Rent Seeking, Market Structure and Growth

Daniel Brou
University of Western Ontario, London, ON N6A 5C2, Canada
[email protected]
Michele Ruta
World Trade Organization, Geneva 21, Switzerland
[email protected]
Abstract
We study the role of productive and unproductive entrepreneurship in economic growth in
a setting where rms compete in both economic and political markets. Specically, rms
compete for market share through cost reducing technological innovation and vie for inuence
over government transfer policy through rent seeking activities. We nd that rent seeking
aects growth in two ways: it allows rms to ignore economic competition, leading to less
innovation, and alters the number of rms that are supported in equilibrium. The former
eect is negative while the latter is ambiguous. We show how these eects depend on various
characteristics of economic and political markets.
Keywords: Political Economy, R&D, Entrepreneurship, Economic Growth, Welfare
JEL classication: D72, L13, O31

We would like to thank Alessandra Casella, Omar Licandro, Massimo Motta, Brendan OFlaherty, Rick
van der Ploeg, Ragnar Torvik, Jaume Ventura, two anonymous referees, and seminar participants at LUISS,
EUI, and Northwestern University for helpful comments. The opinions expressed in this paper should be
attributed to the authors.
1
I Introduction
In the pursuit of prots, rms seek to gain any advantage over competitors. One way to
accomplish this is to invest in research and development (R&D) that may result in the
discovery of new products or improved production technologies. It is this innovative activity
that drives economic growth. The rm, by pursuing its own interests, serves society at large
by introducing new and improved products or providing existing products at lower cost. But
there are other, less productive ways in which rms can seek an advantage over competitors.
What Baumol (1990) describes as unproductive entrepreneurship can take many forms
including lobbying, tax evasion, litigation, or theft and can achieve a variety of goals which
are protable for the rm, but wasteful from the point of view of society. This type of
activity, which is generally referred to as rent seeking, will have important eects on the rate
of innovation, the structure of the market and, consequently, on the rate of growth of the
economy.
There is no clear agreement in the literature about the empirical relationship between
rent seeking and growth. Rent seeking is inherently dicult to measure but Mauro (1995),
using subjective indices of perceived corruption, nds a negative relationship with economic
growth. Bardhan (1997), on the other hand, identies many instances where rent seeking
activities have contributed to growth. Whether rent seeking facilitates or hinders growth
seems to depend on the political and economic environment.
An important consideration is the channel through which rent seeking aects growth
and welfare. The literature on rent seeking and economic performance has focused on two
main channels. The private advantage from an innovation is signicantly greater when the
innovating rm has monopoly power over the new product or improved technology. Rent
seeking activity can then be devoted to ensuring that property rights over the innovation
2
are protected and their value is not diminished by newer technologies. Parente and Prescott
(1994) argue that, when monopoly rights are very strong, rms do not need to innovate in
order to maintain market share. If rms nd it more protable to devote resources to the
protection of existing monopolies than to R&D, overall eort in innovation will fall as a
result of rent seeking.
1
A second way of ensuring monopoly rights is to prevent the adoption
of new technologies. When innovation occurs through a process of Schumpeterian cre-
ative destruction, the owners of pre-existing technologies have an interest in limiting this
process. Mokyr (1998) provides ample historical evidence of this resistance to innovation
and Bellettini and Ottaviano (2005) model this tension as a dynamic common agency prob-
lem between overlapping generations and policy makers. Aghion et. al. (2008) explicitly
model rent seeking activity devoted to obtaining government protection from the entry of
competitors.
Inherent in these arguments is the assumption that the government can grant perfectly
enforceable monopoly rights that restrict the market and create a local monopoly. But it is
also the case that rms inuence government policies that supplement their prots rather
than directly block entry.
2
A variety of government activities have this eect, including
direct subsidies, tax breaks, loan guarantees, below-market-rate loans, and the provision
of subsidized inputs or above market prices through state-owned enterprises. The OECD
(2010) documents that these practices represent a signicant amount of public funds.
3
In
eect, rms use the political process to help them survive against competitors - as witnessed
in the recent nancial crisis. This suggests an alternate channel for rent seeking to aect
growth and welfare: it can alter market structure in ways that the industrial organization
1
Classic works along these lines include Krueger (1974), Murphy, Shleifer and Vishny (1991) and Krusell
and Rios-Rull (1996).
2
Campos et. al. (2010) provide evidence that unproductive activities alter the expected cash ows of
potential entrants and have an impact on the level of competition in economic markets.
3
For example in the European Union, where controls on state aids are strictest, they represented 1%
of EU GDP in 2008. If crisis related measures are included, the number triples.
3
literature argues play an important role in the rate of innovation, including the number and
size of rms.
4
This suggests that a deeper understanding of the relationship between rent
seeking, market structure and innovation is required in order to analyze the eect of rent
seeking on economic growth.
5
We address these issues by introducing rent seeking in a growth model where inno-
vation and market structure are determined endogenously. We dene market structure as
the number and composition of rms participating in the market, recognizing that this is
endogenously determined and depends on the demand and technological characteristics of
markets. Following Peretto (1996 and 1998) we consider oligopolistic markets where each
rm carries out two dierent activities (in addition to rent seeking): producing a dierenti-
ated good and carrying out R&D. The latter results in cost-saving innovation for the rm
as well as non-specic knowledge that can be used in future R&D. This characterization,
combined with the assumption that patent protection is not perfect, implies that rms will
want to carry out R&D in-house and that R&D will have positive spillover eects on the
economy. Firms can also participate in rent seeking by dedicating resources to inuence the
government for direct transfers which are nanced through taxation. The model captures
two important ideas. First, prots from rent seeking supplement economic prots for rms
and represent a (direct or opportunity) cost to taxpayers.
6
Second, modern corporations
4
For a survey of this literature, see Baldwin and Scott (1987). Later work on this topic includes Nickell
(1996), Aghion et. al. (2004) and Aghion et. al. (2005). The literature linking growth theory to industrial
organization is discussed in Aghion and Howitt (2005).
5
The role of market structure has been largely ignored, with two exceptions. Bliss and Di Tella (1997)
study the eect of corruption on the equilibrium number of rms, but do not address the eects on economic
growth. In a related paper, Brou and Ruta (2011), we employ a model of endogenous market structure and
rent seeking to investigate the relationship between economic integration and political integration in the
growth process.
6
These transfers are likely to be unpopular and, therefore, dicult to sustain politically. Though we do
not model it explicitly, we have in mind that they are the result of an agency game between the electorate and
the government - as in Persson and Tabellini (2003) - where the characteristics of the political environment
allow for some rents. In the rest of the paper, we do not focus on the basic question of why these ineciencies
in the political system exist and focus, instead, on their consequences.
4
conduct productive and unproductive activities concurrently.
7
In this context, the choice
between R&D and rent seeking is, therefore, internal to the rm. The model allows us to
study how this choice depends on characteristics of the economic and political markets as
well as how decisions on R&D investment and rent seeking feed back into the structure of
markets and aect economic growth and welfare.
We consider the joint determination of market structure and the rate of growth in a
symmetric model. Market structure has an eect on economic growth because it determines
the level of rivalry in the market and inuences the private costs and benets of innovation.
Furthermore, R&D spending makes up one part of rms total costs and plays a role in
the entry/exit decision. The growth path of the economy depends on the intensity of R&D
competition that this interaction generates. Rent seeking plays an important role in this in-
terdependence for three reasons. Firstly, transfers to rms are nanced by taxing consumers.
Because rms are owned by households, we might expect that any prots from rent seeking
should make their way back into consumers pockets. But in free entry equilibrium, prots
are forced to zero and the size of the economic market is reduced by the total amount of
transfers to rms. This has a negative eect on the return to R&D and induces rms to
reduce expenditures on innovation, thus leading to lower growth. Secondly, rent seeking has
an eect on market structure by changing the incentives to enter the market. In particular,
the direct gains from rent seeking increase prots for the typical rm while the decrease in
market size reduces them. The net eect on prots will determine whether rent seeking in-
duces entry or exit, with further repercussions for growth. Thirdly, prots from the political
market ensure that rms are less dependent on prots from the economic market to remain
7
Laband and Sophocleus (1992) document the vast expenditures on transfer activities by rms (including
those aimed at governments) in the US. Publicly available data on lobby expenditures corroborates their
ndings. For example, AT&T spent more than US$36 million in political contributions to nance the electoral
campaigns of candidates in national elections from 1990 to 2006. Microsoft spent US$15 million and Enron
US$6.5 million (Source: www.opensecrets.org). Ocial data are likely to represent only a fraction of total
spending for rent seeking.
5
competitive. This slack reduces a rms incentive to invest in R&D.
8
As in the previous literature, rent seeking has a negative impact on growth through
the eect of a shrinking market and the slack provided by political prots. Dierently
from previous work in this area, this eect can either be reinforced by the exit of rms
or counteracted by entry. A natural question arises: which of these eects is stronger?
The answer depends on the characteristics of both the political and economic markets. In
the formal literature, there is no consensus on the type of political market that may best
describe the competition for rents. As Bardhan (1997) and Bardhan and Mookherjee (2005)
suggest, the way in which corruption and rents are determined varies on a case by case
basis. To address this issues, we rst consider a general market for rents and describe the
characteristics that will determine the net eect of rent seeking on growth. We then study a
specic example based on the classic rent seeking model by Tullock (1980). By considering
three special cases of this model, we are able to highlight the channels through which rent
seeking aects economic growth.
The paper is organized as follows. Section II presents the formal model and section III
solves it. In section IV, we establish general results on the eects of rent seeking on market
structure, growth and welfare. Section V provides a specic example of a rent seeking game.
Concluding remarks follow.
II The Model
Consider an economy composed of a population of 1 identical consumers with symmetric
preferences over dierentiated goods supplied by oligopolistic producers. Consumers are
8
Our choice of oligopolistic markets matters. Models with localized monopolies, such as the expanding
varieties and quality ladder models of Grossman and Helpman (1991), would not capture all three eects
outlined above.
6
endowed with one unit of labor each, which they supply inelastically. The number of rms
is determined endogenously and rms compete in two markets. In the economic market,
rms engage in two activities. Each rm produces a good which it sells to consumers. At
the same time, each rm undertakes cost reducing R&D. In the political market, rms vie
for a share of the governments scal transfer budget. All three activities undertaken by the
rm require the use of labor as an input. The economic market is modeled following Peretto
(1996).
9
Consumer Behavior
Demand for each good is derived from the behavior of consumers. We assume that consumers
have identical, symmetric preferences over the available varieties of goods. The typical
consumer maximizes lifetime utility,
l(t) =
_
1
t
c
j(tt)
log
_
a

