equality, the disinterested government, and economic growth: · web view2010/10/18 · equality,...
TRANSCRIPT
111Equation Chapter 1 Section 1Equality, the Disinterested Government, and
Economic Growth:
The Case of China*
Daxing He
Ph.D. student
&
Yang Yao
Professor
National School of Development
&
China Center for Economic Research
Peking University
Phone: 86-10-62753103
Fax: 86-10-62751474
January 28, 2010
* We thank Uri Bram, Daniel Bromley, Michael Carter, Eward Friedman, Oded Galor, Yukon Huang, Phillip Keefer, John Knight, Mi Luo, Rui Mao, Gerard Padro-I-Miquel, Debraj Ray, and James Wen for their helpful comments and suggestions. An early version of the paper has been presented in seminars in Brown University, Harvard University, The New School, New York University, University of Oxford, Peking University, Trinity College, World Bank, Wuhan University, and Xiamen University. We thank the participants for their comments. Financial support provided by the Key Research Bases project 04JJD790002, the Ministry of Education, the People’s Republic of China is greatly appreciated.
Equality, the Disinterested Government, and Economic Growth: The Case of China
Abstract: We put forward a political economy explanation to China’s economic success in the
last thirty years. This explanation is centered on the concept of disinterested governments, i.e.,
governments that do not have consistently differentiated interests among the segments of the
society. In a repeated Stackelberg game of a government and two groups of citizens, we show
that under reasonable constraints on the returns to revolutions, a disinterested government
arises if and only if the two groups of citizens have relatively equal political powers that
determine their capabilities to initiate successful revolutions. We also show that the social
output is higher under a disinterested government than under a biased government. We
provide historical evidence to show that the Chinese government has been disinterested in the
last thirty years and attribute it to the absence of significant social inequality in China.
JEL classification: H11, O43, O57
Keywords: Equality, disinterested governments, authoritarian states
1. Introduction
B
etween 1978 and 2008, China’s national GDP recorded an average growth rate of 9.7% per
annum by its official statistics (NBS, 2009). Although more conservative estimations arrive at
lower growth rates,1 China is still one of the thirteen economies that managed an annual
growth rate of 7% or higher for more than 25 consecutive years after World War II (The Growth
Commission, 2008).
T
o a large extent, China’s economic success can be attributed to its almost relentless reform drive
toward a market economy. Although the reform has taken an unconventional path, its
destinations --- in both policies and institutions --- have been clearly converging to those
recommended by mainstream economics. From this point of view, China’s high economic
growth record has nothing miraculous to it (Perkins, 2004). However, this view leaves an
important question unanswered: if the standard economic theories are so right, why have not
most developing countries followed them? Or in other words, why has China been able to adopt
the right growth recipes? This begs a political economy explanation.
C
entral to this paper’s argument is that China’s success lies in its disinterested government. Here
the word “disinterested” is used in one of its three meanings in aesthetics, that is, a person is
unbiased by personal interests when he makes judgment, inquiry, evaluation, and the like on
objective existences.2 Therefore, a disinterested government is defined as a government that
takes a neutral stand when conflicts of interests among different social and political groups
arise. In other words, it is a government that does not consistently represent --- and is not
captured by --- any social or political groups in the society. This does not mean that such a
government is devoid of self interests; quite to the contrary, it can not only have its own
interests, but also be predatory toward society at large. The key is that its predation is “identity-
blind” in the sense it does not care about the social and political statuses of its particular prey.
For that, its policies are more likely to be selective but growth enhancing --- that is, allocating
1 For example, Young (2003) estimates that China’s overall GDP grew by 7% per annum between 1978 and 2000.2 The other two meanings, according to Rind (2002), are “uninterested” and “without self interests” when a person examines an objective existence.
1
resources by the levels of productivity rather than by group identities --- than those of a
government that consistently represents the interests of certain social or political groups.
T
he Chinese government has been disinterested in the last thirty years. A distinctive feature of
the Chinese transition from economic planning to the market is that the transition has been a
voluntary choice made by the ruling Chinese Communist Party (CCP)3 while transitions in other
former communist countries were largely forced upon them by political transformation (Boycko,
Shleifer and Vishny, 1996). The reform, however, has been a process of transferring power from
the CCP elites to ordinary citizens. The party has avoided being trapped by elite dominance. In
the meantime, government policies have been selective based on the premise of economic
growth, not the political or social influences of different groups. The Special Economic Zones
(SEZs), export-oriented growth model, accession to the World Trade Organization (WTO), and
the privatization of state-owned enterprises (SOEs) are but a few examples. In addition, the
government is willing to correct the policies that imply a mismatch between the treatment and
productivity of a certain group. The abandonment of the dual-track price system and fairer
treatment for migrant workers are two examples.
B
eing a disinterested government is a strategic choice by the CCP for its own benefits. We
identify social equality as the most important condition for the Chinese government to opt for a
disinterested position. Although economic inequality is not uncommon, China has a flat social
structure with no significant elite groups, thanks to a series of revolutions in the first half of the
twentieth century and the egalitarian rules implemented in its planning era. Compared with a
socially unequal society, a socially equal society is more likely to produce a disinterested
government. A difference between economic inequality and social inequality is that the latter is
related to identity while the former is not. Identity facilitates the formation and consolidation of
enduring special interest groups because identity is exclusive. A socially unequal society, then, is
often divided into two significant segments, a smaller group of elites and a larger group of
ordinary citizens. Often, the elites have overwhelming economic and political advantages over
ordinary citizens. In such a society, the government is more likely to become biased, i.e., to form
3 We will not distinguish between the Chinese government and the CCP because the latter is the only ruling party in the country.
2
an alliance with the elites and take advantages of ordinary citizens. In contrast, in a socially
equal society no group has overwhelming advantages over the others, so revolutions by the
disfavored groups become credible threats to the government if it does favor some groups of
the society. That is, it does not pay for the government to form an alliance with any group in the
society. Then a sensible choice for the government is to treat social groups equally and adopts
group-blind policies.
W
e formalize the above idea in a repeated Stackelberg game of a government and two groups of
citizens. Our model builds on three key premises:
(1) The government is autonomous in the sense that it is free to choose its allies.
(2) The government in an autocracy does not enjoy absolute power but faces challenges
from social groups.
(3) The government can differentiate social groups in its policies.
T
he key to understanding the first premise is to realize that the government in our model is itself
an independent group and needs to choose an ally from the two social/political groups. In the
case of China, this is equivalent to treating the top CCP leadership as a coherent group. To be
sure, there are fights within the top leadership, but they are not sufficiently serious to lead to a
collapse. The second premise holds for most non-democratic governments including the current
Chinese government. By definition, non-democratic governments do not obtain their legitimacy
from popularly consented selection rules. Military might does not guarantee power in the long
run. Indeed, the buildup of military forces is often elicited by the potential of upheavals in the
country. In China, the CCP might not have faced serious challenges in the initial years after it
took power; but by the mid-1970s, its legitimacy began to falter because of its failures in the
first twenty years of rule. In 1976 and 1989, two major democratic movements broke out. In
today’s China, numerous protests and even upheavals happen each year against the
government’s infringements on individual rights although political movements are at a relatively
low ebb. The third premise holds in both democracies and autocracies and is commonly found
in the literature on government behavior. For example, most of the papers to be reviewed later
in this section are built on this premise. In China, social groups are more commonly defined by
3
regions, urban-rural divide, and sectors. It is easy for the government to differentiate its policies
along these lines.
I
n our model, the two groups of citizens are distinguished by the amount of assets and the size
of political power that each group can rely on to wage a revolt against the government when it
is not the ally of the government or to retain power when it is the ally of the government and
faces the revolt of the other group. Here political power is the indicator for social inequality. The
government taxes both groups and uses the taxes for its own consumption and the provision of
club goods to the two groups. It is called a disinterested government if its policies (tax rates and
the provision of club goods) do not depend on the two groups’ political power. Adopting the
pure stationary Markov perfect equilibrium (MPE) as the solution concept for the game, we find
that a necessary and sufficient condition for the government to be disinterested is that the two
groups of people have relatively equal political powers. This is so because otherwise it pays for
the government to form an alliance with the group with greater political power and charge a
higher tax rate on, and provide a lower level of club goods to the other group without the
danger of being subject to a successful revolution.
T
he social output is higher under a disinterested government than under a biased government.
