A political economy theory of the soft budget constraint

ARTICLE IN PRESS
European Economic Review 53 (2009) 786–798
Contents lists available at ScienceDirect
European Economic Review
journal homepage: www.elsevier.com/locate/eer
A political economy theory of the soft budget constraint
James A. Robinson a, Ragnar Torvik b,Ã
a Harvard University, Department of Government, Littauer, 1737 Cambridge Street, Cambridge, MA 02138, USA
b Norwegian University of Science and Technology, Department of Economics, Dragvoll, N-7491 Trondheim, Norway
a r t i c l e
i n f o
a b s t r a c t
Article history:
Why do soft budget constraints exist and persist? In this paper we argue that the
Received 1 July 2008
prevalence of soft budget constraints can be best explained by the political desirability
Accepted 20 February 2009
of softness. We develop an in?nite horizon political economy model where neither
Available online 21 March 2009
democratic nor autocratic politicians can commit to policies that are not ex post
JEL classi?cation:
optimal. We show that because of the dynamic commitment problem inherent in the
H20
soft budget constraint, politicians can in essence commit to make transfers to
H50
entrepreneurs which otherwise they would not be able to do. This encourages such
O20
entrepreneurs to support them politically. Though the soft budget constraint may
induce economic inef?ciency, it may be politically rational because it in?uences the
Keywords:
probability of political survival. In consequence, even when information is complete,
Political economy
politicians may fund bad projects which they anticipate they will have to bail out in the
Investment
future. We show that, maybe somewhat surprisingly, dictators who are less likely to lose
Development
power, are more likely to use the soft budget constraint as a strategy to gain political
support.
& 2009 Elsevier B.V. All rights reserved.
1. Introduction
Traditional policy analysis in the tradition of Pigou (1920) and Samuelson (1954) saw policymakers as designing policies
to solve market failures, or satisfy normative criteria, subject only to the availability of resources and the nature of
preferences and technology. In the 1970s economists began to realize that even well-intentioned planners were subject to
other types of constraints. Diamond and Mirrlees (1971) examined the nature of optimal policies without lump-sum
taxation, and Gibbard (1973) and Green and Laffont (1979) argued that the incentive compatibility constraints generated by
private information had to be respected. Kydland and Prescott (1977) also showed that optimal inter-temporal policies
might be time inconsistent, making it dif?cult for a planner to commit to even a second-best policy. In the 1980s and 1990s
economists began to merge such ideas with models where policymakers were self-interested and studied how the
interaction between such interests and social welfare led to further deviations from ?rst- or second-best outcomes.
These models have brought us much closer to an understanding of the relationship between market failures and
political failures. Yet many puzzles remain. A central, and fascinating one, is that of the ‘‘soft budget constraint.’’ Originally
introduced by Kornai (1979) in the context of centrally planned economies, the basic notion is that governments and
policymakers are unable to impose a ‘‘hard’’ budget constraint on government owned enterprises or government agencies.
In consequence such enterprises or agencies have incentives to act in inef?cient or pro?igate ways knowing that they will
à Corresponding author. Tel.: +47 73591420; fax: +47 7359 6954.
E-mail addresses: jrobinson@gov.harvard.edu (J.A. Robinson), ragnar.torvik@svt.ntnu.no (R. Torvik).
0014-2921/$ - see front matter & 2009 Elsevier B.V. All rights reserved.
doi:10.1016/j.euroecorev.2009.02.006

---

ARTICLE IN PRESS
J.A. Robinson, R. Torvik / European Economic Review 53 (2009) 786–798
787
be bailed out if things go wrong. Gregory and Harrison (2005) provide a detailed discussion of the soft budget constraint in
the Soviet planned economy.
Since its development, the problem of the soft budget constraint has been recognized to be endemic to most polities,
though clearly being worse in developing economies. This recognition emerges from the fact that all scholars note that soft
budget constraints in Eastern Europe and the former Soviet Republic proved more long lived that central planning. Maskin
and Xu (2001, p. 10) report that ‘‘considerable empirical work indicates that the soft budget constraint syndrome continues
to play an important role in virtually every transition economy, even those that have already undergone many years of
reform’’.
Why do soft budget constraints exist and persist? The central argument in the literature is that soft budget constraints
arise because politicians cannot commit not to re?nance bad projects ex post and cannot distinguish bad from good ex
ante. Given that a project is launched, it will be re?nanced as long as bene?ts cover costs. Previous costs are sunk.
Entrepreneurs know this, and submit bad projects for ?nancing in the ?rst place. This is the key argument in Dewatripont
and Maskin (1995), which has become the canonical model of soft budget constraints. This approach follows the literature
which built on Kydland and Prescott (1977) where policymakers were thought of as well intentioned and thus downplays
any political reason for the existence of soft budget constraints.
Such an approach to understanding the soft budget constraint is interesting. But it is also useful to extend such an
approach, because the overwhelming amount of evidence strongly suggests the role of political motivations in explaining
soft budget constraints. For instance, political scientists who have studied this topic, argue that the main reason for soft
budget constraints to persist is that soft budget constraints serve the interests of politicians—this is precisely the reason
they are not dismantled. This literature leaves unanswered, however, the question of exactly why the interests of politicians
manifest themselves in such a way.
In this paper we develop a fully political economy model of the soft budget constraint. Our starting point, following
Alesina (1988), Osborne and Slivinski (1996), and Besley and Coate (1997), is that politicians cannot commit to policies that
are not ex post optimal. This inability to commit to arbitrary policies hampers the ability of politicians to exchange policies
for support, since voters do not necessarily believe political promises (unlike in the basic Downsian model where perfect
commitment is assumed). Such a political setting is the natural one if one accepts that the problem of the soft budget
constraint is a problem of commitment. Instruments which solve this credibility problem are therefore potentially
attractive politically. We argue that the key thing about the soft budget constraint is that, in effect, it is a credible way of
transferring income to potential supporters: because a policymaker cannot commit to enforce a hard budget constraint, he
can commit to make transfers to citizens.
Nevertheless, this in itself does not make a soft budget constraint politically rational. Instruments which allow all
politicians to make credible commitments to policy are not necessarily attractive unless they improve the position of one
politician relative to another. For example, politicians would like to be able to offer income redistribution to groups to win
their support. In order for this offer to change the expected outcome of a political ?ght, such redistribution has to satisfy
two conditions; (1) it must be optimal ex post for politicians to enact, and (2) it must be something that all politicians
cannot offer.
Such asymmetries arise in many natural ways. Politicians differ in their valuation of welfare of different groups, in their
ability to undertake different policies, in their regional attachment, and in their interaction with different groups. For
instance, Dixit and Londregan (1996) argue that (p. 1134) ‘‘Such differences can arise when each party has its core support
groups of constituents whom it understands well. This greater understanding translates into greater ef?ciency in the
allocation of bene?ts: patronage dollars are spend more ef?ciently’’. Although in reality there may be many reasons for
politicians having different costs or bene?ts in transferring resources to different groups, the exact modelling of these
asymmetries is not crucial for our argument. We therefore simply adopt the familiar mechanism of Dixit and Londregan
(1996) where politicians have lower net costs of transferring resources to core supporters than to other groups.
In this paper we argue that it is the combination of these two things that leads to the prevalence of the soft budget
constraint. Politicians are happy to ?nance projects which are known to be bad in the sense that revenues do not cover
costs and which they anticipate that they will ?nd it optimal ex post to ‘‘bail out’’. This is because such ‘‘bail outs’’
redistribute resources to people or groups to whom they would otherwise ?nd it dif?cult to redistribute to credibly and to
whom other politicians cannot credibly redistribute resources. We refer to such groups as the core supporters of a
politician. We show that the key difference between such bad projects and good projects is that all politicians can commit
to re?nance good projects ex post and thus although they may redistribute resources to potential supporters, they do so
symmetrically and therefore do not give any politician a strategic advantage.
In many countries where soft budgets constraints are prevalent the quality of democracy is questionable at best or they
are pure dictatorships at worst. Thus it is crucial to investigate how the extent of democracy itself in?uences the
attractiveness of the soft budget constraint as a political strategy. Our starting point is the observation by Bueno de
Mesquita et al. (2003, p. 28) that to understand the political survival of dictatorships the key is to investigate how they are
able to generate suf?cient support to cling to power: ‘‘Make no mistake about it, no leader rules alone. Even the most
oppressive dictators cannot survive the loss of support among their core constituents.’’ A main difference among
democracy and dictatorship, however, is that dictators actively use the state apparatus to ensure that even if they are
supported by less than half of the population they remain in power. Thus while the minimum winning coalition under
democracy is half of the voters, the minimum winning coalition in a dictatorship may be considerably lower. To be able to