i=1
c
i
(t)
1

_
1
dt. (1)
subject to the intertemporal budget constraint

= \ +1+:1 1. The individuals
discount rate is j 0 and c 1 is the elasticity of product substitution. Per capita
expenditure is given by 1 =

a
i=1
j
i
c
i
, where j
i
is the price of good i, c
i
is consumption of
good i and : is the number of goods available. 1 represents per capita taxes and \ is the
wage rate that we take as the numeraire and assume equal to unity. Finally, is per capita
asset holding and 1 represents dividends.
10
Households obtain the optimal expenditure plan by setting the time path for expen-
9
Readers interested in a deeper discussion of the derivation of the models equilibrium will nd it in this
reference.
10
In free entry/exit equilibrium, prots will always be zero so that this last term can be omitted without
loss of generality.
7
ditures according to
.
1,1 = : j and maximizing instantaneous utility subject to per
capita expenditure. The resulting demand schedules are given by r
i
= o
i
11,j
i
, where
o
i
j
1c
i
,1
1c
is the the share of the market captured by rm i and 1
_
a
i=1
j
1c
i
1
1
is a price index. It will be useful to characterize demand by the price elasticity of demand,

i
= c (c 1)o
i
,
Production Technology
Each rm produces output with the following technology:
r
i
= .
0
i
(1
a
i
c). (2)
where r
i
is the output of rm i, 1
a
i
is labor used in production, and c 0 is a xed and
sunk cost of production. The rms knowledge (or patent) stock is given by .
i
and o 2 (0. 1)
is the elasticity of cost reduction. The term .
0
i
captures the idea that the marginal cost of
production is decreasing in a rms accumulated stock of knowledge.
Firms invest in R&D in order to accumulate cost reducing innovations that are patented.
These innovations are specic to the rm, but the R&D process produces knowledge that is
useful to other rms. Technological innovations evolve according to the following condition
.
.
i
= 1
:
i
_
.
i
+
a

