groups with diverse
political interests are often geographically or ethnically
only one group quite easy.
this possibility.
fraction
a
as given).
1
It is
In practice, transfers are often
settings.
segmented, making transfers to
straightforward to extend our analysis to allow for
In particular, suppose that
of all tax revenue to itself (where
when group
i is
r
ap > A and a >
in power,
1
—A
it
can allocate a
are parameters taken
This changes the equations in Section 2 in an obvious manner and also raises
the possibility that even the rich
in order to tax the
may
choose a positive level of taxation
when
in
power
poor and transfer some of the resources to themselves.
The key equations from
Section 2 are modified as follows.
24
The poor
will
now
set
a
m
tax rate T given by
ap -
6
A
when they
are not constrained by the threat of a coup,
and the
rich will set a tax rate,
r nd satisfying
,
when they
When
are not constrained by the threat of revolution (and
the state
is (<p
h
equation (2) in Section
,
r
S (9)
2,
= l^T m - T m h r
l
In the state,
(<p
setting r
=
with the only difference that now, 6P (8)
may once
,D), the poor
With targeted
l
,
E) the
all
income
r
at the rate
make
transfer that the poor can
rich will
r nd
=
= ^-T m — r m hp
transfers, however, they
now tax
at the rate r
nd
nd
0).
,
and
can do better than
and target a
x
is
will also
r
{@)
ar
fraction
when they have power).
to the rich
r
x (@) were negative, r = would be better).
The value functions in the nondemocratic regime
(if
(fi
maximum
6,
again change their redistributive strategy
redistribution to the rich (exactly as the rich will do
the
—
.
they can tax
0:
1
D), the returns to poor and rich agents continue to be given by
in order to prevent a coup.
d
aT <
if
= j^T
nd
—
of the
Therefore,
T nd h T
>
change because in the state
and redistribute a fraction a r of the revenues
to themselves. Hence:
V\v\ E) =
^T
=
where x r (#)
coup constraint
is
now found
nd
is still
-
h
l
+X
r nd h r
is
is
increase in
differs
As a
from
result,
ap and
stakes (ap raises
is
- g)VV, E) + qV\n\ E)\
[(1
as defined above,
and xp {6)
and the
(6),
critical
=
(21)
^Tw -
r nd K?.
The
value for the cost of a coup
(8)
W
=
1-/3(1-,)
because
it
has the additional term
also because with targeted transfers, the cost of
higher.
to tax
+
to be
Expression (22)
and
{e)
given by equation
¥>(*,«,')
ator,
l
more
in
ar
!p(0,q,s)
raise ~fi(0, q, s),
— 8 and ar
T
raises
x
r
)-
making coups more
With targeted
valuable; the rich are forced to
democracy and are able to receive net
make
in the
democracy to the
greater than 7p(9,q,s).
is
r
x >
More
likely,
numer-
rich,
generally,
— 8r
,
both an
because they raise the
transfers, controlling the
power
greater net transfers to the poor in a
transfers in a nondemocratic regime.
the rich are willing to pay a greater cost to regain power.
Therefore,
This result also puts one
of our basic results from Section 2 into perspective; greater inequality increases political
25
instability because
it
increases the stakes in the political game, giving greater incentives
to each group to fight to regain power (a point also emphasized
A
less
similar
argument implies that the
critical value for
than Jl(6,q) as given by (11) in Section
transfers, political
regimes are more
such as
many
So,
2).
h
fi
~jt(9,q),
is
now
lower
when the government can
power has more economic value, and
between
oscillations
(i.e.,
direct
different
This implies that politics will be more unstable in countries,
likely.
in Africa,
,
by Madison, 1788).
where ethnic or geographic segregation of
different
groups makes
targeted transfers possible.
This extension also emphasizes that our model can be applied to political conflict
between any two groups, rather than only between the rich and the poor.
the two parties can be thought of as rural
urban
interests in conflict over transfer
two ethnic groups, with the one
of resources to urban areas, or
resources from the other group,
vs.
For example,
in
power transferring
and the group out of power contesting
political power.
Conclusions
6
In this paper,
we have developed a simple theory
of political transitions.
Our theory
emphasizes the role of the threat of revolution and social unrest in leading to democratization
and the
desire of the rich elite to limit redistribution in causing switches to
nondemocratic regimes. Inequality emerges as a crucial determinant of
ity as
it
political instabil-
encourages the rich to contest power in democracies, and also often encourages
social unrest in
nondemocratic
societies.