---

ARTICLE IN PRESS
788
J.A. Robinson, R. Torvik / European Economic Review 53 (2009) 786–798
study not only democracies, but also autocracies, we include this insight into our modelling. We show that, maybe
somewhat surprisingly, dictators who are less likely to lose power, are more likely to use the soft budget constraint as a
political strategy to gain political support.
The ability of incumbent politicians to launch projects that only they can credibly re?nance in the future creates an
incumbency bias. Moreover, it introduces an interesting inter-temporal structure to the model. If an incumbent politician
launches a project today which only he can re?nance tomorrow, this encourages his core supporters to support him
because they anticipate that he will bail them out tomorrow, thus increasing their utility. In addition, if such a politician
remains in power then he can launch further projects in the next period with payoff in the period after that. This further
increases the bene?t to core supporters from maintaining the politician in power. A one shot game will not capture these
inter-temporal effects, and thus we develop an in?nite horizon model.
While in Dewatripont and Maskin (1995) the soft budget constraint is something politicians would want to
escape if they credibly could, in our model the soft budget constraint may arise as something politicians desire
even when information is complete. Many case studies point out that soft budget constraints may serve political
purposes. For instance, Gimpelson and Treisman (2002) ?nd that in Russia (p. 172) ‘‘regional governments boost public
employment by hiring partisans and clients and extract greater federal aid’’ and that (p. 178) ‘‘Central politicians
responded with bailouts because they knew, too, that regional voters would, quite rationally, have punished them
if they did not’’. Kitschelt et al. (1999) discuss the widespread use of clientelistic policies in post-communist
countries. One example is Bulgaria where politicians build clientelistic networks (p. 203) ‘‘especially in the sectors of
state-run enterprises and collectivized agriculture’’ and where ‘‘quasi-private business groups in the BSPs
sphere of supporters successfully extracted cheap credits from a compliant government-controlled central bank . . . and
sold foreign commodities at high world market prices to unavailable, debt-accumulating state-owned companies’’.
These scholars see the soft budget constraint as arising out of a clientelistic exchange of redistribution for political
support.
The politics of soft budget constraints and patronage is not, however, unique to eastern European transition economies.
A large number of studies emphasize such political strategies are prevalent in African countries. The ?rst European colony
in Africa that became independent was Ghana in 1957.1 The Nkrumah government launched a policy of active involvement
in the economy, but the economic effects of the policy of patronage was disastrous; massive public investments did not
yield any payoff in terms of increased growth. Killick (1978, p. 248) argues that to understand the poor economic
development in Ghana one need to ask ‘‘why the creation of new state enterprises was allowed to outstrip the resources
devoted to project planning, why incompetent managers were tolerated and why interfering politicians were not
disciplined’’. He goes on to argue that
‘‘‘Political interference;’ emerges as a logical result of the use of state enterprises to reward party activists and to
extend the area of political control. And inattention to economic ef?ciency in the planning and operation of
enterprises becomes explicable if the creation of such enterprises is accepted as an end in itself and as an
ostentatious display to impress the electorate’’.
Political motivations for the soft budget constraint have previously been proposed by Shleifer and Vishny (1994),
Boycko et al. (1996), and Desai and Olofsga?rd (2006), who assume political bene?ts of excess labor in public ?rms that
result in soft budget constraints. A difference from these models is that in the present paper these political bene?ts
emerge as a result.2 Our model is related to models where the incumbent chooses policy to bind his own hands in
order to in?uence the outcome of an election (e.g., Milesi-Ferretti, 1995). As in such papers we study a dynamic
model of voting and commitment. Although we study different questions, our model also relates to Robinson and Torvik
(2005), because as in that paper a key mechanism is that politicians differ in what commitments they can make to different
groups of citizens. In contrast to Robinson and Torvik (2005), the present paper develops an in?nite horizon model,
allowing us to capture dynamic effects. Maybe more important, however, motivated by the case study literature the
present paper departs from an assumption of prefect democracy. We study how the degree of democracy affects the
political desirability of soft budget constraints.3 Finally, our model is related to Dixit and Londregan (1995) where agents do
not undertake ef?cient investments because politicians cannot commit not to tax away the future pro?ts by these
investments. In our model, by contrast, politicians choose policy today so that they are able to commit to a particular policy
in the future.
1 For other African examples of politically motivated soft budget constraints see Barkan and Chege (1989) on Kenya, Tangri (1999) on Zambia,
Nkurunziza and Ngaruko (2002) on Burundi, and Hayward (1987).
2 Shleifer and Vishny (1994), and Boycko et al. (1996) have no political competition or voting in their models. Desai and Olofsga?rd (2006) assume that
some voters cannot observe whose policies are being implemented to increase employment, making it attractive for incompetent politicians to increase
employment through subsidies. In a similar vein Coate and Morris (1995) explain how inef?cient redistribution may emerge in a political equilibrium
when voters do not know for sure if the implemented policies are ef?cient or inef?cient. In contrast our model has complete information.
3 In addition to the result that dictators who are less likely to lose power, are more likely to use the soft budget constraint as a strategy to gain political
support, this also clari?es that ideological heterogeneity, which does not have any effect in a perfect democracy probabilistic voting model, affects the
outcome under imperfect democracy.