)6=i

1 + o(: 1)
.
)
_
1
:
i
2
i
. (3)
where 2
i
is the stock of knowledge available to rm i in the innovation process. Taken
together, equations (2) and (3) imply that rms can only use the knowledge they have
produced, .
i
, in nal good production, but knowledge produced by other rms helps in the
production of new rm-specic knowledge. The parameter 2 (0. 1) determines the share of
8
privately developed R&D that becomes publicly available. The parameter o determines how
quickly congestion sets in. If the rm allocates 1
:
i
units of labor to R&D in an interval of
time dt, it produces
.
.
i
new patents. The R&D technology exhibits overall increasing returns
to scale and constant returns to scale in knowledge.
11
Prots from the economic market are
dened as :
c
i
j
i
r
i
1
a
i
1
:
i
.
The Political Market
We model a polity where the government collects lump sum taxes from consumers and
uses this common pool of tax revenue to provide transfers to rms. The key assumption
is that the government allocates the revenues to rms in response to their rent seeking
eorts.
12
Formally, we assume that rms dedicate a given amount of labour to inuencing
the government. We posit that rent seeking cannot be separated from the other activities of
the rm - in other words, only rms engaging in production can inuence the government.
13
This is meant to capture the fact that modern corporations engage in both R&Dand inuence
activities.
The government must satisfy the following budget constraint in each period,
1 11 =
a

i=1
Q
i
. (4)
where Q
i
represents the transfer to rm i, 1 is the per capita tax on consumers and 1 is
11
The model can be reinterpreted as having quality-improving innovation, where quality is a continuous
variable.
12
As discussed in the introduction, we wish to introduce a general rent seeking structure. This representa-
tion will allow us to consider the eect of various dierent rent seeking games. In section V we will provide
a specic example.
13
That rms are willing to both lobby the government and participate in production can also be explained
with a small modication to the specication of the model. If the xed cost of production, c, is interpreted
as a xed and sunk cost of running a rm (so that a rm wishing to just lobby the government would also
have to pay it), then economies of scope imply that a rm can lower costs by participating in both markets.
9
the overall budget, that can either be xed ex ante or endogenously determined. We dene
a rent seeking technology as a function,
Q
i
= , (1
Q
i
. ) . (5)
which translates the rms eorts in rent seeking, 1
Q
i
, into rents. This general representation
will allow us to capture a variety of plausible relationships between rent seekers and the
government. The vector contains variables that describe the way that rent seeking eort
is converted into transfers. This may include exogenous measures of political competition
such as the level of government bargaining power, the level of government malfeasance or the
political cost of raising taxes. The vector may also include endogenous measures of political
competition such as the rent seeking behavior of other rms (1
Q
i
) or the number of rms
in the market (:), as is widely discussed in the literature (see for instance Ades and Di Tella
(1999)). At this stage, the only restriction we place on the function , is that it is increasing
and concave in 1
Q
i
. Prots from the political market can be dened as :
j
i
Q
i
1
Q
i
.
It is useful to discuss the dierences between the market for consumption goods (the
economic market) and the market for rents (the political market). Firms compete for favours
from the government in a way that is fundamentally dierent from the way in which they
compete for market share. In the economic market, rms are oligopolists, oering dierenti-
ated goods and engaging in Bertrand competition. Competition takes place over prices and
rms can capture market share from their competitors by investing in cost-reducing R&D.
In the political market, rms compete with each other by trying to obtain the same good -
government granted rents. This is, in eect, a game of distribution where the government
need not have an inherent preference for one rm over the other. The economic and political
markets are linked, however, by the fact that the total rents paid out by the government
10
must be raised through taxation. This has a negative income eect on consumers, who re-
duce their demand for all goods. Section III provides a formal treatment of these ideas and
studies the eects of rent seeking on market structure and economic growth.
Firm Behavior
For rm i, the present discounted value of net prots is
\
i
(t) =
_
1
t
c

R

t
v(c)oc
[:
c
i
(t) + :
j
i
(t)] dt. (6)
The rm will maximize \
i
subject to technological and institutional constraints (2), (3) and
(5), total demand, and taking as given the number of active rms and competitors pricing,
innovation and rent seeking strategies. The initial knowledge is assumed to be equal for all
rms.
In symmetric equilibrium, rm-specic subscripts can be dropped. The elasticity of
substitution is = c (c 1),: and the stock of knowledge available to each rm when
innovating is given by 2 = [1 + (: 1), (1 + o(: 1))] . c(:).. where c(:) represents
the productivity of labour in R&D. Each rms optimizing behavior is characterized by the
Bertrand-Nash price strategy, j = [,( 1)] .
0
, and the optimal level of R&D activity,
1
:
=
11o( 1)
:

:
c
. (7)
Equation (7) captures some important characteristics of R&D in partial equilibrium.
The term 11,: represents the gross-prot eect that depends on total sales per rm and
the oligopoly mark-up. The term o( 1) is the business-stealing eect - by investing in
cost reducing innovations, rms lower prices and expand their market share. Spillovers
11
(represented by the term c) have two distinct eects, one negative and one positive. R&D
makes their competitors more productive, but they also benet from the R&D undertaken
by other rms.
When choosing the optimal amount of resources to devote to rent seeking activity, each
rm equates the marginal benet of one additional unit of labor used in rent seeking with its
marginal cost.
14
The optimal amount of resources employed in rent seeking can be written
as 1
Q
i
= 1
Q
i
(). Substituting the optimal rent seeking strategy into Q
i
, we obtain the
equilibrium amount of rents that rm i can extract from the government. In symmetric
equilibrium instantaneous prots are given by
:
c
+ :
j
=
11 [1 o( 1)]
:
+
:
c
+ Q1
Q
c. (8)
The above discussion describes the Nash equilibrium for a given market size, 11, and
number of rms, :. Since we assume that all rms start out with the same stock of knowledge,
rms will accumulate knowledge at the same rate and symmetry will be preserved at all times.
III Market Structure and Growth
In order to establish equilibrium market structure, we rst consider the behavior of rms
taking the number of competitors and the price of labor as given. Firms maximize their stock
market value through the choice of pricing strategy, R&D expenditure, and rent seeking
expenditure. The entry/exit decision then determines the number of active rms. Once the
behavior of rms is established, market clearing conditions determine the general equilibrium
14
Notice that the rm does not consider the eect of its rent seeking activity on total expenditures. This
need not be viewed as an assumption, but rather a result of the fact that rms are price takers in the labor
market and take the number of rms as given. Any prots made by rms from the rent seeking process will
be paid to consumers as dividends and any cost of R&D will be paid to consumers as wages. From the point
of view of the rm, the net eect of rent seeking on expenditures would be zero.
12
of the economy.
Free Entry and General Equilibrium
Consider the entry/exit decisions of rms.
15
In the absence of entry/exit costs, the number
of rms is a jumping variable and \ = 0 at all times. Moreover, stock prices must satisfy
the arbitrage condition, :\ = :
c
+ :
j
+
.
\ . These conditions imply that prots must equal
zero at all time. Together with condition (7), the zero prot condition denes the Nash
equilibrium with free entry, given the interest rate, :, and the size of the market, 11.
By summing the zero prot condition across rms and imposing labor market clearing,
per capita expenditures are shown to be 1 = 1 :Q,1. Notice that rent seeking activities
represent a waste of resources and reduce per capita equilibrium expenditures. Since : and
1 are constant along the balanced growth path it must be the case that : = j. The optimal
investment strategy (7) can now be written as a function of exogenous parameters and the
number of rms. Substituting this into condition (8) and setting it equal to zero yields the
zero prot condition,
1[1 o((:) 1)]
:(:)
_
1
:Q()
1
_
+ :
j
() +
j
c(:)
= c. (9)
This condition implicitly determines the equilibrium number of active rms.
16
Notice that
the market structure of this economy depends on the interaction of rms in the economic
and political markets. In particular, rent seeking activities provide an extra source of income
to rms (:
j
), but also reduce sales because of the eect of higher taxes on equilibrium
15
For analytical convenience, in the rest of the analysis we treat the number of rms as a continuous
variable.
16
If the vector does not contain :, total prots - the left hand side of (9) - are strictly decreasing in :
and there is a unique equilibrium. There are also a large class of rent-seeking games for which the left-hand
side is decreasing in : so that the equilibrium number of rms is also uniquely determined. We consider an
example in section V.
13
expenditures.
17
Condition (9) is depicted in the top panel of Figure 1.
[Figure 1 about here]
Equilibrium Growth and Welfare
Firms take the number of competitors as given and choose the optimal level of R&D accord-
ing to condition (7). The equilibrium number of rms, in turn, is endogenous and determined
by the zero prot condition (9). The number of rms is a jumping variable and is constant
along the balanced growth path. This implies that the total stock of knowledge available to
each rm grows at the rate

2,2 = c1
:
. Along the balanced growth path, both consumer
expenditures and the number of rms are constant. The rate of cost reduction then deter-
mines the growth of output and consumption. Using the zero prot condition (9) and the
denition of growth yields the equilibrium growth rate of the economy
q

C,C = o

2,2 = o
oc(:) [(:) 1] [c :
j
()] j
1 o [(:) 1]
. (10)
This condition is a modied version of the rms R&D decision, where rms have perfect
foresight, correctly perceive the eect of parameter changes on their prots and, based on
this, choose whether to be active or not. It therefore shows, for any given number of rms,
the incentive to pursue the innovation that drives economic growth. Together with the
equilibrium number of rms determined by condition (9), this determines the economys
equilibrium growth rate as depicted in the lower panel of Figure 1.
Welfare in this economy is closely related to economic growth. Integrating the con-
17
These results would be slightly dierent if it were rms rather than consumers who paid the tax. The
market size eect would not arise. Furthermore, in symmetric equilibrium, the amount of taxes paid by each
rm would exactly equal the transfers received and
p
= 1
Q
. All other results and comparative statics
would be the same. We prefer the assumption that taxes are paid by consumers because rms extract some
benet from rent seeking.
14
sumers lifetime utility function yields
l =
1
j
_
1
c 1
log : + log
1