This is shown intuitively by noticing first that the best stage utility a group can obtain with a
successful revolution, i.e., a revolution that eliminates the current government and the other
group, is its autarkic utility --- that is, utility it can obtain when the government ceases to exist
--- because the new government installed after the revolution can survive by just giving the
victorious group that level of utility. When the two groups have relatively equal political powers,
both would start a revolution if they obtained a level of utility lower than their autarkic level. As
a result, the government gives each group its autarkic utility and the sum of outputs is the
autarkic social output. In all other cases, each group has a chance to initiate a successful
revolution and the final reward of the revolution (the autarkic output) is shared by them with
their probabilities of a successful revolution as the weights, albeit discounted by their
impatience. As a result, the sum of their outputs falls short of a convex composition of their
autarkic outputs. This result also implies that there exists an optimal social structure with a
4
relatively equal distribution of political powers under which the social output is the highest.
T
here is a rich theoretical literature relating retarded economic growth to inequality through
negative government policies. Most of it, however, studies democracies. For instance, Alesina
and Rodrik (1994), Benabou (2000), and Hassler, Rodriguez-Mora, Storeletien, and Zilibotti
(2003) study the relationship between the inequality of wealth and government’s redistributive
policies; Esteban and Ray (2006) demonstrate how inequality of wealth may distort government
allocation through asymmetric lobbying; Renzo (2007) focuses on how asymmetric political
powers may lead to short-sighted government policies; and Galor, Moav, and Vollrath (2009)
show how equal land distribution is conducive to growth by promoting human capital
accumulation.
E
steban and Ray (2006)’s work has direct bearing on our paper. They treat lobbying as a signal
sent by private agents to the government showing their productive worthiness for preferential
treatments. However, high wealth enables agents to launch louder lobbies. Because the
distribution of productivity does not match the distribution of wealth, inequality distorts
government allocation of resources even if it seeks to maximize economic efficiency. In a sense,
our model provides a parallel theory on autocracies. In Esteban and Ray’s model, economic
equality reduces the strength of lobbying serving as a signal for productivity, while
social/political equality reduces the strength of revolutions that forces the government to take
biased policies in ours. In both cases, equality makes the government more likely to allocate
resources to match the productivity of social groups. One of the differences, however, lies in the
fact that their model is built on the premise that productivity is private information to individual
agents and equality reduces the distortion caused by the government’s inability to obtain
accurate information, while our model assumes perfect information but shows that equality
improves efficiency by freeing the government from the fear of losing power. Another difference
is that we do not assume efficiency-maximizing government. The social output generated under
a disinterested government, therefore, is only the second best. Of course, the subjects of study
--- democracies in their study and autocracies in ours --- naturally make the two models
different in many other aspects.
5
O
ur model also shares the spirit of Galor, Moav, and Vollrath (2009)’s in that both ascribe
underdevelopment to unequal social structure. However, we stress the role played by the
competition among different social groups whereas they stress the role of a specific group, the
landed class, which is associated with a sector (agriculture) with a low demand for modern
inputs like human capital.
T
here is a small but growing literature on the performance of autocracies. While most studies
focus on the persistence of autocracies (e.g., Acemoglu, Ticchi, Vindigni, 2007; Padro-I-Miquel,
2007; and Acemoglu and Robinson, 2008), two recent papers have studied why some
autocracies have obtained better records of economic growth than others. Gehlbach and Keefer
(2008) observe that autocracies performed better in terms of economic growth when the ruling
party had a longer history. They interpret this finding as evidence for the positive role of party
institutionalization. Specifically, their theoretical model takes within-party information sharing
as the most distinctive feature of party institutionalization. Party members are informed of the
behavior of the leader and can punish the latter by obstruction. As a result, the leader becomes
less predatory on party members who then become more likely to invest in the economy.
Besley and Kudamatsu (2007) further provide a detailed documentation of the economic
performance of autocracies after World War II and put forward an explanation for successful
economic records. In their model, a selectorate comprised of a group of insiders selects the
leader. In an agency model, they show that when the selectorate’s power does not depend on
the leader’s being in office, its threat of replacing the leader becomes credible and the leader
chooses strategically to adopt growth-friendly policies. In a sense, this story is one of party
institutionalization, as proposed by Gehlbach and Keefer (2008).
H
owever, Besley and Kudamatsu (2007), together with Padro-I-Miquel (2007), have a close link
with our paper in terms of modeling. Padro-I-Miquel (2007) studies why some failing
autocracies could last for a long time. His explanation is one of “rule by fear”. In his model,
people in the ruling group have to deliberate between replacing a bad ruler and the repression
of the competing group if the internal struggle leads to the loss of power of the ruling group.
6
One of his key assumptions is that the probability of losing power is high when people in the
ruling group replace their ruler. In a sense, this is equivalent to the assumption that the two
groups of citizens have equal political power. In contrast, Besley and Kudamatsu (2007) are at
the other extreme. Their condition for a better autocracy, i.e., that the survival of the ruling
group does not depend on whom its leader is, is equivalent to the assumption that the ruling
group has superior power over the competing group. Put together, these two papers suggest
that equality between groups is actually bad for the society.
O
ur model differs from these two papers by treating the government as a third group and
focusing on between-group, instead of within-group, dynamics. Treating the government as an
autonomous group is consistent with the roles of the rulers in several autocracies (like South
Korea under Park Chae-Chi) as well as CCP’s role in China. Our result that between-group
equality helps growth differs from the results of Besley and Kudamatsu (2007) and Padro-I-
Miquel (2007). Those two papers treat between-group dynamics as the background of their
models and thus ignore some of their important implications. For example, Besley and
Kudamatsu fail to explain why the ruling group does not become excessively predatory on other
groups if it has absolute advantage over them. On the other hand, Padro-I-Miquel does not
address why the other group cannot force the ruler (and the ruling group) to behave better if it
can take any available opportunity to replace the ruling group. The reason that between-group
equality does not lead to a better government (ruler) in his model is that the competing group
cannot directly influence the behavior of the leader of the ruling group; instead, it undermines
the ability of the ruling group members to supervise their leader. While our model is not aimed
at replacing the arguments of those two papers, we do bridge the above gaps left by them.4
T
he rest of the paper is arranged as the follows. In Section 2, we provide the setup of our model,
which is solved in Section 3 to obtain the key results for the disinterested government and its
efficiency. In Section 4, historical evidence is provided to show how the Chinese government in
the last thirty years has behaved in a manner that is both disinterested regarding the interests
of different social groups and selective to promote economic growth. In Section 5, we conclude
4 A unifying model may be possible to accommodate both within-group and between-group dynamics. But for our current purpose, the model presented in this paper may be sufficient.
7
the paper with a discussion on some of the alternative explanations for China’s economic
growth.
2. Model setups
W
e start with a closed economy with discrete time and an infinite horizon. There are a
government and two social groups in the economy. We abstract from the complexities in the
government and simply assume that it is represented by one person. The two groups, labeled P
and E, have a measure of population of 1 and α, respectively.5 Individuals in each group are
identical. In each period, nature assigns each individual an endowment of capital whose amount
does not vary across periods. Let and denote the amount of capital that individuals in
Group P and Group E receive in period t, respectively. Capital is taxable, and the government
levies a tax directly on each individual which it can then use either for its own consumption or
for provision of services to individuals in the two groups. We assume that the government
cannot discriminate among individuals in the same group, but can do so between the two
groups. In other words, the government has to impose the same tax rate on and provide the
same level of services to individuals in the same group. Here we can take the services as club
goods.6 Let and be the tax rates and and be the levels of services for Group P
and Group E in period t, respectively. We then summarize the government’s policy in period t by
. Naturally, the sum of government services cannot exceed the sum of
taxes. That is,
.
E
ach individual deploys the capital left with him and the services provided by the government to
5 If not otherwise indicated, uppercase letters indicate groups and lowercase letters indicate individuals in the groups.6 For an example of such treatment, see Qian and Roland (1998), p.1147.
8
produce a uniform product by the following production function:
(1) , 0 < , and .
Outputs cannot be stored or invested, but can only be consumed in the current period. We
assume that individuals are risk neutral.
T
he government is also risk neutral and derives utility from the total amount of taxes, but incurs
disutility from its provision of services to individuals. So its net utility in period t is
(2) .
Obviously, this net utility cannot be less than zero, which we take as the government’s
reservation utility. The government and the individuals have a common discount factor
.