---

ARTICLE IN PRESS
J.A. Robinson, R. Torvik / European Economic Review 53 (2009) 786–798
789
2. A model of politically ef?cient soft budget constraints
We consider an in?nite horizon society with two politicians A and B and a unit mass of entrepreneurs that are also the
voters (or under autocracy the potential supporters of the dictator). The starting point of our model is the two period model
of the soft budget constraint in Dewatripont and Maskin (1995). Entrepreneurs have no capital themselves and submit
projects for ?nancing to the politician in power. Projects generate observable returns R one period after being launched
given that they are re?nanced. The politician can extract these returns. Projects require one unit of capital per period they
are ?nanced.
By holding power politicians receive some exogenous rents X, get the eventual returns on projects net of investment
costs, and have the right to decide policy. The crucial assumption in our model is that different politicians have different
costs or bene?ts in transferring resources to different groups of citizens, for instance that other things equal politicians
would rather give transfers to their own region than to another region, to members of their own party rather than to
members of another party, to their own social class rather than to another social class, to their own ethnic group rather
than to another ethnic group, to their own clan rather than to another clan, and so on. Much empirical evidence support
such an assumption. For instance, in Kenya Barkan and Chege (1989) ?nd that when Moi replaced Kenyatta, expenditures
on road construction shifted from the Kenyatta political base to the Moi political base; within six years the share going to
the core supporters of Moi increased from 32% to 67% of the total, while the share going to the Kenyatta core supporters
decreased from 44% to 16%. Keefer (2002) study public investments in the Dominican Republic, and reports that (p. 27)
‘‘hundreds of projects that were begun by the government of Joaquin Balaguer, two governments before, were then
paralyzed under the Leonel Ferna´ndez government. Other observers noted that incomplete projects from the Ferna´ndez
government were similarly halted under Mej?´a’’.
To account for political asymmetries in a simple and transparent way, we model the transfers from politicians to
entrepreneurs required to ?nance a project in the same way as Dixit and Londregan (1996), where transfers occur via a
leaky bucket and where this leakage depends on the group and the party. The cost for a politician of transferring one unit of
capital to a core supporter is 1 þ j units of resources, while the cost of transferring to a non-core supporter is 1 þ y units of
resources; joyo1. The assumption that yo1 implies that politicians are not too different with respect to their costs of
transferring resources to different groups. Without loss of generality we simplify the Dixit–Londregan formulation by
setting j ¼ 0. Each politician has a fraction pp 1 of the voters in his core support group, while a fraction 1 À 2p of the voters
2
do not belong in any core support group.4 All voters freely decide who to support and thus even members of the core group
of a politician have to be persuaded to support him, they do not automatically do so.5 In a democracy a politician keeps
power if he receives at least half of the votes. In an autocracy the incumbent only needs the support of a fraction Q o 1 of
2
the people. The more autocratic the regime, the lower is the minimum winning coalition Q .
Entrepreneurs receive some exogenous income w related to their productivity, with the corresponding utility WðwÞ. Y is
the ?nal period additional income of a project being re?nanced, with the corresponding utility Wðw þ YÞ. To economize on
notation we de?ne Wðw þ YÞ À WðwÞ ¼ E. Thus with W00o0, E is higher in poor countries than in rich ones. Politicians and
entrepreneurs have a discount factor b 2 ð0; 1Þ and all agents aim to maximize the expected present discounted value of
utility. As in Lindbeck and Weibull (1987) and Dixit and Londregan (1996) entrepreneurs also have preferences over
ideology. Each voter j has an ideological (per period) bias dj toward politician A. We assume that dj is uniformly distributed
on the interval ½Àð1=2sÞ; 1=2s? with density s40, and that ideological preferences remain constant over time.6 Each
individual is also subject to an aggregate shock in favor of politician A, denoted c, which is a random variable uniformly
distributed on the interval ½Àð1=2hÞ; 1=2h? with density h40 (and measured in next period utility units). Each period a new
drawing from the popularity distribution is undertaken independently of the popularity shock of the previous period. The
expected next period (per period) utility of entrepreneur j is given by
WðwÞ þ E þ djG,
(1)
where E represents the utility gain of projects re?nanced and G is a dummy variable that takes the value of unity if the
entrepreneur votes for politician A and 0 otherwise (and the expected value of c is 0). Thus we employ a standard
probabilistic voting model based on Lindbeck and Weibull (1987) and Dixit and Londregan (1996). A difference is that we
4 Our mechanism does not hinge on how we model political the asymmetries. In the Dixit–Londregan formulation the costs of transferring resources
differ between politicians. An alternative interpretation of partisan politics is that the bene?ts differ when politicians care differently about the wellbeing
of different voters, as in Robinson and Torvik (2005) who assume that politicians are altruistic. In any case, the crucial thing is that the net cost of
transferring resources to own core supporters is lower than for transferring resources to other voters.
5 We abstract from any other type of policy instrument in the analysis. Note however that if politicians had access to transfers, since we shall analyze
only Markov perfect equilibria, they would not be able to credibly promise to use them to gain support. Once in power, since politicians value their own
welfare higher than the welfare of voters, they would give no transfers. Citizens understand this, and in our model promises of transfers will therefore not
be credible.
6 To save on notation we assume that ideology is symmetrically distributed and the same in the different groups. We could easily have allowed the
core supporters of a politician to have a distribution of ideology that is skewed toward that politician. However, as shown in, for example, Sections 8.1–8.3
in Persson and Tabellini (2000), in a model with a uniform distribution of ideology the marginal effect on the number of political supporters from
economic policy is unaffected by such an assumption, and thus we leave it out.