+
q
j
+ log 1
_
. (11)
Welfare is increasing in the number of goods available, in the rate of growth of consumption
(which is equal to the rate of cost reduction, q) and in total expenditures.
IV The Eects of Rent Seeking
Rent seeking, like market structure and growth, is an endogenously determined feature of
this model. In order to analyze the implications of rent seeking on the economic outcomes
of interest we will compare the outcomes derived in the previous section to those that would
arise if the possibility of rent seeking were eliminated from the model.
Market Structure
In the absence of rent seeking, the equilibrium number of rms is determined by condition
(9) with 1
Q
= Q = 0. The change in the left-hand side of (9) when rent seeking is eliminated
is
:
j
()
[1 o((:) 1)]
(:)
Q(). (12)
This term represents, for a given number of rms, the total contribution of rent seeking
activity to the rms prots. In other words, it represents the shift in the :
c
+ :
j
c
schedule in the top panel of Figure 1. We see that rent seeking has two eects on the
equilibrium number of rms, which work in opposite directions. The rst term represents
the increase in prots from rent seeking activity, the second represents the loss of prots
from a reduction in consumers equilibrium expenditure brought about by higher taxes.
15
We can see that the eect of rent seeking on market structure depends on characteristics
of both the economic and political markets. The protability of rent seeking, :
j
, relative to
the size of rents, Q, provides an incentive for entry. Economic market conditions shape this
comparison. The term [1 o( 1)] , captures the fraction of revenues from the economic
market that the rm keeps as prots. The higher this term, the more costly is the loss of
market size due to rent seeking. More precisely, notice that this term is higher for lower
values of c (the elasticity of product substitution) and o (the elasticity of cost reduction).
Using the terminology of Bliss and Di Tella (1997), we can refer to c and o as deep
competition parameters, because they are demand and technology parameters, respectively,
that determine the structure of economic market in the absence of rent seeking. The negative
impact of rent seeking on rms prots is larger for those industries that are inherently less
competitive (i.e. with low c and o). This shows the following:
Proposition 1. Rent seeking can either increase or decrease market entry. For any
given political structure, the negative impact of rent seeking on the equilibrium number of
rms will be larger for an industry characterized by low elasticity of product substitution ( c)
and low elasticity of cost reduction ( o).
Equilibrium transfers, Q, and prots from rent seeking, :
j
, depend on characteristics of
the political market and on the eects of competition. In general, if rents are large, but the
cost of obtaining them is also large so that political prots are small, then rent seeking will
have a negative eect on prots and rms will exit from the market.
16
Growth and Welfare
To see the impact of rent seeking on growth, rst dene the growth rate (10) when there is
no rent seeking (i.e. when 1
Q
= Q = 0) as q
avc
.
18
The change in the growth schedule can
be expressed as:
q q
avc
=
oc[ 1]
1 o [ 1]
:
j
. (13)
For any given market structure, this term represents the direct eect of rent seeking on
growth, which is always negative. In terms of Figure 1, it represents the shift in the q(:)
schedule in the second panel. When prots generated by rent seeking are high, production
plays a lesser role in the rms equilibrium prots. As a result, there is less of an incentive to
invest in R&D and growth is lower for any given number of rms: the growth schedule shifts
down. In other words, political prots provide rms with slack in the economic market. The
greater are political prots, the less pressure rms feel to cover their xed costs through
economic competition. As long as prots from rent seeking are positive, the growth schedule
of the economy will be lower because rent seeking reduces the incentive to innovate for any
given number of rms.
[Figure 2 about here]
Notice that if the political market were to generate no direct prots for a rm, the
growth schedule would not be aected. This does not, however, mean that rent seeking
would not have any eect on economic growth. As shown in the previous subsection, rent
seeking aects the incentives to enter or exit the market. The number of rms plays an
important role in determining the incentive to innovate and the growth rate of the economy.
When the prots from rent seeking are larger than the cost in terms of a reduced market
18
Assuming 1 0 (c 1) is sucient to prove the symmetry of the Nash equilibrium (see Peretto, 1998,
proposition 1). Under this assumption, one can show that the growth rate of the economy without rent
seeking (q
nrs
) is positive only for a suciently large number of rms and always increasing in :.
17
rent seeking will induce entry. The increase in competition for consumer spending implies
an increase of knowledge spillovers, a positive business stealing eect and a negative gross
prot eect. The rst two eects always dominate the third and will result in higher R&D
spending. The negative eect on growth discussed above will be mitigated, though the total
eect on growth is ambiguous. This possibility is depicted in Figure 2. Alternately, when
the cost of rent seeking more than osets any benets, rent seeking will induce exit and
growth will be further reduced. This case is depicted in Figure 3.
[Figure 3 about here]
The eect of rent seeking on welfare is closely linked to growth. If we again dene
consumers lifetime utility (11) in the absence of rent seeking as l
avc
, then the dierence
l l
avc
=
1
j
_
q q
avc
j
+ log(1
:Q
1
)
_
(14)
shows that the impact on welfare, for a given number of rms, depends on the impact on
growth together with an (always negative) income eect.
The above results are summarized in
Proposition 2. Rent seeking has two eects on long-run growth. First, rents reduce
incentives to invest in R&D for a given number of active rms by creating slack in the
economic market (direct eect). Second, rent seeking alters the equilibrium number of active
rms, further changing the incentives to innovate (indirect eect). There is an additional
negative eect on welfare through the decrease in expenditures.
Graphically, the direct eect of rent seeking is to shift the growth schedule down for any
value of :. Rent seeking also alters the equilibrium structure of the market, : (the indirect
eect). The interaction of these two eects - i.e. the intersection of the two schedules in
Figures 2 and 3 - determines the long run rate of growth of the economy. It is important
18
to note that market structure and the growth rate depend importantly on what happens in
both the economic market and the political market. The inherent level of competition (in
the sense of the deep parameters) in each market aects the number of rms that can be
supported in equilibrium. Furthermore, political factors directly inuence the incentives of
rms to engage in R&D. The overall eect of rent seeking on growth will depend on the
specic characteristics of the economic and political markets. We turn our attention to this
in section V.
V A Model of Rent Seeking
Until now we have focused on a very general structure for the market for rents. This section
introduces a specic model of rent seeking developed by Tullock (1980). The rent a rm
obtains is assumed to be an increasing function of that rms share of total rent seeking
activity. The general form of this model can be expressed as:
Q
i
=
1
A
Q
i

)
1
A
Q
j
1. (15)
In terms of the general rent seeking function described in section III, the vector of political
market characteristics is now =
_
1
Q
i
. 1. `

, where 1
Q
i
is the vector of rent seeking
eort of rm is : 1 competitors, 1 is the exogenous government budget, and ` 0
is a parameter measuring the responsiveness of government to rent seeking eort. In each
period there is a xed budget (1 0) nanced by lump sum taxation which the government
allocates to rms in response to their rent seeking eort. We therefore have that in each
period 1 = 11 =