I
n any period, the government comes from one of the two groups of citizens.7 The group where
the government comes from is called the incumbent group (or simply the “incumbent”), and
the other group is called the challenger group (or the “challenger”). The statuses of the two
groups may switch over time. Also following Padro-I-Miqual (2007), we assume that the
incumbent has the veto power on policies issued by the government. That is, the government is
constrained by its own social origin. We adopt this assumption to introduce an asymmetry
between the two social groups.8 The assumption does not mean that the government cannot
switch loyalty. In our model, the government may favor the challenger if it is politically more
powerful than the incumbent. Likewise, the incumbent does not necessarily support the
government at all times. As we will see, the incumbent may use its veto power to induce the
challenger to start a revolution when it is in an inferior political position to defend its own
7 Admittedly, the way how the government is produced can be important for its behavior, as Besley and Kudamatsu (2007) and Padro-I-Miquel (2007) have rightly emphasized. Here we ignore this issue because we focus on between-group dynamics.8 In the Chinese case, there is an asymmetry between the working class, the social group that the CCP initially derived its legitimacy from, and the other social groups. The CCP states in its charter that Marxism is its guiding ideology and emphasizes workers’ welfare in its official documents. On the other hand, however, the CCP is becoming more and more pragmatic and its policies often do not favor workers, as SOE privatization has demonstrated (Section 4).
9
0,1
interests.
I
n period 0, Group P and Group E are each endowed with a certain level of capabilities of social
mobilization, or political power, . Both groups can use their political powers against
each other or the government. For example, the challenger can use its power to initiate a
revolution, while the incumbent can use its power to help the government suppress the
revolution. In subsequent modeling, we treat vP as the probability of success for the challenger
to initiate a revolution and link vE to the incumbent’s probabilities of survival and suppression
when the challenger initiates a revolution. These are the two most important parameters in our
model. When they are close to each other (in a sense to be precisely defined later), we say that
the society is socially equal; otherwise, the society is said to be socially unequal.
10
, (0,1)P E
Figure 1. The Stage Game
n each period, the government and the two groups play a Stackelberg game. Figure 1 depicts the sequence of actions of the game. The government acts first by announcing a policy
accept (A) or to reject (NA) the policy. Since the incumbent has veto power, the government ceases to function if rejection happens. In this case, each individual engages in autarkic production that uses only his own capital if the
challenger does not start a revolution. We denote the decision of the incumbent by
, and
After observing the government policy and the incumbent’s decision, the challenger decides
whether to start a revolution. We use R for revolution and NR for no revolution. If a revolution
happens, the government also stops functioning; in addition, production and consumption stop
for one period.9 We denote the decision of the challenger and its consequences by
.
9 This assumption is intended to capture the fact that revolutions are destructive.
11
Government,Gt
Incumbent,A
Incumbent,NA
Challenger,R
Challenger,NR
Challenger,R
Challenger,NR
A successful revolution may result in redistribution of capital and destroy the political power of
the incumbent. Details will be provided shortly when we present the dynamics of the game.
F
igure 2. Revolution and its consequences
R: revolution; NR: no revolution; S: revolution succeeds; F: revolution fails; L: losing political power; H: retaining political power.
T
he stage game is repeated in subsequent periods. Since output cannot be stored or invested,
the dynamic game is an ordinary repeated game. To discuss the transitory process between
periods, let us assume that Group E is the incumbent and Group P is the challenger in an
arbitrary period t. If P decides not to start a revolution, the game enters the next period. If it
starts a revolution, then the ensuing events are shown in Figure 2. The figure starts with P
12
starting a revolution (R). With probability the revolution would succeed (S), and with it
would fail (F). If the revolution turns out successful, Group P substitutes for Group E as the
incumbent and forms its own government. Group E has a probability of to retain its political
power, or to permanently lose it. If it retains power, Group E will become the challenger in
the next period. On the other hand, if Group P does not succeed in its revolution, the current
government remains in power. Group P may permanently lose its political power with a
probability , or retain its power with a probability ; in the latter circumstance, Group P
may start a new revolution again in the future. Here the critical assumption is that the political
power of the incumbent not only determines whether it could survive in a successful revolution,
but also determines whether the challenger could survive in an unsuccessful revolution. This of
course is a simplified assumption. What we want to capture is that the incumbent has an
advantageous stand over the challenger. In our model, the power of the government comes
solely from the power of the group it comes from. In reality, the government usually
monopolizes the use of violence, so its favored groups have some edges over the disfavored
groups. Our assumption reflects this asymmetry.
I
f no revolution happens in period t + 1, the game enters the next period; otherwise, it repeats
what Figure 2 shows. We make an assumption that once a group permanently loses its power,
the other group seizes the capital of the members of the losing group and distribute it equally
among its own members. That is, the losing group no longer exists. In this case, the new
government will give the winning group its autarkic utility because the winning group will reject
any lesser level. Naturally, this is a Markov equilibrium. The utility level of the winning group is
thus if it happens to be Group E, or
13
P 1 P
E
1 E
E 1 E
if it happens to be Group P.
T
o avoid the collective action problem,10 we assume that each group has a charismatic leader
whose decisions the group members follow without any doubts. This leader shares the same
utility function of any other individual in the same group. Therefore, we will not distinguish
between the leader with either the group as a whole or any group member.
3. The existence of the disinterested government and its efficiency
F
ollowing the literature (e.g., Padro-I-Miquel, 2007; Renzo, 2007), we adopt the pure stationary
Markov perfect equilibrium (MPE) for our solutions. Let , denote the status of political power
of a group in period t such that
Let denote the relationship of a group with the government in period t such that
Then the state of the society in period t can be summarized by
. The government’s policy decision on tax rates and service
provision is , the incumbent’s decision to accept or reject the government policy is
10 See Padro-I-Miquel (2007), p.1263 for discussions of this issue.
14
, and the challenger’s decision whether to start a revolution is
. The probability transformation function is . Let
denote the repeated Stackelberg game defined in Section 2. The joint
solution of the following three Bellman equations for the government and the two groups
characterizes the solution to the game:
(3) ,
(4) ,
(5) .
Let the joint solution be denoted by .
Definition 1. Pure stationary MPE 11
S
uppose that the incumbent is group E in period t. If , , and
, then is a pure stationary MPE for
.
3.1 Revolutions and utility levels
W
ithout loss of generality, let us continue to assume that in an arbitrary period t Group E is the
11 See Padro-I-Miquel (2007) for a similar definition.
15
incumbent and Group P the challenger. In fact, focusing on the pure stationary MPEs makes the
starting period irrelevant in our analysis. Therefore, we will omit the subscript t in subsequent
analysis. Then we start our analysis by determining the levels of discounted utility of the two
groups and the government under different scenarios. For that, let , j = E, P, i = p, e, G,
denote the expected utility at any given period for the challenger, the incumbent, and the
government when the incumbent is Group E or Group P, respectively.
T
he first scenario is when Group E, the incumbent, accepts the government policy, and Group P,
the challenger, does not start a revolution. Then the discounted utility of Group P, Group E, and
the government in period t are, respectively,
(6) ,
(7) , and
(8) ,
in which the superscript (A,NR|E) denotes the situation that Group E accepts the government’s
policy and Group P does not start a revolution given that Group E is the incumbent. This rule of
notation will be applied throughout all the other scenarios. When it is used separately in the
superscripts, E denotes that the incumbent is Group E. All the subscripts denote the players.
T
he second scenario is when Group E accepts the government policy, but Group P starts a
revolution. This case is more complicated than the first scenario. Revolutions stop production so
everyone’s utility is reduced to zero. If the revolution is successful, then following our setups in
Section 2, Group P either takes all the capital of Group E (so each member has amount
of capital) in the case that Group E loses its political power, or becomes the incumbent under
the circumstance that Group E does not lose its political power. So Group P enjoys either the
16
discounted utility under autarky, , or the discounted utility when it becomes the
incumbent, . If the revolution is unsuccessful, Group P is either expelled with its capital
stripped off by Group E, in which case its members’ utility level drops to zero forever, or remains
as the challenger in which case the game starts again and everyone in the group continues to
have an expected sum of discounted utility of . Therefore, the discounted utility of
revolution for Group P when Group E is the incumbent, , is
(9) .
Similar analysis arrives at the following utility for Group E under revolution when it is the
incumbent:
(10) .