---

ARTICLE IN PRESS
790
J.A. Robinson, R. Torvik / European Economic Review 53 (2009) 786–798
extend the probabilistic model to consider also non-democratic regimes, and in addition that we extend the probabilistic
voting model to an in?nite horizon economy.
We consider both poor and good projects. Poor projects are projects that do not yield a positive return. Thus for poor
projects bRo1 þ b for a core supporter and bRoð1 þ bÞð1 þ yÞ for others. Here 1 þ b is the present discounted cost of a
project operated by a member of the core group of a politician and bR is the present discounted revenues. Projects that do
not ful?ll these inequalities are termed good projects.7 We assume that in addition to eventual good projects entrepreneurs
may have access to, all entrepreneurs have access to one poor project in each period.
2.1. Policies
We characterize the symmetric pure strategy Markov perfect equilibria of the model. In a Markov equilibrium actions at
a given play of the stage game can only be conditioned on the payoff-relevant state of the game at that point and not the
entire history of play. Here the state of the game is captured by the identity of the incumbent politician and existing
projects in their second period. The restriction to Markov perfect equilibria implies that strategies played within a period
must be subgame perfect which means that all actions must be credible. This introduces the problem of commitment in a
natural way. Therefore, entrepreneurs realize that for policies to be implemented they have to be ex post optimal for the
politician in power. The timing of the stage game in period t is as follows.
1. At the start of the period whichever politician won political power starts out as the incumbent.
2. He must decide whether or not to re?nance the projects launched in period t À 1.
3. He must decide what new projects to launch in period t.
4. Agents receive their period t payoffs.
5. At the end of the period entrepreneurs decide which politician to support, and the incumbent retains or looses power.
We start out with analyzing poor projects. The politician who wins political power must decide if such projects should be
launched, as well as if existing projects should be terminated or re?nanced. We start out by considering the following
equilibrium strategies: when in power a politician launches poor projects with 1oRo1 þ y for his core supporters only and
he re?nances core supporters only. We term this policy l. Note that policy l is symmetric in the sense that both politicians
act in the same way towards their core supporters, and against the core supporters of the other politician. We study the
appropriate Bellman equations, and later we show that the only possible poor projects that can be ?nanced in equilibrium
are those with 1oRo1 þ y.
Whether or not launching poor projects is an equilibrium depends on the availability of good projects, since these
in?uence the bene?ts from being in power. However, since the effects of payoffs from such good projects are identical to
the effects of X in the Bellman equations we do not explicitly introduce the payoffs from good projects. We treat them as
contained in X, and we also return to the rami?cations of good projects later.
In general, de?ne Vilða; b; mÞ as the return to politician i ¼ A; B when politicians follow policy l, a ¼ PA if the core
supporters of politician A have non-completed projects and a ¼ 0 otherwise, b ¼ PB if the core supporters of politician B
have non-completed projects and b ¼ 0 otherwise, and politician m ¼ A; B has political power. Let pðlÞ denote the reelection
probability in a democracy or the survival probability in an autocracy of an incumbent following policy l. This probability
will be endogenously determined later. Note that it is independent of who was the incumbent in the previous period as
previous period projects end before next period (irrespective of whether they are re?nanced or not). 1 À pðlÞ is the
probability for an opposition politician to win political power under policy l.
Consider ?rst the case where politician A wins political power. If politician B was the previous incumbent, the static
payoff to politician A as a consequence of winning is the rents of power X net of the unit ?nancing cost for the p
entrepreneurs in his core support group that he launches poor projects for. If keeping power also for the next period the
politician gets VAlðPA; 0; AÞ, while otherwise he gets VAlðPA; 0; BÞ. Hence his payoff is
VAlð0; PB; AÞ ¼ X À p þ bðpðlÞVAlðPA; 0; AÞ þ ð1 À pðlÞÞVAlðPA; 0; BÞÞ.
(2)
When politician A was the incumbent and wins political power his core supporters will be re?nanced. Thus in this case the
?rst period static payoff to the incumbent also includes the return R net of the re?nancing cost for his p core supporters.
VAlðPA; 0; AÞ ¼ pðR À 1Þ þ X À p þ bðpðlÞVAlðPA; 0; AÞ þ ð1 À pðlÞÞVAlðPA; 0; BÞÞ
¼ V A
l ð0; PB; AÞ þ pðR À 1Þ,
(3)
where the second line follows immediately from (2).
Should politician B win political power the static payoff for politician A is 0 regardless of whether he was the incumbent
or not, as he receives neither rents nor payoffs from projects. Furthermore, when politician B wins power the next period
7 Note that these conditions are not the same as those for social ef?ciency. These would also include the bene?ts to the entrepreneurs E. Nevertheless,
the interesting feature of the soft budget constraint seems to understand how policymakers fund loss making enterprises and the private bene?ts are not
relevant to whether or not a project make losses for the government since they cannot be expropriated.