i
Q
i
.
This model is consistent with two aspects of rent seeking that are commonly described
in the literature. First, rent seeking is a directly unproductive activity in that rms dedicate
19
real resources to obtain a prot without producing any good or service (along the lines of
the directly unproductive prot-seeking of Bhagwati (1982)). Secondly, competition between
dierent special interests reduces the returns to rent seeking (along the lines of Becker
(1983)).
19
This choice further simplies the analysis by imposing that the amount available
for distribution by the government is xed: rms cannot inuence the level of taxation, only
their share of the pie. More generally, this specication can be thought of as the reduced form
of political model as in Persson and Tabellini (2003) where the parameters are the outcome
of an agency model of political accountability. In particular, ` measures how responsive the
government needs to be to certain special interests. The Tullock model also has the benet
of sharing important characteristics with other interesting rent seeking games. Baye and
Hoppe (2003) show that under certain conditions, a Tullock game is equivalent to models of
innovation tournaments and race patents.
The game has a symmetric pure strategy Nash equilibrium if and only if ` :,(:1),
so we restrict our attention to this case.
20
The rent seeking literature is concerned with the
total rent seeking outlays relative to the value of the contested rent - i.e. the level of rent
dissipation. The model specied in (15) allows us to consider two possibilities: full and partial
rent dissipation.
21
It is important to note that even though the case of over-dissipation is
ruled out in this game, it does not rule out the possibility that in general equilibrium rent
seeking will have a negative eect on growth and welfare greater than just the reduction in
income. As we show, rent seeking draws valuable resources away from more productive uses.
19
Theoretically the eect of competition on rent seeking is ambiguous. However, Ades and Di Tella
(1999) nd that corruption is higher in economies with fewer number of rms, thus supporting the idea that
competition reduces rent seeking.
20
Whenever ` :,(: 1), there is a continuum of Nash Equilibria in mixed strategies, but no pure
strategy equilibrium. A general characterization of the equilibria is not available in the literature except for
the limit case of ` !1.
21
That over-dissipation is not possible is not a result of our restriction on `. Baye, Kovenock and de Vries
(1994) show that the equilibrium mixed strategies when ` :,(: 1) cannot allow for over-dissipation in
expectation because this would imply that rms would have a negative expected payo and would be better
o not participating in rent seeking activity at all.
20
In this sense, our results are similar to those in the resource curse literature. Typically,
the decrease in income, growth or welfare that results from an increase in natural resources
is not (fully) explained by the assumption of a political contest with over-dissipation, but
rather by a re-allocation of productive resources (e.g. Tornell and Lane (1999), Sachs and
Warner (1995)). Torvik (2002) uses a related rent seeking function to show that a shift in
entrepreneurial eort toward rent seeking can explain the resource curse.
Full Rent Dissipation
Full rent dissipation occurs when ` = :,(:1).
22
Formally this implies that in a symmetric
equilibrium 1
Q
= Q = 1,: and :
j
= 0. Condition (12) then becomes
=
1[1 o((:) 1)]
:(:)
< 0. (16)
Since prots, for any given :, are lower when rent seeking is allowed, the number of rms will
be lower in a rent-seeking society. The intuition is that rents reduce the size of the economic
market (as in the general case). Moreover, political competition between rent seeking rms
fully dissipates their value to rms. This model allows us to isolate the market size eect of
rent seeking since, in equilibrium, rent seeking generates only this cost. Firms nd themselves
in a prisoners dilemma. They would all be better o if none of the rms participated in rent
seeking, but if this were the case, the incentive for any one rm to inuence the government
22
The analysis is the same for the case where ` ! 1. This is referred to as the all-pay auction because
the rm putting forth the greatest rent seeking eort is certain to receive the entire budget and all rms pay
the cost of rent seeking whether they win the contest or not. Baye, Kovenock and de Vries (1999) show that
there is a symmetric equilibrium where the expected payo for each rm is zero. There is also a continuum
of asymmetric equilibria, but all of these equilibria are payo equivalent so little is lost from focusing on the
symmetric equilibrium.
A similar result would obtain if we assume a lobbying a la Dixit, Grossman and Helpman (1997). Since
transfers are lump-sum, the government can costlessly transfer funds between the rms. This allows it to be
able to play them o against each other and extract the entire surplus. Each rm will increase its bid for
a transfer in the hopes of beating out its competitors up to the point where there is no benet from bidding.
21
is high because having no competition guarantees high returns from rent seeking. All rms
realize this and the result is positive rent seeking expenditures in every period with lower
total prots for every rm.
To see the eect on economic growth notice that condition (13) is equal to zero, since
:
j
= 0. There is no direct eect on economic growth because rent seeking does not generate
additional prots for the rm - does not provide any slack - and therefore does not alter
the incentive to engage in R&D for a given market structure (i.e. number of rms). If the
market could support the same number of rms as without rent seeking, the same level of
growth would be achieved. However, with the smaller market size resulting from the taxation
required to nance government transfers, a smaller number of rms will nd it protable to
enter the market. Faced with less competition, each active rm will have less of an incentive
to engage in R&D. The result is an unambiguous reduction in growth, as depicted in Figure
4. This has important implications for the welfare of consumers. Lower expenditures due
to taxation shift the equilibrium utility curve (10) down while a lesser variety of goods
to consume results in higher prices and reduced growth, leaving consumers unambiguously
worse o.
Proposition 3. In a Tullock economy with full rent dissipation, rent seeking lowers
long-run growth through its negative indirect (market structure) eect. As political prots
are zero, no direct eect on growth is present. Welfare is reduced.
This model is of particular interest as a starting point because it isolates the market size
eect. The traditional (i.e. direct) eect shifting down the growth schedule disappears.
Rent seeking leads to lower growth in equilibrium only through its indirect (market structure)
eect. Since rms compete away all benets from rent seeking and spending on their goods
is reduced, prots are lower and the market sustains a smaller number of rms.
[Figure 4 about here]
22
Partial Rent Dissipation
The stark result presented in the previous sub-section is important, but it was achieved
by shutting down some important channels through which rent seeking may aect growth.
Eectively, competition for transfers was so tough that rms could not generate positive
prots from rent seeking. But if we allow for rms to gain some advantage through rent
seeking, it may induce entry into the market and possibly oset the market size eect
described above.
23
Consider the case where ` :,(: 1). We will establish that the impact of rent
seeking on the economy depends on the interaction between economic and political market
parameters. Notice that the larger is `, the more the government is responsive to rent seeking.
This parameter can depend on institutional variables such as transparency, accountability of
bureaucrats and politicians and, more generally, on the checks and balances on government
activity. We can think of ` as a deep parameter of competition in the political market.
The symmetric equilibrium rent and prot from rent seeking are given, respectively,
by Q = 1,: and :
j
=
_
1,:
_
[1 `(: 1) ,:]. It is important to highlight the role that
checks and balances on the government play in this model. As ` increases, devoting resources
to rent seeking is more protable, because the government will provide larger rents for each
additional worker employed in rent seeking. However, since this is true for all rms it induces
more rent seeking by all rms. In equilibrium, this reduces the ability of each rm to extract
resources from the government.
The eect of rent seeking on entry/exit is captured by rewriting condition (12) to reect
23
It is has been argued, following Posner (1975), that free entry into the rent seeking game will ensure
full rent dissipation - total expenditures on rent seeking equal to total rents. This result does not apply to
the present model because rent seeking is not the only thing that rms do. Free entry will ensure that total
prots are pushed to zero, but this does not necessarily imply that prots from rent seeking will be zero.
23
equilibrium values:
=
1
:
_
1 `
_
: 1
:
_