And the government’s utility is
(11) .
The third scenario is when Group E rejects the government policy and Group P does not
start a revolution. In this case, both groups fall back to autarky for one period and then start
over again in the next period. So the levels of utility of their members are, respectively,
(12) , and
(13) .
The government stops functioning for one period, so its discounted utility is
(14) .
T
he last scenario is when Group E rejects the government policy, but Group P starts a revolution.
Since revolutions determine the final outcomes, the three players’ utility levels are those
17
defined in equations (9) – (11), respectively.
T
he cases when Group P is the incumbent in period t are similar and we do not repeat them
here.
3.2 The existence of the disinterested government
T
he nature of the stage game gives the government first-mover advantage. Comparing the
government’s utility levels in (8), (11), and (14), we find that the solution to the game is boiled
down to the government’s maximization problem subjected to the constraints of no rejection or
revolution. Our first task is thus to find the minimal utility --- the reservation utility as we will
call it --- that each group should get in order to induce a response of neither rejecting nor taking
a revolution. We call the two cases the government wants to avoid “the exit points”, that is, the
points at which the current government ceases to function. To determine the two groups’ levels
of reservation utility, we need to consider four cases:
(15) , and
,
(16) , and
,
(17) , and , and
(18) , and .
In the case of (15), both groups obtain higher levels of utility when the exit point is revolution
than when it is rejection, so their reservation utility levels are and , respectively. To
understand why the government should give the incumbent its utility in a revolution, notice
that the second inequality does not rule out the case > , that is, the incumbent
obtains a higher level of utility when a revolution happens than when it accepts the government
policy and no revolution happens. In this case, the incumbent will induce a revolution by
18
intentionally rejecting the government policy. By the first inequality of (15), the challenger will
indeed revolt. In the case of (16), the levels of reservation utility are still and . The
reservation utility of the incumbent is , not , because the incumbent’s rejection of
the government’s policy will only induce the challenger to revolt. By parallel reasoning, we
know that the two groups’ reservation utility levels are and , respectively, in the
cases of (17) and (18).
T
he government does not need to give the two groups levels of utility higher than their
reservation utility. In the case of (15) and (16), then we have , and
. The latter result implies . That is, revolting in
one period is better than autarky in all periods. In other words, revolution is a credible threat
that deters the government from giving the challenger only its autarkic utility. Therefore, the
government’s constraints in the case of (15) and (16) can be summarized as
(19)
, and .
Similarly, the government’s constraints in the case of (17) and (18) can be summarized as
(20)
, and .
After determining the reservation utility of each group, the decision of the government can be
described by the following program: 12
12 See similar treatments in Klein, Krusell, and Rios-Rull (2008) and Padro-I-Miquel (2007).
19
(21)
When the incumbent is Group P at the beginning of the game, the government’s program is
similar.
To facilitate the exposure, we define a group’s revolution utility as its life-time discounted
sum of utility when a revolution happens in the current period, and a group’s autarkic utility as
its life-time discounted sum of utility when it engages in autarkic production in its life time. With
that we present the following proposition:
Proposition 1: The social dominance of autarky over revolutions
The sum of the two groups’ revolution utility is smaller than the sum of their autarkic utility.
Proof: In the appendix.
The intuition of this proposition is the following. The final prize of a successful revolution for
either group --- that is, the one-period aggregate utility (output) that one group can have from a
revolution that seizes the capital of the other group --- converges to the social output at autarky
, which is also the sum of the stage utility of the two groups under autarky. However,
this prize must be (virtually) shared by the two groups because both of them have some
chances to start a successful revolution. Their shares are otherwise proportional to their
respective probability to exterminate the other group if they are perfectly patient. As a result,
the sum of the two groups’ revolution utility is smaller than the autarkic social output when
people are not perfectly patient.
A
natural extension of Proposition 1 is that the two groups’ revolution utility cannot both be
equal to or higher than their autarkic utility. Then, the stage utility of the two groups under a
revolution cannot both be equal to or higher than their autarkic stage utility.
A disinterested government is formally defined as:
Definition 2: The disinterested government
20
A disinterested government is a government whose policies are identity-blind. Specifically,
the government’s tax rates and the levels of club goods do not depend on the political powers of
the two groups. In a pure stationary MPE, this can be expressed as:
Opposite to a disinterested government is a biased government whose policies depend on
the two groups’ political powers. Under the disinterested government in definition 2, an
individual’s stage utility is only related to one’s endowment. Denote this utility by . It is
equal to
. That is, each individual has the same function form for
his stage utility under a disinterested government.
It can be shown that the program in (21) has a unique pure stationary MPE. Instead of
giving a full characterization of this equilibrium, however, here we are only interested in the
necessary and sufficient conditions under which a disinterested government emerges from the
equilibrium.13
Proposition 2: The existence of the disinterested government
Suppose that the incumbent in period 0 is Group E. Then there exists a unique pure
stationary MPE in which the disinterested government emerges if and only if the distribution of
political power of the society satisfies the social equality condition
(22)
.
In addition, we have
a. The government adopts the policy , and obtains the utility of
13 Interested readers can contact the authors for a full solution of the program.
21
in each period;
b. Group E (the incumbent) accepts the policy and each of its members obtains his
autarkic utility in each period;
c. Group P (the challenger) does not revolt and each of its members also obtains his
autarkic utility in each period.
P
roof: In the Appendix.
T
o provide the intuition for this proposition, we first explain why equation (22) is called the social
equality condition. To start with, it is easy to show that and increase in each other. This is
easy to understand because one group has to become stronger when the other group does.
Less understandable is that decreases in and ke/kp, or put in the other way, increases in
and kp/ke. Here the claim has a lot to do with our assumption that the size of the population is
irrelevant to a group’s political power. Then a larger population and a relatively larger stock of
per-capita capital only increase the level of a group’s sum of wealth, which in turn increases the
benefits of the other group’s tempted revolutions. When the sum of capital of one group equals
the sum of capital of the other group, has to be larger than by a gap of , a
premium that Group E enjoys over Group P by its initial position as the incumbent. This gap
vanishes when approaches 1, that is, when individuals do not discount future utility.
T
hat each individual of the two groups gets his reservation utility ki in each period follows from
the constraints in (20). This is straightforward for Group E because the constraints say that the
government just gives this group its autarkic utility ke. For Group P, the constraints imply that it
22
cannot get utility larger than its autarkic utility kp. On the other hand, it gets utility less than kp
only if it has a smaller political power defined in (22), in which case its threat of revolution is not
credible.
T
he social output of each period is , namely, the autarkic social output. This does not
mean that the government does not tax the two groups. As result (a) in Proposition 2 shows,
the government does tax people and in turn provides services to them. By doing this, it confers
a positive gain while giving the individuals their autarkic utility.
I
t is worth noting that the amount of services the government provides to each group is
proportional to its members’ capital endowment whereas the tax rate is the same for both
groups. Therefore, a disinterested government does discriminate between the two groups by
their stocks of productive assets. That is, its policy is identity-blind, i.e., free from the two
groups’ political powers, but is not efficiency-blind. The two groups make peace with different
levels of utility because their political powers are close to each other so neither group can force
the other group to get less than what it could get in autarky.
3.3 The efficiency of the disinterested government
It turns out that the social output under the disinterested government is the highest level
that the society can achieve.
Proposition 3: The efficiency of the disinterested government
The society yields the highest aggregate output in each period under the disinterested
government defined in Proposition 2.
Proof: This proposition follows Propositions 1 and 2. Proposition 2 establishes that the
disinterested government is unique, so we only need to compare this disinterested
government’s social output with those of the biased governments. For that, we notice first that
the aggregate output is equal to the sum of utility of individuals in the two groups. In addition,
the utility of each individual is the same in any period under a stationary MPE. This allows us to
convert the comparison to one comparing the discounted sum of utility under the disinterested
23
government and that under any biased government. Then we can consider three cases. The first
is the trivial case where both groups obtain utility less than their autarkic utility under a biased
government, which automatically establishes the proposition. Second, if (20) holds, the
incumbent produces its autarkic output, and the challenger produces an output no more than
its autarkic output under any kind of government. Therefore, social output is the highest under
the disinterested government following Proposition 2. The third and less obvious case is when
(19) holds, in which case the challenger produces an output higher than its autarkic output
while the incumbent does not. However, both groups obtain their revolution utility in this case,
and by Proposition 1, the sum of their outputs (utility) cannot be larger than the sum of outputs
in autarky. Q.E.D.