Therefore, inequality
is
However, in very unequal nondemocratic
frequent regime changes.
generally conducive to
societies,
the rich
choose a repressive strategy in order to prevent democratization and revolution.
cases,
is
democracy
limited,
is
more
and inequality
likely to
is
be consolidated
if
may
In
all
the level of inequality in the society
likely to lead to political instability, either in the
form of
frequent regime changes or repression of social unrest.
Inequality
also likely to lead to fiscal volatility, as the redistributive regime changes
is
with the political system of the society.
to
more
redistribution.
regimes, there
be used to
is
Unequal
societies switch
no redistribution.
stabilize
Nevertheless, inequality does not necessarily lead
between regimes and in nondemocratic
Our theory suggests that
asset redistributions
may
both democratic and nondemocratic regimes, but the anticipation
of a radical asset redistribution, such as a land reform,
consolidated democracy because the elite
may mount
reforms.
26
may
destabilize
an otherwise
a coup specifically to avoid the
References
Acemoglu, Daron and James. A. Robinson (1997) "Why Did
the Franchise?
the West Extend
Democracy, Inequality and Growth in Historical Perspective,"
CEPR
Discussion Paper No. 1797.
Ades, Alberto and Thierry Verdier (1996) "The Rise and
Theory of Economic Development and Social Polarization
CEPR Discussion
A
Rent-Seeking Societies,"
Paper No. 1495.
and Roberto Perotti (1996) "Income
Alesina, Alberto
Instability
in
Fall of Elites:
and Investment," European Economic Review,
Alesina, Alberto
and Dani Rodrik (1994)
Distribution, Political
40, 1203-1225.
"Distributive Politics
and Economic
Growth," Quarterly Journal of Economics, 109, 465-490.
Amsden, Alice H. (1985) "The
State and Taiwan's Economic Development," in
Peter B. Evans, Dietrich Rueschemeyer and
In,
Theda Skocpol
eds.
Bringing the State Back
Princeton University Press, Princeton NJ.
Benabou, Roland (1999) "Unequal
NBER Working Paper,
Societies,"
forthcoming
American Economic Review.
Benabou, Roland and Efe A.
Redistribution:
The
Ok
(1998) "Social Mobility and the
POUM Hypothesis," NBER Working Paper
Benhabib, Jess and Aldo Rustichini (1996)
of
Economic Growth,
1,
for
#6795.
"Social Conflict
and Growth" Journal
125-142.
Brown, Marion (1989)
Thiesenhusen ed.
Demand
"Radical Reformism in Chile:
1964-1973," in William C.
Searching for Agrarian Reform in Latin America,
Unwin Hyman,
Boston.
Cardoso, Eliana and
Trends, and Conflicts
Ann Helwege
The MIT
Press,
(1992) Latin America's Economy: Diversity,
Cambridge MA.
Chalker, Cynthia, H. (1995) "Elections and Democracy
in
Costa Rica," in Mitchell
A. Seligson and John A. Booth eds. Elections and Democracy in Central America Revisited,
University of North Carolina Press, Chapel
Cohen, Youssef (1994)
NC.
Radicals, Reformers
and Reactionaries, University of Chicago
Press, Chicago IL.
Collier,
in
Ruth
B. (1998) Paths Towards Democracy: The Working Class and
Elites
Western Europe and South America, forthcoming book, Cambridge University Press.
27
Dahl, Robert A. (1971) Polyarchy Yale University
Press,
Dorner, Peter (1992) Latin American Land Reforms
versity of Wisconsin Press,
Di
Telia,
Economy
Political
Madison
in Theory
Haven.
and
Practice, Uni-
WL
Guido and Rudiger Dornbusch (1989)
of Argentina, 1946-1983," in
Economy
New
"Introduction:
The
Guido Di Telia and Rudiger Dornbusch
Political
eds.
The
of Argentina, 1946-1983, University of Pittsburgh Press, Pittsburgh
PA.
Ellman, Matthew and Leonard Wantchekon (1998) "Electoral Competition under the Threat of Political Unrest," Unpublished Paper, Department of Political Science,
Yale University.