1 o((:) 1)
(:)
_
. (17)
The term in square brackets is negative when the governments responsiveness to rent seeking
is high or when the elasticity of substitution between consumption goods is low. A sucient
condition for the term to be negative is ` (c 1)(1 + o),c. This condition will hold for
higher values of ` and lower values of c. When ` < (c 1)(1 + o),c, the initial (no rent
seeking) equilibrium number of rms matters. In particular, there is a cuto level of : given
by ` = (( :) 1)(1 + o),( :) such that for any initial :
avc
< : the term in the square
brackets will be negative and rent seeking will result in exit. This cut o level : is increasing
in ` and decreasing in c so that the condition is again more likely to hold for higher values
of ` and lower values of c.
As established in Proposition 1, when consumers cannot substitute freely between goods,
rms can charge a higher price and extract greater prots from every unit of revenue. If
this is the case, rent seeking is very costly because the reduction in market size has a more
pronounced eect on prots. Exit is also more likely for high values of `. As the quality
of institutions falls, rents increase but so does the cost of obtaining them. For this reason,
rent seeking is more likely to have a negative (general equilibrium) eect on prots. In
order for rent seeking to have a positive impact on entry it must be the case that ` is small
relative to c. Recalling that we referred to these as deep parameters of competition in
the the political and economic markets, respectively, this can be interpreted as saying that
rent seeking promotes entry only when political markets are relatively less competitive than
economic markets. Only when this is the case does the increase in prots from rent seeking
activities more than oset the loss of market size.
24
To see the eect on economic growth notice that the shift in the growth schedule is
q q
avc
=
oc[ 1]
1 o [ 1]
:
j
. (18)
Again, since rent seeking generates positive prots for the rm, the growth rate of the
economy is lower for any given :. It can be shown that the growth rate of this economy
still is everywhere increasing in : (if positive). More competition in the political market
reduces political prots, :
j
, thus providing further incentive for rms to invest in innovation.
The total eect of rent seeking on growth again depends on the negative direct eect and
the indirect (market structure) eect. When rent seeking results in exit, growth will be
unambiguously lower.
Proposition 4. In a general Tullock economy, rent seeking increases the equilibrium
number of rms if governments responsiveness to rent seeking activity is low ( ` < (c
1)(1 +o),c) and the number of rms supported by the market in the absence of rent seeking
is small ( :
avc
:, where : is implicitly determined by ` = (( :) 1)(1 + o),( :)), and
decreases the number of rms otherwise. In the rst case, rent seeking has an ambiguous
eect on long-run growth as the indirect eect of rent seeking is positive while the direct eect
is negative. In the second case, rent seeking unambiguously lowers long-run growth as both
its direct and indirect eect are negative. There is an additional negative eect on welfare
through the decrease in expenditures.
Two important implications can be drawn from this analysis. First, it is the interaction
between demand characteristics from both the economic and political markets that determine
the impact on growth. Secondly, rent seeking will have a positive growth eect only if it
induces enough entry into markets that the negative direct eect is canceled out. This
would occur only when rms have little market power over consumers (c high), and are able
25
to extract high rents from every dollar raised by the government (` low), and the market
supports few rms. In other words, rent seeking is good for growth only if there are few rms
engaging in low R&D and they cannot take advantage of consumers or impose too great of
a cost on society by extracting rents.
VI Conclusions
A rm can choose from a variety of strategies in order to maximize prots. Economists
have often focused on the way rms behave in economic markets, but a rm can also earn
prots in the political market through rent seeking. We provide a formal analysis of how this
type of unproductive activity can aect economic growth. Firstly, the prots generated by
rent seeking provide the rm with relief from economic competition and dull the incentive
to innovate. Secondly, the market structure sustained in equilibrium is altered, further
distorting the incentives to innovate. The former eect is consistent with the nding that
aggregate measures of rent seeking and corruption are negatively correlated with economic
growth, as in Mauro (1995). The importance of the latter eect is corroborated by Campos
et al (2010), who nd that rent seeking activities aect the cash ow of rms both through
the costs involved and the rents received. Furthermore, the change in cash ow aects the
entry decision of potential competitors and the protability of incumbent rms. Perhaps
more importantly, the relative impact of rent seeking on rm entry is conditional on the
quality of institutions, as we show in our discussion of the Tullock rent seeking example.
The market structure eect we identify is also broadly consistent with the work of Aghion
et al (2008), who nd that the quality of institutions can have a dierent eect on entry
and growth depending on characteristics of the economic market. Although their focus is on
democracy as a measure of institutional quality and distance from the technological frontier
26
as the economic characteristic, this result is similar to our ndings.
The model also provides additional insight into the role that government (scal) policy
plays in economic growth. This role is not dependent on whether government expenditures
are useful in production or subject to congestion, but rather based on how competitively rms
vie for the allocation of the scal budget and the implications this has on the incentives
to innovate. Moreover, the model has two clear testable implications. First, in countries
where governments are more responsive to rent seeking (i.e. where checks and balances on
government activity are weak), rent seeking works as an implicit barrier to entry for new
rms, as rents are oset by the large cost of obtaining them. In this case, rent seeking is
predicted to have a stronger negative eect on growth due to the resulting lack of competition.
Second, for a given quality of political institutions, in sectors characterized by low elasticity