Because Proposition 2 shows that the disinterested government can only emerge under
the social equality condition (22), Proposition 3 implies the following proposition regarding the
optimal social structure:
Proposition 4. Optimal social structure
The social output is the highest when the distribution of political power satisfies the social
equality condition (22).
Figure 3. Social outputs under different social structures
24
Notes: The parameters are kp = 1, ke = 0.6, α = 1, β = 0.9.
In a sense, the social equality condition defines a knife-edge equilibrium for the optimal
social structure and the disinterested government. However, there exists a region in which the
social output gradually increases toward the output under the setting of optimal social structure
and the disinterested government. Figure 3 shows this under one set of parameters. There is
clearly a plateau of outputs. This plateau runs from northwest to southeast. Over the plateau,
the incumbent gets its autarkic utility and the challenger gets its revolution utility, which
increases as the plateau moves southeast. At its southeastern edge (denoted by DG), both
groups get their autarkic utility and the social output reaches its highest (equals to 1.60). Except
for its south end, changes in the social output are moderate on the plateau because both
groups have relatively equal political powers.
The two edges of the plateau are high cliffs. Beyond the northwestern cliff, the incumbent
still gets its autarkic utility, and the challenger still gets its revolution utility, but the latter drops
significantly. When the incumbent’s political power is small, this decline is only caused by the
decrease in the challenger’s own political power. When the incumbent’s political power
becomes larger, the probability of the challenger being completely eliminated by the
incumbent’s revenge against the former’s revolutionary attempts increases, and ultimately
approaches one, in which case the challenger’s utility drops to zero. This is why the cliff is not
significant at its southern tip but becomes more pronounced when it moves northward. Beyond
the southeastern cliff, the challenger gets its autarkic utility, whereas the incumbent can no
longer sustain its autarkic utility but instead gets a lower revolution utility. As the challenger’s
power increases, however, its chance of securing a complete revolution depriving the
incumbent of its assets increases, so the challenger’s utility (and social output) increases. But
because it is a future gain, this utility is always subjected to a discount and never surpasses the
sum of utility (social output) under the disinterested government.
The plateau can be fully quantified; we have analytical solutions to its edges, government
policies, and social outputs. But our main point has been made clear in the last paragraph
without presenting those results.14 That is, the disinterested government (and the optimal social
structure) is not isolated; the plateau allows the society to move toward it continuously from
14 Interested readers can approach the authors for those results.
25
one side with the social output increasing gradually.
4. The disinterested government and economic growth in China
I
n this section, we apply our theoretical model to explain China’s economic success in the last
thirty years. Our first task is to provide evidence that the Chinese government has been
disinterested in the last three decades. For that, we will concentrate on two real-life
implications of Propositions 2 and 3. First, a disinterested government adopts selective and
growth-enhancing policies; second, a disinterested government corrects incongruence between
its policies and the levels of productivity of different social groups. Our second task is to link the
disinterestedness of the Chinese government with China’s relatively equal social structure. This
link is implied by Proposition 2 and 4.
B
efore proceeding, we first identify our unit of analysis. Our focus is the central government, or
more specifically, the politburo of the CCP. On fiscal terms, China is a decentralized country and
local governments have much freedom to decide what they want to do with their local
economies. Nevertheless, the central government can largely control local governments
through the centralized political system (Xu, 2008). The current politburo consists of 25 core
CCP members who occupy the top positions in the party, the central government, and the four
major cities, Beijing, Shanghai, Tianjin, and Chongqing. Within the politburo, the standing
committee of nine people is the most powerful and makes the final decisions, but the other
members, with various degrees of capacity, also have considerable discretional power and can
influence the decisions of the standing committee. Naturally, the incumbent social group that
the politburo comes from is the 68 million ordinary CCP members. But for the ease of narrative,
we will often refer to the politburo just by the CCP. Other social groups are delineated by
geography (e.g., rural versus urban and coastal versus inland), sectors (e.g., state-owned
enterprises versus private firms, natural monopolistic sectors versus competitive sectors, etc.),
occupations (e.g., workers versus managers, migrants versus locals), income (the rich versus the
poor), and other factors that assign benefits and costs to the population in certain government
26
policies and reform measures.
W
e realize that a complete account of the Chinese experience perhaps requires a book. Therefore,
we will focus on using examples to show that the CCP has been able to resist the pressures from
both the elites --- including its own senior members --- and the populace, and adopt selective
but growth-enhancing policies in the process of institutional transition and economic
development during the period 1978-2008. Those policies inevitably created winners and losers,
but the configuration of winners and losers changed over time and their impacts were limited
on government policies. The CCP is able to do so because the Chinese society has been made
equal by a series of revolutions in the first half of the twentieth century. Under this equal social
structure, there have been no interest groups emerging that are either capable of capturing the
government or strong enough for the government to rely on.
F
inally, we notice that to apply our theory to the Chinese case, we need the assumption that the
CCP is as pragmatic as the government assumed in our model. The Chinese society had been
made equal shortly after the CCP obtained power in 1949, but the first thirty years of the CCP’s
rule were mostly characterized by political turmoil and economic disasters. Although there were
many causes leading to those dire consequences, Mao Tsedong’s radicalism was by and large
the most important cause. Therefore, our first task in the subsection below is to show that the
CCP had become pragmatic when reform started at the end of the 1970s.
27
4.1 Evidence of the disinterested government
T
he “growth consensus” formed at the end of the 1970s opened the door for the CCP to
transform itself from a radical party to a pragmatic party. This consensus came out of the CCP’s
realization that continuing with the Stalinist socialism would not only retard China’s rise in the
world, but also threaten the CCP’s legitimacy as the ruler of the country. The essence of this
consensus is to put economic growth at the center of all government and societal endeavors.
For that, the CCP quietly dropped its commitment to an egalitarian society and began --- in
Deng Xiaoping’s words --- “to let some people get rich first”.
T
he growth consensus has accomplished several things. The first was that it unified the party and
the country. After the dark years of the Cultural Revolution, people got tired of the political
rhetoric and needed a new direction to head for. But the radical elements still controlled the
party after Mao’s death. Shifting the party’s emphasis to economic growth preempted the
objection of the radical camp. The second accomplishment was in the area of system changes.
Since the old Stalinist model did not work, new models had to be introduced. Once the
objective was set, institutions became a tool to realize it. The third accomplishment was related
to the change of the CCP’s ideology. Starting with allowing some people to get rich first, the
growth consensus has watered down the CCP’s old convictions. After thirty years of evolution,
today the CCP is hardly a party built on political ideology. Pragmatism seems to be the only
principle that it adheres to.
W
ith a pragmatic attitude, the CCP was ready to become disinterested with regard to the society.
Next, through a narrative of examples, we will provide evidence for the first implication of our
theory regarding the disinterestedness of the Chinese government, that is, it tends to adopt
selective but growth-enhancing policies.
W
e start with China’s open-door policy. China’s opening to the outside world started with the
creation of the SEZs. This move was selective because the SEZs enjoyed many preferential
28
treatments that other parts of the country could not even imagine (Ge, 1999). But the SEZs
were critical for China to reach the outside world and learn to run a market economy at that
time. The export-led growth model was adopted in the mid-1980s and led China away from the
even development strategy in the planning period to an uneven development strategy that
focused on the growth of the coast. The coast’s share of the central government’s investment
increased from 39.5% in the period 1953-1978 to 53.5% in 1979-1991. It dropped a bit in the
1990s, but picked up again to 52.9% in the period 1999-2005 (Yao, 2008). Today, nearly 90% of
China’s exports come from the nine coastal provinces (Tong, 2008). The coast enjoys
geographical, historical, and technological advantages over inland provinces and the central
government’s choice conforms to the logic of the new economic geography. China’s integration
into the world system culminated in 2001 when China, after 13 years of Marathon negotiations,
finally joined the WTO. It was widely believed before China’s accession that China would have to
undergo painful structural adjustments in agriculture, automobile industry, and services if it
joined the WTO. Amidst the debate, the central government actually sped up China’s
negotiations with the WTO members, especially the United States. Accession to the WTO has
greatly accelerated China’s growth of exports, which is one of the most important engines for
the country’s growth and employment. Exports grew by an annual rate of 28.9% between 2002
and 2007 whereas the rate was 14.5% in the previous decade (Tong, 2008).