---

ARTICLE IN PRESS
J.A. Robinson, R. Torvik / European Economic Review 53 (2009) 786–798
791
probability of politician A winning political power becomes that of an opponent; 1 À pðlÞ. Hence we have
VAlðPA; 0; BÞ ¼ VAlð0; PB; BÞ ¼ bðð1 À pðlÞÞVAlð0; PB; AÞ þ pðlÞVAlð0; PB; BÞÞ.
(4)
From (2)–(4) we then ?nd
ð1 À bpðlÞÞðX À p þ bpðlÞpðR À 1ÞÞ
VA
.
(5)
l ð0; PB; AÞ ¼
1 À bðb þ 2pðlÞð1 À bÞÞ
Symmetric equations hold for politician B.
Note that poor projects are always economically loss making to initiate; the expected payoff of a project is À1 þ
bpðlÞðR À 1Þ which must always be negative when bRo1 þ b. However, they can still be politically ef?cient and to show this
we proceed to determine pðlÞ.
There are three groups of entrepreneurs—the two groups of core supporters each of size p and the group of
entrepreneurs of size 1 À 2p that are not core supporters of either politician. Consistent with the above notation, let
UAlða; b; BÞ denote the return to a core supporter of politician A when politicians follow policy l, a ¼ PA if core the supporters
of politician A have non-completed projects and a ¼ 0 otherwise, b ¼ PB if the core supporters of politician B have non-
completed projects and b ¼ 0 otherwise, and politician B wins political power. The expected future value of the aggregate
shock c is 0. Consider an entrepreneur with ideological bias dj in the core support group of politician A. If politician A is the
incumbent the next period expected static payoff for the entrepreneur is WðwÞ þ E þ dj if the incumbent keeps power.
Furthermore, the present incumbent will also be the next period incumbent with the reelection or survival probability pðlÞ,
while the probability of losing power will be 1 À pðlÞ. Thus
UAlðPA; 0; AÞ ¼ WðwÞ þ E þ dj þ bðpðlÞUAlðPA; 0; AÞ þ ð1 À pðlÞÞUAlðPA; 0; BÞÞ.
(6)
In case the incumbent loses power, a core supporter of politician A is not re?nanced and since we measure the ideological
bias in favor of politician A, his next period expected static return is WðwÞ. Furthermore, the probability that politician A
wins political power at the end of next period is the probability that an opposition politician wins, namely 1 À pðlÞ. Thus
UAlðPA; 0; BÞ ¼ WðwÞ þ bðð1 À pðlÞÞUAlð0; PB; AÞ þ pðlÞUAlð0; PB; BÞÞ.
(7)
In case politician B is the current incumbent the expected payoffs are
UAlð0; PB; AÞ ¼ UAlðPA; 0; AÞ À E
(8)
and
UAlð0; PB; BÞ ¼ UAlðPA; 0; BÞ.
(9)
Consider ?rst the case where politician A is the incumbent. The expected future net gain in utility of a core supporter of
politician A if A remains in power is
ð1 þ bð1 À pðlÞÞÞE þ dj
UA
.
(10)
l ðPA; 0; AÞ À UA
l ðPA; 0; BÞ ¼
1 À bð2pðlÞ À 1Þ
Consider next the case where politician B is the incumbent. In this case the net gain for a group A core supporter should
politician A rather than politician B take power is given by
bpðlÞE þ dj
UA
.
(11)
l ð0; PB; AÞ À UA
l ð0; PB; BÞ ¼ 1 À bð2pðlÞ À 1Þ
Note that the right-hand side of (10) is higher than the right-hand side of (11). This is intuitive. The difference between
them stems from the fact that if A is the incumbent then he chooses to implement projects in the current period. This
implies that if he wins political power then a member of his core group will get E next period. In addition if A retains power
tomorrow he will initiate further projects in the next period, which are again re?nanced should he also win power in the
next round, and so on. However, if B is in power in the current period, then even if A is in power tomorrow, since no projects
bene?tting members of A’s core group will have been started in the current period, there cannot be any payoff E tomorrow.
However, if A gets back in power tomorrow he will then initiate projects, generating possible future bene?ts. This argument
explains why the right-hand side of (10) is higher than the right side of (11). This immediately implies that even in a
democracy the reelection probability pðlÞ will be different from the election probability of an opposition politician, 1 À pðlÞ.
The symmetric equations for core supporters of politician B can easily be found (remembering that dj is de?ned as the
ideological bias toward politician A). Also, by inserting E ¼ 0, the corresponding equations for non-core supporters can be
found.
We then reach the following proposition:
Proposition 1. A strategy where the politician in power launches poor projects with 1oRo1 þ y for core supporters only, and
re?nances core supporters only, increases his probability of remaining in power compared to a strategy of not launching and
re?nancing projects.

---

ARTICLE IN PRESS
792
J.A. Robinson, R. Torvik / European Economic Review 53 (2009) 786–798
Proof. Assume that politician A is the incumbent. Denote by NA the number of core supporters of the incumbent that also
A
support him. From (10) we know that those politician A core supporters that support politician A are those with a higher
ideological bias than the dj de?ned by Àc ¼ ð½1 þ bð1 À pðlÞÞ?E þ djÞ=ð1 À bð2pðlÞ À 1ÞÞ. NA is then given by
A
Z 1=2s


1
p
s di ¼ p
þ sð1 þ bð1 À pðlÞÞÞE þ sð1 À bð2pðlÞ À 1ÞÞc
(12)
2
Àð1þbð1ÀpðlÞÞÞEÀð1Àbð2pðlÞÀ1ÞÞc
The number of core supporters of politician B that support incumbent A is given by
Z 1=2s


1
NB
s di
.
(13)
A ¼ p
¼ p
À sbpðlÞE þ sð1 À bð2pðlÞ À 1ÞÞc
bp
2
ðlÞEÀð1Àbð2pðlÞÀ1ÞÞc
The number of non-core supporters that support the incumbent, NA, is given by
Z 1=2s