Engerman, Stanley
ing, Suffrage,
L., Elisa
Mariscal and Kenneth L. Sokoloff (1997) "School-
and the Persistence of Inequality
Paper, Department of Economics,
in the Americas, 1800-1945,"
UCLA.
Gavin, Michael and Roberto Perotti (1997) "Fiscal Policy
NBER Macroeconomics Annual,
Unpublished
in Latin
America"
edited by Benjamin Bernanke and Julio Rotemberg,
MIT
Press, Cambridge.
Grossman, Herschel
(1991) "A General Equilibrium Theory of Insurrections,"
I.
American Economic Review,
Grossman, Herschel
81, 912-921.
(1993) "Production, Appropriation and Land Reform,"
I.
American Economic Review,
84, 705-712.
Haggard, Stephan (1990) Pathways from
the
Newly
the Periphery:
The
Politics of
Industrializing Countries, Cornell University Press, Ithaca
Haggard, Stephan and Robert R. Kaufman (1995) The
Growth in
NY.
Political
Economy
of
Democratic Transitions, Princeton University Press, Princeton NJ.
Handy, James (1984)
Gift of the Devil:
A
History of Guatemala,
Dumont
Press,
Toronto.
Ho, Samuel (1978) Economic Development
New Haven CT.
Horowitz, Andrew W.
of Taiwan, 1860-1970, Yale University
Press,
of
(1993) "Time Paths of Land Reform:
Reform Dynamics," American Economic Review,
A Theoretical
Model
83, 1003-1010.
Jarvis, Lowell, S. (1989) "The Unraveling of Chile's Agrarian Reform, 1973-1986,"
in
William C. Thiesenhusen
ed. Searching for
Agrarian Reform in Latin America, Unwin
Hyman, Boston.
Kaufman, Robert R. and Barbara
of Latin
American Populism,"
in
Stallings (1972) "The Political
Economy
Rudiger Dornbusch and Sebastian Edwards eds.
28
The
Macroeconomics of Populism in Latin America, University of Chicago Press, Chicago
Kolb, Glen L. (1974) Democracy and Dictatorship
Press,
in Venezuela, 1945-1958, Shoestring
Hamden CT.
Levine, Daniel H. (1989) "Venezuela:
Democracy,"
Diamond,
in L.
Linz,
Juan
sity Press,
Linz,
(1978)
J.
Linz and S.M. Lipset eds.
Crisis,
Breakdown and
Consolidation, Johns Hopkins University Press, Baltimore
Seymour M. (1959) "Some
Lipset,
Development and
Political Legitimacy,"
Madison, James (1788) The
New
Democracy
Reequilibration, Johns
and Alfred Stepan (1996) Problems
J.
Sources, and Prospects of
in Developing
Hopkins Univer-
MD.
Baltimore
Juan
J.J.
The Nature,
Lynne Reimer, Boulder Co.
Areas: Latin America,
of the
IL.
Constitution,
New
of Democratic Transition
MD.
Social Prerequisites for
American
Federalist:
A
and
Democracy: Economic
Political Science Review, 53.
Collection of Essays Written in Favor
York.
Meltzer, Allan H. and Scott F. Richard (1981) "A Rational Theory of the Size
of Government," Journal of Political
Economy,
89, 914-927.
Moene, Karl O. and Michael Wallerstein (1997)
tion,"
Unpublished Paper, Department of
Moore, Barrington (1966) The
Political Science,
"Inequality and Redistribu-
Northwestern University.
Social Origins of Dictatorship
and Democracy:
Lord and Peasant in the Making of the Modern World, Beacon Press, Boston
Muller,
tions,"
Edward N. and
American
MA.
Mitchell A. Seligson (1987) "Inequality and Insurrec-
Political Science Review, 81, 425-451.
O'Donnell, Guillermo (1973) Modernization and Bureaucratic Authoritarianism:
Studies in South American Politics, University of California, Institute for International
Studies, Berkeley
CA.
O'Donnell, Guillermo and Phillip C. Schmitter (1986) Transitions from Authoritarian Rule:
Tentative Conclusions about Uncertain Democracies, Johns Hopkins
University Press, Baltimore
Perotti,
of
Roberto (1996) "Growth, Income
Economic Growth,
Paige, Jeffrey
1,
M.
in Central America,
Distribution and Democracy" Journal
149-188.
(1997) Coffee and Power: Revolution and the Rise of Democracy
Harvard University Press, Cambridge
John D. (1971) Political
Press, Cambridge MA.