of product substitution, that is where active rms benet from market power, rent seeking
lowers entry as the cost of losing market size dominates the benet of receiving a net rent.
In these sectors, rent seeking is predicted to have stronger negative eects on growth.
The set of policies which rms try to aect through rent seeking activity and the set
of means they employ to do so are both very large. The rst includes - but is not limited
to - taxes, subsidies (to production, R&D, etc.), barriers to entry and regulation, while the
former includes lobbying, bribery and informal inuence. No general theoretical framework
can capture all these important features and new insights will likely emerge from future work
in this area. However, we believe that our model is a useful rst step. It provides insights
into how the competition for a widely used policy - direct transfers to rms - aects economic
growth through several channels.
27
References
Ades, A. and Di Tella, R. (1999), Rents, Competition and Corruption, American Eco-
nomic Review 89, 982-993.
Aghion, P., Alesina, A. and Trebbi, F. (2008), Democracy, Technology, and Growth, in
E. Helpman (ed.), Institutions and Economic Performance, Harvard University Press,
Cambridge, MA.
Aghion, P., Bloom, N., Blundell, R., Grith, R. and Howitt, P. (2005), Competition
and Innovation: An Inverted U Relationship, Quarterly Journal of Economics 120,
701-728.
Aghion, P., Blundell, R., Grith, R., Howitt, P. and Prantl, S. (2004), Entry and Produc-
tivity Growth: Evidence for Microlevel Panel Data, Journal of the European Economic
Association, 2, 265-276.
Aghion, P. and Howitt, P. (2005), Growth with Quality Improving Innovations: An
Integrated Framework, in P. Aghion and S. Durlauf (eds.), Handbook of Economic
Growth, North-Holland, Amsterdam.
Baldwin, W. and Scott, J. (1987), Market Structure and Technological Change, Harwood
Press, New York.
Bardhan, P. (1997), Corruption and Development: A Review of the Issues, Journal of
Economic Literature 35, 1320-1346.
Bardhan, P. and Mookherjee, D. (2005), Decentralization, Corruption and Government
Accountability: An Overview, in S. Rose-Ackerman (ed.), International Handbook on
the Economics of Corruption, Edward Elgar Publishers, Cheltenham.
Baumol, W.J. (1990), Entrepreneurship: Productive, Unproductive and Destructive,
Journal of Political Economy 98, 893-921.
Baye, M.R. and Hoppe, H.C. (2003), The Strategic Equivalence of Rent Seeking, Inno-
vation and Patent Race Games, Games and Economic Behavior 44, 217-226.
Baye, M.R., Kovenock, D. and de Vries, C.G. (1994), The Solution to the Tullock Rent-
Seeking Game when R2, Public Choice 81, 363-380.
Baye, M.R., Kovenock, D. and de Vries, C.G. (1999), The Incidence of Overdissipation
in Rent-Seeking Contests, Public Choice 99, 439-454.
Becker, G.S. (1983), A Theory of Competition among Pressure Groups for Political
Inuence, Quarterly Journal of Economics 101, 371-400.
Bellettini, G. and Ottaviano, G. (2005), Special Interests and Technological Change,
Review of Economic Studies 72, 43-56.
Bhagwati, J.N. (1982), Directly Unproductive, Prot-Seeking (DUP) Activities, Journal
of Political Economy 90, 988-1002.
28
Bliss, C. and Di Tella, R. (1997), Does Competition Kill Corruption?, Journal of Political
Economy 105, 1001-1023.
Brou, D. and Ruta, M. (2011), Economic Integration, Political Integration or Both?,
Journal of the European Economic Association 9, 1143-1167.
Campos, N.F., Estrin, S. and Proto, E. (2010), Corruption as a Barrier to Entry: Theory
and Evidence, IZA Discussion Paper 5243.
Dixit, A., Grossman, G.M. and Helpman, E. (1997), Common Agency and Coordination:
General Theory and Application to Government Policy Making, Journal of Political
Economy 105, 752-769.
Grossman, G.M. and Helpman, E. (1991), Innovation and Growth in the Global Economy,
Cambridge University Press, Cambridge.
Krueger, A.O. (1974), The Political Economy of the Rent Seeking Society, American
Economic Review 64, 291-303.
Krusell, P. and Rios-Rull, J.V. (1996), Vested Interests in a Positive Theory of Stagnation
and Growth, Review of Economic Studies 63, 301-329.
Laband, D. and Sophocleus, J. (1992), An Estimate on Expenditures on Transfer Activity
in the United States, Quarterly Journal of Economics 107, 959-983.
Mauro, P. (1995), Corruption and Growth, Quarterly Journal of Economics 110, 681-712.
Mokyr, J. (1998), The Political Economy of Technological Change: Resistance and In-
novation in Economic History, in M. Berg and K. Bruland, (eds.), Technological Rev-
olutions in Europe, Edward Elgar Publishers, Cheltenham.
Murphy, K. M., Shleifer, A. and Vishny, R.W. (1991), The Allocation of Talent: Impli-
cations for Growth, Quarterly Journal of Economics 106, 503-530.
Nickell, S. (1996), Competition and Corporate Performance, Journal of Political Economy
104, 724-746.
OECD (2010), Competition, State Aids and Subsidies, Global Forum on Competition,
DAF/COMP/GF(2010)5.
Parente, S.L. and Prescott, E.C. (1994), Barriers to Technological Adoption and Devel-
opment, Journal of Political Economy 102, 298-321.
Peretto, P. (1996), Sunk Costs, Market Structure and Growth, International Economic
Review 37, 895-923.
Peretto, P. (1998), Technological Change, Market Rivalry, and the Evolution of the
Capitalist Engine of Growth, Journal of Economic Growth 3, 53-80.
Persson, T. and Tabellini, G. (2003), The Economic Eects of Constitutions, MIT Press,
Cambridge, MA.
29
Posner, R. (1975), The Social Costs of Monopoly and Regulation, Journal of Political
Economy 83, 807-827.
Sachs, J.D. and Warner, A.M. (1995), Natural Resource Abundance and Economic
Growth, NBER Working Paper 5398.
Tornell, A. and Lane, P.R. (1999), The Voracity Eect, American Economic Review 89,
22-46.
Torvick, R. (2002), Natural Resources, Rent Seeking and Welfare, Journal of Develop-
ment Economics 67, 455-470.
Tullock, G. (1980), Ecient Rent Seeking, in J.M. Buchanan, R. D. Tollison and G.
Tullock (eds.), Toward a Theory of the Rent Seeking Society, Texas A&M University
Press, College Station, TX.
30
Figure 1: Equilibrium market structure and growth
n

e
+
p

n*
n
g
g(n)
n*
Figure 2: The effects of rent seeking on market
structure and growth (when rent seeking induces entry)
n

e
+
p

n
g
g(n)
n*

e(nrs)
g
nrs
(n)
n
nrs
n

e
+
p

n
g
g(n)
n*

e(nrs)
g
nrs
(n)
n
nrs
Figure 3: The effects of rent seeking on market
structure and growth (when rent seeking induces exit)
n

e
+
p

n
g
g(n)=g
nrs
(n)
n*

e(nrs)
n
nrs
Figure 4: The effects of rent seeking on market
structure and growth (full rent dissipation)