T
he privatization of SOEs witnessed another dramatic episode of the CCP’s transformation from a
radical party to a pragmatic party whose major aims revolve around economic growth. Between
1995 and 2004, the period when privatization was at its highest tide, the SOE sector reduced
40% of its employment. Nearly 50 million SOE workers lost their jobs; more than 20 million
workers lost their jobs in 1998 alone.15 Although the large-scale unemployment was not caused
by privatization, but rather came as a result of worker redundancy in the SOEs,16 both public
protests and academic debates pointed fingers at privatization. This put the CCP in a
conundrum: if it supported privatization, it would lose support from its own power base --- the
working class; if it gave up privatization, transition to the market would came to a halt and China
15 All figures come from Shen and Yao (2008), Chapter 4.16 Huang and Yao (2007) even find that privatization slowed down employment retrenchment due to privatized firms’ better performance than SOEs.
29
would lose the chance to catch up with the developed world. The solution the CCP adopted to
solve this conundrum was to quietly continue privatization and in the meantime to do
everything possible to reemploy laid-off workers. This strategy has worked. By the mid-2000s
when SOE privatization moved to its end, most of the unemployed workers had found new jobs
or had been put on government welfare programs.
P
erhaps, the most controversial case of the Chinese government’s selective policies is the urban-
rural divide. With urban per-capita income standing at 3.5 times rural per-capita income,17 China
has the largest urban-rural income differential in the world. While there are many causes for
this large urban-rural divide, 18 central to our argument are two of them. One concerns the role
played by the gap of political power between the urban and rural residents, and the other
concerns the role played by the productivity difference of the city and the countryside. It is
undeniable that rural people enjoy less political power than their urban counterparts in many
aspects. For example, four rural citizens are needed for every one urban citizen needed to elect
one delegate to the National People’s Congress (NPC), China’s legislative body, and today rural
migrants are still denied permanent residency and related benefits in their recipient cities.
However, how the uneven distribution of political power affects government policies is still
unclear. Delegates to the NPC only meet once a year and their role in legislation and supervision
are limited. The power --- if the NPC has any power beyond just a rubber stamp --- rests with
the standing committee of the NPC which is mostly comprised of retired government officials. In
reality, the central government does respond to the needs of rural people. The recent New
Socialist Countryside Movement is an example. Within this movement, a new cooperative
medical care scheme has been established, nine-year education has been made free, and
investment in infrastructure has been increased. The lack of real democracy does enable the
CCP to ignore rural people’s numerical advantage, which perhaps is one of the critical factors
that distinguish China from a democratic country such as India where the urban-rural divide is
moderate. But urban people also lose many advantages under the Chinese regime --- for
example, they cannot utilize their advantage of geographic concentration to initiate successful
17 Unless otherwise indicated, figures in this paragraph are from the official website of the National Statistical Bureau of China: www.stat.gov.cn. 18 For a comprehensive treatment of this issue, see Riskin, Zhao, and Li (2002).
30
lobbying. On the other hand, the productivity gap between the countryside and the city may
better explain why the government concentrates investment in the city. As the recent literature
on the new economic geography has convincingly shown, cities have strong agglomeration
effects. Just like the concentration of the central government’s investment along the coast, the
concentration of its investment in cities is also an efficiency-sensitive policy.
N
ext, we provide two examples to validate the second implication regarding the disinterestedness
of the Chinese government, that is, it should be willing to correct mismatches between
government policies and group productivity. One example is the dual-track price system. This
system was a compromise reached in 1984 between radical abandonment of and continuous
adherence to socialist planning. Under this system, SOEs were given the opportunity to sell their
products and buy inputs in the free market after they fulfilled their planned quotas. The market
prices were higher than the quota prices. The dual-track price system had opened up a wide
door for economic incentives to play a role in SOEs’ decision making. Lau, Qian, and Roland
(2000) believe that this system brought Pareto improvements to China. However, the dual-track
system also had serious downsides. One of them was that the gaps between the market prices
and the quota prices, sometimes extraordinarily high, created huge leeway for rent seeking.
Enterprises and government officials who controlled the quotas of key inputs could easily get
rich by selling their quotas to other enterprises and individuals. That is, a strong interest group
comprised mostly of top CCP members and SOE managers was created by the system. However,
the dual-track system also created, though unintentionally, new elements that were a result of
the market track only. The township and village enterprises (TVEs) and private firms were
among these elements. They did not have access to planned resources such as bank credits and
key inputs, and solely relied on the market to survive. Nevertheless, they became important
players in the Chinese economy by the early 1990s. For example, TVEs contributed 40% to
China’s industrial growth and 40% of its exports (Lin and Yao, 2001). That is, the new elements
and the old elites were not treated equitably in terms of their contributions to the national
economy. Our model then predicts that this would not persist under a disinterested
government. This was indeed what happened in reality; the dual-track price system was
abandoned in 1994, noticeably a time when the beneficiaries of the dual-track price system
31
were mostly insiders of the CCP, i.e. the incumbent group.
G
overnment policy toward migrant workers is another example. Free mobility of labor was
hampered by various barriers created by the central and local governments in the 1990s. The
household registration, or hukou system, has been in place since 1958.19 Many local
governments had set up rules to protect the jobs of local workers,20, while migrant workers
were often treated as second-rate citizens deprived of basic medical and safety protections.
They had to work long hours but were not guaranteed to get their due salaries. However, it had
become more and more clear that migrant workers were indispensible in the national economy
after China became “the world’s factory” at the end of the 1990s. In the meantime, grassroots
rights movements had joined hands with the intelligentsia to fight for better treatment for
migrant workers. Here the two social groups in the dispute were urban dwellers and rural
migrant workers. Urban dwellers gained undue benefits by the government’s suppression on
migrant workers’ rights, and migrant workers got less than what they had contributed to the
national economy. Grassroots rights movements and intellectuals’ petitions were equivalent to
the revolution in our theoretical model. They were perceived by the CCP as elements leading to
social unrests that would undermine its legitimacy. Against this background, the Hu Jintao-Wen
Jiabao government, upon its institution in 2003, began to abandon most of the discriminatory
policies toward migrants. This policy change can be viewed as being forced by migrant workers’
equal political power to threaten the CCP’s rule as urban dwellers’ power to protect its rule.
4.2 Social foundations for the disinterested government
I
n this subsection, we link the disinterestedness of the Chinese government to China’s equal
social structure. China’s modernization drive started in the mid-1800s when the imperial China
had to open its door to foreign powers. The 1911 Xinhai Revolution ended the Manchurian rule
of imperial China and established a republic. However, the social structure was more or less
kept intact. In particular, the landed class still dominated the vast countryside, and the warlords
19 The first constitution of the People’s Republic of China, announced in 1954, stipulated that the citizen has the freedom of migration (Article 90). The current constitution, announced in 1982, eliminates this article.20 See Zhao (2005) for a review of China’s migration policy.
32
and their crony industrialists monopolized the economy. It took the revolution of 1949 led by
the CCP to level out Chinese society. Through a thorough yet in many cases brutal land reform,
the landed class was effectively eliminated and land distribution was equalized. Large monopoly
businesses originally tied to the Kuomintang government were nationalized. Later in the
Socialist Transformation Movement of 1956, the remaining private businesses were either
nationalized or forced into state-private joint ventures. In the next twenty years, the Chinese
society was further equalized. Peasants were organized into communes and urban dwellers
were employed either by state-owned enterprises or collective enterprises. Accumulation of
productive wealth was made impossible and people were only paid by their labor. As a result,
the Gini coefficient of per-capita income was only 0.28 by 1978 (Riskin, Zhao, and Li, 2002).
Despite a high barrier between the countryside and the city, social mobility existed within the
countryside and within the city, respectively.