1
NA ¼ ð1 À 2pÞ
s di ¼ ð1 À 2pÞ
þ sð1 À bð2pðlÞ À 1ÞÞc .
(14)
2
À½1Àbð2pðlÞÀ1Þ?c
The reelection or survival probability, pðlÞ, is given by the probability that the total number of supporters exceeds the
minimum winning coalition Q .
pðlÞ ¼ PrfNAA þ NBA þ NAXQg,
which can be simpli?ed to
8




1 1
9
>
h 1
À Q
>
<
>
>
À Q
s 2
= 1
s 2
pðlÞ ¼ Pr cX À pE À
¼
þ phE þ
.
(15)
1 À bð2pðlÞ À 1Þ
>
>
2
1 À bð2pðlÞ À 1Þ
:
>
>
;
From this the probability can be found as
??????????????????????????????????????????????????????????
s
?


h 1
1 þ 2bphE À
ð1 À 2bphEÞ2 À 8
À Q
1
s 2
pðlÞ ¼
þ
,
(16)
2
4b
where we show in Appendix A that stability implies ð1 À 2bphEÞ2 À 8ðh=sÞðð1=2Þ À Q Þ40.
Consider next the case where politicians do not ?nance any projects, and denote the reelection or survival probability in
this case pð0Þ. Now the post election income of all entrepreneurs is independent of the election outcome. The probability of
political survival can then be found as
?????????????????????????????????
s
?


h 1
1 À
1 À 8
À Q
1
s 2
pð0Þ ¼
þ
.
(17)
2
4b
It is easy to see that pðlÞ4pð0Þ, and the proposition follows. &
To make the intuition behind Proposition 1 transparent, consider ?rst the special case of democracy, i.e. Q ¼ 1. Then the
2
reelection probability is simply given by 1 þ phE. The reelection probability is affected in three ways by ?nancing poor
2
projects for core supporters. First, the reelection probability of the incumbent politician A increases as the next period
income for core supporters of A is higher if he rather than the opposition politician B wins the election. In the ?rst case they
are re?nanced, in the second they are not.
Second, core supporters see that an increased reelection probability has value also for future periods. In addition to a
higher next period static payoff the election of politician A has the effect of increasing the probability of being ?nanced in
the future. This effect is stronger the more core supporters that are ?nanced, as this makes the reelection probability
higher. For the core supporters of politician A it is thus good news if many poor projects are launched. This dynamic effect
adds to the increased next period static payoff, and thus increases the reelection probability further.
Third, core supporters of the opposition politician realize that a higher probability of reelection of the incumbent
decreases the chance that they will receive ?nancing of poor projects in later periods. For the core supporters of politician B
it is thus bad news if many poor projects are launched by an incumbent politician A. This dynamic effect decreases the
reelection probability of the incumbent.
However, since the group of politician A and politician B core supporters is of the same size p, these two latter dynamic
effects are of exactly the same size, and the net effect on the reelection probability constitutes the ?rst static effect of
increased next period income for core supporters.
Next, consider the case of non-democracy, i.e. Q o 1. Then, since incumbents need less supporters to survive politically,
2
the ‘‘reelection’’ or survival probability increases compared to that under democracy. We note that the increase in the
survival probability is stronger the more ideologically heterogenous the population is (s low). The intuition for this is that
an ideologically heterogenous population increases the number of ‘‘extreme supporters’’ for both the dictator and the

---

ARTICLE IN PRESS
J.A. Robinson, R. Torvik / European Economic Review 53 (2009) 786–798
793
challenger. Under democracy this does not affect the reelection probability as the effect is symmetric, but under autocracy
such an increase in ‘‘extreme supporters’’ on both sides is to the advantage of the dictator. The reason for this is that the
minimum winning coalition Q is less than 1, and thus the probability that the number of supporters exceeds Q is higher the
2
more extreme supporters there are.
It is worth noting how the formula for (16) is in?uenced by our simplifying assumption that all members of both core
groups have poor projects to be ?nanced. Imagine that only a fraction q of them did, with a fraction 1 À q having good
projects. In this case under democracy (16) would be pðlÞ ¼ 1 þ qphE emphasizing that the presence of good projects does
2
not in?uence the reelection probability. Intuitively, it is only the poor projects which will in?uence the election outcome
since they will only be re?nanced by one of the politicians. Good projects will be ?nanced (and re?nanced) by both types of
politicians and will therefore not in?uence the election outcome (which we show in Proposition 4).
Also, note that in the model the incumbent pays the full cost of ?nancing bad projects which means that the ?nancing is
not hurting other voters through lower spending or higher taxation. If voters had to pay part of this cost our mechanism
would still be valid, and could actually produce a stronger incumbency bias. The reason for this is the combination of sunk
costs and rational forward looking voters. Whichever politician is elected the cost of launching bad projects would have to
be paid. However, by electing the incumbent the net cost is lower as he will make a surplus on re?nancing existing projects
in case he wins political power while the opposition will not. Finally, if a politician without any core support group entered
into political competition he would in such a setting have a political advantage among non-core supporters, which would
weaken the political advantage of politicians with core supporters. However as politicians with core supporters could still
have an advantage among the core supporters, the political survival probability may still be higher by launching bad
projects. Thus our mechanism could be present also in such a case.
We now proceed to show that the only poor projects that can affect the reelection or political survival probability are
those with 1oRo1 þ y. First, note that so far we have only assumed that politicians play the strategies associated with
policy l and drawn the implications for the reelection or survival probability. We now need to justify that (i) it is credible for
an incumbent to promise re?nancing of the loss making projects he initiates, (ii) it is not credible for the opposition to
promise to re?nance these projects, and (iii) no other poor projects can affect the political survival probability.
(i) A promise by the incumbent to re?nance projects should he remain in power is credible. Given that he launches a
project a politician will re?nance (if in power) when R41 given that previous costs are sunk.
(ii) A promise by the opposition to re?nance is not credible. Given that Ro1 þ y re?nancing non-core supporters is loss
making. Entrepreneurs realize that a promise of re?nancing should the opposition take power is not ex post optimal
for the opposition, and hence such promises are not credible.
(iii) Consider ?rst a poor project where Ro1. It is not credible for any politician to promise to re?nance such a project.
Consider next a poor project with R41 þ y. If launched in a period such a project will be re?nanced by any politician
winning power, as the investment cost from the previous period is sunk. Hence when poor projects have Ro1 or
R41 þ y the decision to re?nance is independent of who is in power. It is then straightforward to show that the
political survival probability is pð0Þ.
The following proposition is now evident:
Proposition 2. The only way for an incumbent to increase the probability of keeping power by poor projects is to launch projects
with 1oRo1 þ y.
Thus, launching poor projects for core supporters may be an ef?cient political strategy to increase the probability of
political survival. Such projects allow the incumbent to credibly promise to some entrepreneurs that their income will be
higher if the incumbent rather than the opposition wins political power. In this way the incumbent is able to tie the
continuation utility of some entrepreneurs to his own political success. The gain in supporters from own core supporters is
higher than the loss in supporters from the core supporters of the opposition politician — the incumbent is able to utilize
the advantage of deciding policy to produce an incumbency bias.
This mechanism is also valid under autocracy, and is a formalization of the intuition in Bueno de Mesquita et al. (2003,
p. 59) who note that ‘‘To depose an incumbent, a challenger needs to convince a suf?cient number of members of the
winning coalition to defect him. On the surface this appears to be a relatively easy task. All the challenger has to do is to
promise these members of the existing coalition more rewards than they currently receive. Unfortunately for the
challenger, such a promise lacks long-term credibility.’’ Furthermore (p. 60) ‘‘The incumbent leader does not face this
problem of credibility as severely, because her current supporters understand that they will continue to receive private
bene?ts as long as they remain loyal’’.
Note, however, that for poor projects to produce an incumbency bias they have to be suf?ciently poor. Marginally poor
projects in the sense that 1 þ yoRoðð1 þ bÞ=bÞ do not suf?ce to increase the reelection probability as they will be
re?nanced by both politicians should they win power.
The incumbency bias rests on the fact that redistribution will be shifted from core supporters of the old politician to
core supporters of the new politician when there is a change in power. Much case study evidence supports this result. The
regressions in Gimpelson and Treisman (2002) shows that (p. 149) ‘‘Public employment tended to fall after the election of a