Powell,
University
MD.
MA.
Mobilization of the Venezuelan Peasant, Harvard
29
Persson, Torsten and Guido Tabellini (1994)
Theory and Evidence," American Economic Review,
Prezworski,
Adam
"Is
Inequality Harmful for
84, 600-621.
(1991) Democracy and the Market, Cambridge University Press.
Rock, David (1987) Argentina 1516-1987: From Spanish Colonization
University of California Press, Berkeley
Roemer, John E. (1985)
and the Tsar," Econometrica,
in
New
to Alfonsin,
CA.
"Rationalizing Revolutionary Ideology:
A
Tale of Lenin
53, 85-108.
Roemer, John E. (1998) "Why
Argument
Growth?
the Poor do not Expropriate the Rich:
An Old
Garb," Journal of Public Economics, 70, 399-424.
Rodriguez, Francisco (1998)
"Inequality, Redistribution
and Rent-Seeking," Un-
published Paper, Department of Economics, University of Maryland at College Park.
Rodrik, Dani (1999) "Democracies Pay Higher Wages," forthcoming
terly
in the
Quar-
Journal of Economics.
Rueschemeyer, Dietrich, Evelyn H. Stephens and John D. Stephens (1992)
Capitalist
Development and Democracy, University of Chicago Press, Chicago
Rustow, Dankwart C. (1970)
Model," Comparative
IL.
Toward a Dynamic
"Transitions to Democracy:
Politics, 2, 337-363.
Sachs, Jeffrey D. (1990) "Social Conflict and Populist Policies in Latin America,"
in R.
Brunetta and C. DeH'Aringa eds. Labor Relations and Economic Performance,
Press,
New
York.
Skidmore, Thomas E. (1967)
Democracy, Oxford University Press,
Politics in Brazil,
Juan
J.
of
Democracy
Stepan, Alfred (1978)
ica,
J.
An Experiment
in
in Argentina, 1916-1930,"
Linz and Alfred Stepan eds. The Breakdown of Democratic Regimes: Latin
America, The Johns Hopkins University Press, Baltimore,
Juan
1930-1964-'
New York NY.
Smith, Peter H. (1978) "The Breakdown
in
NYU
"Political Leadership
MD.
and Regime Breakdown:
Brazil," in
Linz and Alfred Stepan eds. The Breakdown of Democratic Regimes: Latin Amer-
The Johns Hopkins
University Press, Baltimore,
Therborn, Goran (1977) "The Rule
MD.
of Capital
and the Rise of Democracy," New
Left Review, 103, 3-41.
Tornell,
Aaron and Andres Velasco (1992) "The Tragedy
Economic Growth:
Political
Why
Commons and
Does Capital Flow from Poor to Rich Countries?" Journal of
Economy,l00, 1208-31.
Trudeau, Robert (1993) Guatemalan
racy,
of the
Politics:
Lynne Reimer, Boulder CO.
30
The Popular Struggle for Democ-
Valenzuela, Arturo (1989)
Democratic Regime," in
L.
Diamond,
veloping Countries: Latin America,
Vilas, Carlos
M.
Origins, Consolidation
"Chile:
J.J.
and Breakdown of a
Linz and S.M. Lipset eds. Democracy in De-
Lynne Reimer, Boulder CO.
(1995) Between Earthquakes and Volcanos: Market,
the Revolutions in Central America,
Monthly Review
Press,
State,
and
New York NY.
Wallerstein, Michael (1980) "The Collapse of Democracy
in Brazil: Its
Economic
Determinants," Latin American Research Review, 15, 3-40.
Wantchekon, Leonard (1998)
bility:
The 1994
"Strategic Voting in Conditions of Political Insta-
Elections in El Salvador," Unpublished Paper, Department of Political
Science, Yale University.
Weyland, Kurt G. (1996) Democracy
without Equity: Failures of Reform in Brazil,
University of Pittsburgh Press, Pittsburgh PA.
Wright, Randall D. (1996) "Taxes, Redistribution and Growth," Journal of Public
Economics, 62, 327-338.
Yashar, Deborah
J.
(1997) Demanding Democracy: Reform and Reaction in Costa
Rica and Guatemala, 1870's-1950's, Stanford University Press, Stanford CA.
31
inn
Date Due
Lib-26-67
3 9080 01918 3588