T
he first thirty years of the People’s Republic were mainly characterized by economic failures,
political turmoil, and human tragedies. Nevertheless, the equal social structure, although it was
largely a result of ill-conceived social engineering plans, had laid a foundation for the pragmatic
CCP leadership in the reform era to become disinterested relative to the society. Most
importantly, the absence of social classes reduced the CCP’s incentives to rely on any classes to
maintain its power. In its long struggle with the Kuomintang, the CCP gained power by relying
on the working class, especially the landless peasantry. It continued to emphasize its class roots,
and class struggle poisoned every corner of the country and every aspect of people’s life in the
first thirty years of the CCP’s rule. The new pragmatic leadership in the reform era realized the
dire consequences of class struggle, and in the course of reform has totally changed its political
outlook from a revolutionary party to a party with no clear political convictions (Wang, 2006). In
the meantime, the party’s membership has increased from 35 million in the early 1980s to 68
million today. Economic liberalization has not weakened, like some theorists would predict (e.g.,
Nee, 1989; Nee and Lian, 1994), but has strengthened the party’s rule. This has been made
possible because the party has aligned its own interests with, and tied its rule to economic
growth.
33
4.3 Related evidence from Taiwan and South Korea
T
he linkage between an equal social structure and the government’s neutrality with regard to the
society is not confined to China. As Woo-Cumings (1997) notes, governments in Taiwan and
South Korea were relatively free to adopt economic policies that enhanced the two economies’
long-term growth prospects in their early stage of economic development because the two
societies were made relatively equal by the Japanese colonists between 1895 and 1945. On the
one hand, Taiwan and Korea were designated as suppliers of agricultural goods in imperial
Japan’s version of the Great East Asian Commonwealth so urban industrialists were suppressed
in those two places. On the other hand, the Japanese colonists intentionally restricted the
growth of the landed class in both places because they feared that this class would become a
brewer for nationalist sentiments and organized upheavals against their colonial rule. “This
discontinuity had a powerful leveling effect, equalizing incomes more than in most developing
countries and providing a fertile ground for instituting effective interventionist states, which
were given a relatively free hand to forge a developmental coalition as they saw fit.” (Woo-
Cumings, 1997; p. 331).
T
he contrast between what Chiang Kai-shek could do in mainland China and Taiwan shows clearly
how an equal social structure can lead to selective but growth-enhancing policies. When he was
in the Mainland, Chiang had to rely on the landed class, large industrialists, and warlords to
maintain the rule of the Kuomintang, especially in its fights with the communists who
represented landless peasants and industrial workers. As a result, his policies were constrained.
Land reform is a case in point. As soon as Chiang consolidated power in the spring of 1927, the
Kuomintang government began to enact two laws, the Tenant Protection Law and the Tentative
Regulations on Tenancy, both aiming at restricting rents and protecting tenant rights in tenancy
contracts. However, they were put into effect only in 1932. There were two waves of
government efforts trying to implement those two laws in the 1930s, but both failed. While the
laws set an upper limit of 37.5% of output for land rents and outlawed fixed contracts and
prepayments, a nationwide survey found that the rent in most tenancy contracts was higher
than 50% and fixed contracts and prepayments were common (Yang, 2005).
34
T
aiwan was different from the Mainland in two key regards. One was that Chiang and his
Mainland followers were outsiders who did not have any social ties with Taiwanese people,
which made them free of the interference of special interests. The other was that the
Taiwanese society was more equal than the Mainland, which rendered unnecessary Chiang’s
attempt to rely on any groups for supports. One of the indicators that Chiang was unconstrained
in Taiwan was that he ordered Chen Cheng, the governor of Taiwan, to start land reform even
before he moved to Taiwan in early 1949. In addition, the Kuomintang government, like its
communist counterpart in the Mainland, adopted a policy of accumulating initial industrial
capital by extracting from agriculture. Indeed, the Taiwanese industrial sector extracted more
surpluses from agriculture through the price scissors than its Mainland counterpart in the
1950s. Between 1951 and 1960, 27% of Taiwan’s capital formation was contributed by net
capital outflow from agriculture in the form of price scissors.21 In the same period, however, the
Mainland government was actually paying higher prices for agricultural products than the
market (Wu, 2001).
5. Alternative explanations
O
ur explanation for China’s economic success is centered on the concept of disinterested
governments and we link it to China’s equal social structure. In the recent literature, Yingyi Qian
and his coauthors (e.g., Qian and Weingast, 1997; Qian and Roland, 1998), Naughton (2004) and
Jefferson and Zhang (2008) also provide political economy explanations to China’s economic
success.
Q
ian emphasizes the role of fiscal decentralization as a commitment device constraining the
behavior of the central government and a competition device constraining the behavior of the
local governments. But there are criticisms in the literature that Qian overemphasizes the role
of fiscal decentralization. For instance, after comparing the performance of fiscal
21 Calculated from estimates provided by Table 3 in Lee (1971), p. 29.
35
decentralization in China and Russia, Blanchard and Shleifer (2000) believe that fiscal
decentralization alone is insufficient to generate market-preserving and growth-enhancing
policies and China’s success has a lot to do with its political centralization. Along this line, one
can also ask the question how fiscal decentralization is sustained in China because political
centralization, at least in theory, gives the central government power to recentralize the system.
Indeed, the 1994 fiscal reform did lead to some degree of recentralization (Zhang, 2008).
N
aughton provides an explanation based on the CCP’s ability to maintain an effective
bureaucracy. Compared with most other developing countries, the Chinese bureaucracy is
indeed more effective in many aspects. In particular, it is very capable of carrying out orders
related to economic growth, for two reasons. First, economic growth brings higher income to
government officials because their incomes are tied to the growth records of their jurisdictions.
Second, economic growth is the most significant among factors determining the promotion of
government officials (Li and Zhou, 2005). It is undeniable that an effective bureaucracy has
helped China’s economic success. However, we have doubts whether it is a fundamental cause
for China’s economic growth. China has a long history of a strong civil service starting in the Sui
dynasty when the keju system was introduced to select government officials, yet China has not
sustained high rates of economic growth in its history --- at least not in most part of the last 500
years. The story of growth-based promotion has a lot to recommend, but it does not tell us why
the CCP top leadership wants economic growth in the first place.
J
efferson and Zhang distinguish two sources of legitimacy bestowed on governments, one from
procedures and the other from performance. They believe that the CCP has adopted growth-
enhancing policies because it does not have procedural legitimacy but instead has to seek
performance-based legitimacy. This is indeed part of the story provided by our model because
the government has to guarantee reasonable economic performance in order to pacify the two
social groups. However, none of the authoritarian governments has procedural legitimacy and
most of them have to obtain legitimacy by delivering tangible benefits to the people, yet few of
them have achieved high rates of economic growth. That is, seeking performance-based
legitimacy alone is insufficient to generate sustained economic growth.
36
T
he studies by Gehlbach and Keefer (2008) and Besley and Kudamatsu (2007) have also provided
implications for the case of China although they are not intended just to explain China’s
economic success. Gehlbach and Keefer’s emphasis on party institutionalization is relevant to
the Chinese reality. The CCP has indeed been institutionalized in many aspects, especially
compared with the time under Mao’s rule, including peaceful succession, more regular and
more rational decision processes, more power sharing within the party, and strengthened
disciplines of party members. Those changes have definitely helped the CCP adopt better
economic policies. However, this explanation does not tell us why the CCP should enact policies
that benefit the whole society instead of policies that only benefit its members. On the other
hand, Besley and Kudamatsu’s model implies that the CCP leadership adopted growth-
enhancing policies because it had a strong hold of power. However, treating the CCP’s power as
exogenous may miss a key element of the explanation. The CCP might have enjoyed high
popularity in the early years after it took power, but its revolutionary dividends were depleted
by the time reform began because of its failures and mistakes in its first thirty years of rule. Its
current legitimacy largely comes from its ability to deliver continuous economic growth. This is
what our model shows by treating the government’s rule as an endogenously determined state.
While our explanation by no means supersedes any of the above explanations, we do
believe that we provide new insights to the study of economic growth in China and developing
countries in general. Economic growth in many developing countries is hampered by two
interwoven impediments, interest group capture and a predatory government. If interest groups
act like thieves when they try to capture the government, a predatory government is a bandit. It
is difficult to eliminate either of them; a more practical approach is to turn the government
from a reckless bandit to a sedentary bandit who has a longer time horizon. Our paper identifies
social equality as a condition to make that transformation happen and at the same time to
constrain the ability of interest groups to capture the government. Since most developing
countries are characterized by social inequality, our paper, like several papers reviewed in the
introduction (especially Esteban and Ray, 2006; and Galor, Moav, and Vollrath, 2009) points out
a possible direction for improvement.
37
References
Acemoglu, Daron; Ticchi, Davide; and Vindigni, Andrea (2007). “Emergence and Persistence of
Inefficient States.” working paper.