---

ARTICLE IN PRESS
794
J.A. Robinson, R. Torvik / European Economic Review 53 (2009) 786–798
new governor, who presumably trimmed the patronage appointments of his predecessor.’’ With the change of political
power in Ghana in 1966 Killick (1978, p. 238) notes that the new government decided ‘‘to lay off nearly 40,000 redundant
workers in various state agencies’’.8 There is also direct evidence that the policy of patronage indeed raises the reelection
probability. According to the analysis in Treisman (1999, p. 81) ‘‘Where regional government spending increased relatively
more, the vote was subsequently higher for Yeltsin and his reformist allies, controlling for the previous level of regional
support for them.’’
Whether or not he wins or loses political power, the incumbent incurs an economic loss on poor projects. However, they
may still be rational to undertake because the probability of remaining in power increases and when in power the politician
receives rents. Consider the incumbents’ alternative strategy of not ?nancing the projects. We have already denoted this by
policy 0. In this case the per period return from winning power is independent of history and given by X. VA0ðAÞ is the payoff
to politician A from winning political power, VA0ðBÞ the payoff to politician A from losing political power, and pð0Þ the
political survival probability. Then
VA0ðAÞ ¼ X þ bðpð0ÞVA0ðAÞ þ ð1 À pð0ÞÞVA0ðBÞÞ
(18)
and
VA0ðBÞ ¼ bðð1 À pð0ÞÞVA0ðAÞ þ pð0ÞVA0ðBÞÞ.
(19)
Consequently
ð1 À bpð0ÞÞX
VA
.
(20)
0 ðAÞ ¼ 1 À bðb þ 2pð0Þð1 À bÞÞ
Note that from (5) and (20) it follows that if pðlÞ ¼ pð0Þ poor projects will never be launched (as X À p þ bpðlÞpðR À 1Þ is
always less than X). However, we already know that pð0ÞopðlÞ, since when no projects are launched, the income of all
entrepreneurs is independent of who wins political power. This leads to the following proposition:
Proposition 3. Poor projects with 1oRlo1 þ y are more likely to be launched for core supporters of the incumbent
(i) the higher the rents of power X,
(ii) the more responsive voters are to economic policy (high h),
(iii) the poorer the country (high E),
(iv) the less democratic the country (low Q ), and
(v) the more ideologically heterogenous the population is (low s) given that democracy is imperfect ðQ o 1Þ.
2
Proof. Consider VAlð0; PB; AÞ from (5) and VA0ðAÞ from (20). These both include the rents from being in power today, the
eventual costs of ?nancing projects today, and the future expected utility from following policy l or 0 respectively. Since the
rents from power are already secured an incumbent will follow policy l instead of policy 0 when the net gain of doing so is
positive, i.e. VAlð0; PB; AÞ À X4VA0ðAÞ À X. By substituting from (5) and (20) and then for pðlÞ and pð0Þ from (16) and (17), the
proposition follows by straight forward calculations that we relegate to Appendix A.
&
In the existing economics literature the soft budget constraint is viewed as the outcome of a commitment problem; if bad
projects are ?nanced initially politicians cannot credibly commit not to re?nance them as long as returns cover next period
costs. Thus, since bad entrepreneurs know that they will be re?nanced, they submit poor projects in the ?rst place. If
politicians knew that the projects were poor, they would never have been ?nanced. By contrast, in our theory politicians
?nance poor projects exactly because they are known to be poor. This is also the result of commitment, but in our case this
is viewed as an opportunity rather than as a problem by the politicians; by ?nancing bad projects for core supporters
politicians can credibly commit to re?nance the projects in case they win political power while the opposition cannot.
As seen from Proposition 3 poor projects are more likely to be chosen, and the soft budget constraint more prevalent,
when the rents of being in power X are high. Then there is more to gain by in?uencing the probability of political survival.
This may explain why the soft budget constraint is typically a problem in countries with bad institutional quality. In such
countries the rents from being in power may be high as politicians tend to view the state ?nances as their own resources.
Also, it may explain the serious problem of soft budget constraints in natural resource abundant countries.
It is interesting to note that the higher is h, the more likely it is that poor projects are ?nanced. Thus, the more
responsive entrepreneurs are to economic factors, the worse economic outcomes may be. The reason for this is that when
entrepreneurs are responsive they are easier to buy with inef?cient redistribution. This result contrasts with most other
8 What ?rst seemed to be a shift in policy away from state enterprises towards laissez faire, however, turned out not to be. Soon after the ratio of
public to private sector employees was back to the levels seen during the Nkrumah administration, and when the NRC came to power in 1972 Killick
(1978, p. 317) reports that ‘‘The government announced its intention to reactivate various state enterprises left uncompleted or abandoned after the
overthrow of Nkrumah, but of far greater signi?cance for the role of the state was the compulsory acquisition of a 55 percent shareholding in the timber,
mining and oil industries . . . It seemed, then, that economic policy had become full circle. The expansion of state control and participation begun under
Nkrumah was resumed’’.

---

ARTICLE IN PRESS
J.A. Robinson, R. Torvik / European Economic Review 53 (2009) 786–798
795
theories of the ef?ciency of electoral competition, where politicians are more strongly inclined to adopt ef?cient economic
policies the more responsive voters are to economic policy, see e.g. Persson and Tabellini (2000).
In the same manner standard political economy models suggest that if utility is concave in income then electoral
competition shall be stronger, and policy more ef?cient, in poor countries. Part (iii) of Proposition 3 delivers the opposite
result. If in a poor country the utility gain of being ?nanced is high, it makes inef?cient redistribution attractive because it
strongly affects the political survival probability. Thus poor countries may be especially prone to soft budget constraints
emerging as a political strategy.
Part (iv) of Proposition 3 suggests that the less democratic the regime the more likely that soft budget constraints
emerge. A regime that maintains power despite of little support in the population has a high survival probability. But why
should a regime less likely to loose political power use more inef?cient redistribution? There are, as can be seen from the
discussion of Eq. (29) in Appendix A, two reasons for this: ?rst, the higher is the probability of staying in power, the higher
is the expected payoff for the incumbent—and in turn when the expected rent of political power is high a marginal increase
in the political survival probability is more valuable. Second, the higher is the probability the incumbent stays in power, the
more likely the projects launched this period are re?nanced and the higher is the expected payoff for core supporters. In
turn, when the expected payoff for core supporters is high, a marginal increase in the survival of the incumbent has a large
effect on their utility. Thus the higher is the probability the incumbent keeps power in the ?rst place, the stronger is the
marginal effect on this probability of adopting policy l relative to policy 0.
A more ideologically heterogenous population does not affect the decision to launch poor projects in a democracy. A
smaller s has no effect on the reelection probability since the effect is symmetric. However, as stated in part (v) of
Proposition 3 in a non-democratic regime ðQ o 1Þ a more ideologically heterogenous population makes poor projects more
2
likely. The intuition for this is that a more heterogenous population increases the number of ‘‘extreme supporters’’ of the
autocrat and the opposition by an equal amount. But since autocrats need support of less than half of the population to stay
in power, this increase in ‘‘extreme supporters’’ is to the largest advantage of the autocrat, and his survival probability
increases. Thus a lower s sparks off a rise in the survival probability, and the same effect as with a lower Q comes into play,
making poor projects more politically attractive.
Finally, note that, as we discussed earlier, the presence of good projects makes it more valuable to be in power and has
the same effect as an increase in X. Thus the presence of rents from good projects makes it more attractive for incumbents
to launch bad projects.
We now turn attention to good projects. We then have the following:
Proposition 4. Good projects cannot affect the political survival probability.
Proof. If a good project is launched in a period it will be re?nanced by any politician holding power in the next period since
R4ð1 þ bÞ=b ) R41 þ y. Thus for good projects the decision to re?nance is independent of who wins political power.
The decision to launch projects may differ between politicians. Politician A faces a lower cost of projects to his core
supporters, and vice versa for politician B. Thus for a suf?ciently low payoff of good projects, there exists an equilibrium
path where politicians only launch projects for core supporters, but projects are always re?nanced independent of who
wins political power.
Consider ?rst the case where projects are suf?ciently pro?table that they will be launched by all politicians. In this case
the income of all entrepreneurs in all periods is independent of who wins power, and the political survival probability is
pð0Þ. Consider next the case where politicians only launch projects toward core supporters. The decision to re?nance is still
the same for both politicians. Thus in any period next period income is independent of who wins political power. Future
income, however, is not. Politician A core supporters will be better off if politician A wins as then they have additional
projects launched in the future, and vice versa for politician B core supporters. Politician A core supporters are thus more
likely to support politician A, while politician B core supporters are more likely to support politician B. The point to note,
however, is that these effects are symmetric; what politician A wins among his core supporters he looses among the
politician B core supporters, and vice versa. It can be veri?ed that this intuition is correct by using the same techniques as
those used to prove Proposition 1 to ?nd that also in this case the political survival probability is pð0Þ, which is the same as
if no good projects had been ?nanced. The proposition then follows.
&
Thus it is exactly the bad quality of poor projects that makes them politically appealing.9 By adopting poor projects, an
incumbent ensures that he can credibly offer to re?nance them, while the opposition cannot. Good projects do not have
this asymmetric feature since all politicians can credibly commit to re?nance them and they thus have a symmetric effect
on political support.
9 Proposition 4 as stated does depends on the assumption that yo1. The incentive to fund inef?cient projects comes from the fact that the politicians
are competing, which can only happen if they are not too dissimilar. If they are very dissimilar, so that it is very costly for one politician to deliver
resources to the supporters of the other, then in effect there is no competition and Proposition 4 would need to be re-stated. For example, if politician A is
not threatened with politician B trying to redistribute to his supporters, then he may be able to reward his supporters with socially ef?cient projects. Our
main mechanism still applies but in this case the model implies that relatively poor projects would be preferred to relatively better ones. Here, relatively
poor projects would be adopted, but not necessarily loss making projects.

---

Document Outline

• A political economy theory of the soft budget constraint
  • Introduction
  • A model of politically efficient soft budget constraints
    • Policies
  • Concluding remarks
  • Acknowledgments
  • References

---