Acemoglu, Daron; and Robinson, James (2008). “Persistence of Power, Elites and Institutions.”
American Economic Review, Vol. 98, Issue 1, pp.267-293.
Alesina, Alberto; and Rodrik, Dani (1994). “Distributive Politics and Economic Growth.”
Quarterly Journal of Economics, May, Vol.109, No.2, pp.465-90.
Benabou, Roland (2000). “Unequal Societies: Income Distribution and the Social Contract.”
American Economic Review, Vol.90, No.1, pp.96-129.
Besley, Timothy; and Kudamatsu, Masayuki (2007), “Making Autocracy Work.” LSE STICERD
Research Paper No. DEDPS 48.
Blanchard, Oliver; and Shleifer, Andrei (2000). “Federalism with and without Political
Centralization: China versus Russia.” Manuscript, Department of Economics, Harvard
University.
Boycko, Maxim; Shleifer, Andrei; and Vishny, Robert (1996). Privatizing Russia. Cambridge: The
MIT Press.
Esteban, Joan; and Ray, Debraj (2006). “Inequality, Lobbying, and Resource Allocation.”
American Economic Review, Vol.96, No.1, pp. 257-279.
Galor, Oded; Moav, Omer; and Vollrath, Dietrich (2009). “Inequality in Land Ownership, the
Emergence of Human Capital Promoting Institutions, and the Great Divergence.” Review of
Economic Studies, vol. 76, no. 1, pp. 143-79.
Garnaut, Ross; Song, Ligang ; Tenev, Stoyan; and Yao, Yang (2005). China's Ownership
Transformation : Process, Outcomes, Prospects. Washington D.C.: The World Bank.
Ge, Wei (1999). Special Economic Zones and the Economic Transition in China. Singapore, New
Jersey, London and Hong Kong: World Scientific.
ehlbach, Scott; and Keefer, Philip (2008). “Investment without Democracy: Ruling-Party Institutionalization and Credible Commitment in Autocracies.” Manuscript, The World Bank.
Hassler, John; Rodríguez Mora, José V.; Storesletten, Kjetil; and Zilibotti, Fabrizio (2003). “The
Survival of the Welfare State.” American Economic Review, Vol.93., No.1, pp.87-112.
38
Huang, Lingwen; and Yao, Yang (2007). “SOE Restructuring and Employment: Evidence from 11
Cities.” Journal of Economic Research (In Chinese), V42, No.3, pp. 57-69.
Jefferson, Gary; and Zhang, Jun (2008). “China’s Political Reform: A Property Rights
Interpretation.” Manuscript, Department of Economics, Brandeis University.
Klein, Paul; Krusell, Per; and Rios-Rull, Jose-Victor(2008). “Time Consistent Public Policy.” Review
of Economic Studies, Vol.75, No.3, pp.789-808.
au, Lawrence J.; Qian, Yingyi; and Roland, Gérard(2000). "Reform without Losers: An Interpretation of China's Dual-Track Approach to Transition."
ee, Teng-hui
i, Hongbin; and Zhou, Li-An (2005). “
in, Justin Yifu; and Yao, Yang (2001). "Chinese Rural Industrialization in the Context of the East Asian Miracle." Chapter 4 in
University Press.
ational Bureau of Statistics (NBS) (2009).
aughton, Barry (2004).
ee, Victor (1989). “
ee, Victor; and Lian, Peng (1994). “Sleeping with the Enemy: A Dynamic Model of Declining Political Commitment in State Socialism.”
Olson, Mancur (1982). The Rise and Decline of Nations. New Haven and London: Yale University
Press.
adro-I-Miquel
Perkins, Dwight (2004). “China’s Economic Development in an International Perspective.” Paper
presented in the conference “China’s Economic Reform and Growth”, China Center for
Economic Research, Peking University, October 2004.
39
ian, Yingyi; and Weingast, Barry R. (1997). “Federalism as a Commitment to Preserving Market Incentives.”
ian, Yingyi; and Roland, Gérard (1998). “Federalism and the Soft Budget Constraint.”
enzo, Marina Azzimonti (2007). “On the Dynamic Inefficiency of Governments.” Manuscript.
ind, Melise (2002). “The Concept of Disinterestedness in Eighteenth-Century British Aesthetics,”
iskin, Carl; Zhao, Renwei; and Li, Shi (2002).
hen, Yan; and Yao, Yang (2008).
he Growth Commission (2008).
ong, Jiadong (2008).
ang, Hui (2006).
oo-Cumings, Meredith (1997). “The Political Economy of Growth in East Asia: A Perspective on the State, Market, and Ideology.” In Masahiko Aoki, Hyung-Ki Kim, and Masahiro Okuno-Fujiwara, eds.
Asian Economic Development: Comparative Institutional Analysis
u, Li. “A Study of the Size of the ‘Price Scissors’ in China in 1949-1978.”
u, Chenggang (2008). “The Institutional Foundation of China’s Economic Growth.” Working paper, Department of Economics, University of Hong Kong.
ang, Guangbin (2005). “An Institutional Analysis of the Nanjing National Government’s Failure.”
Yao, Yang (2008). “The Political Economy of Government Policies toward Regional Inequality in
China.” In Yukon Huang and Alessandro Magnoli Bocchi, editors. Reshaping Economic
40
Geography in East Asia. Washington, D.C.: The World Bank, 2008.
Young, Alwyn (2003). “Gold into Base Metals: Productivity Growth in the People's Republic of
China during the Reform Period.” Journal of Political Economy, Vol. 111, No. 6, pp. 1220-
1261.
Zhang, Jun (2008). “Decentralization and Recentralization: The Trajectory of Economic Reform
and Growth in China.” Conference paper, “Thirty Years of Reform and Growth:
Retrospective and Prospects”, Peking University, October 24-25, 2008.
Zhao, Zhong (2005). “Migration, Labor Market Flexibility, and Wage Determination in China: A
Review.” The Developing Economics, XLIII, No. 2, pp. 285-312.
41
Appendix
In this appendix, we provide proofs of Propositions 1 and 2.
Proof of Proposition 1: We discuss the two cases in (19) and (20) separately. In the case of (19), the levels of utility of
the two groups are
(A1)
(A2)
Similarly, when the incumbent is Group P, the levels of utility of the two groups are
(A3)
(A4) .
If , then, similar to (19), we have
(A5) .
Then from (A1)—(A5), we have
(A6.1) , and
(A6.2) .
That is, each group gets the same level of utility regardless who is the incumbent at the
beginning of the game. The social output under revolution is
42
(A7)
I
f , then, similar to (20), we have
(A8) .
From (A1)—(A4) and (A8), we have
(A9)
This means that
(A10)
And the utility of Group E equals to the level defined in (A6.2). So we have
(A11)
This concludes our proof for the case of (19). For the case of (20), we first notice that, the
43
social output with one-period of revolution is obviously smaller than the Social output under
autarky if . On the other hand, if , then we have
(A12)
(A13)
From those two results, we know that
(A14)
(A15)
.
From (A14) and (A15), we know that
44
(A16) .
If the inequality holds, then like the case of (19), group P’s reservation utility when it is the
incumbent is . Therefore, the value of , which is determined by (A1), will be the
value shown in (A6.1). Combining (A5) and (A6), we obtain similar result like the one in (A11). If
the equality holds, then from (A14) and (A15), we have , i.e.,
(A17) .
In this case , we claim that , If not, we have . Then, we can calculate
from (A6) or (A9), that both lead to
(A18)
,
which contradicts with (A17).
Q.E.D.
45
Proof of Proposition 2:
Notice first that in the case of (19), we know from Proposition 1 that
(A19)
, and
(A20) .
Similarly, in the case in (20), we have
(A21) , and
(A22)
when ; or
(A23)
, and
(A24)
when .
Next we show that the disinterested government cannot exist in the case of (19), in which the
group utility is determined by (A6). First notice that the ratio of endowments, , is a
constant number. Then, the utility of the individual is a linear function of its own endowment,
46
i.e., , for some constant . Being disinterested requires . However, in the
case of (19), we have
(A25
a)
(A25b)
and from the proposition 1, we know that and , so the government cannot be
disinterested. The disinterested government can only occur in the case of (20), which implies
(A26)
.
It becomes the social equality condition (22) after some algebraic transformation. The other
results of Proposition 2 follow straightforward by solving the program in (21).
Q.E.D.
47