Transcript ADI_7:Organized Political Corruption - uni

Anticorruption and the Design of Institutions 2010/11
Lecture 7
Organized Political
Corruption
Prof. Dr. Johann Graf Lambsdorff
Literature
ADI 2010/11
 Lambsdorff, J. Graf (2007), The New Institutional Economics of
Corruption and Reform: Theory, Evidence and Policy. Cambridge
University Press: 81-108.
 Myerson, R. (2008), “Perspective on Mechanism Design in Economic
Theory”, American Economic Review Vol. 98 (3): 600-602.
 Olson, M. (1993), "Dictatorship, Democracy, and Development," The
American Political Science Review, Vol. 87(3): 567-576.
 Shleifer, A. and R.W. Vishny (1993), ”Corruption.” Quarterly Journal
of Economics, Vol. 108: 599–617.
Disorganized Corruption
ADI 2010/11
 Businesspeople might sometimes be confronted with many public
servants extorting payments.
 For example, setting up a new business may require various permits,
none of them can be missed. The organization of these multiple
payments becomes a daunting task.
 Shleifer and Vishny [1993] model this situation: In Russia private
investors negotiate with various departments (fire, police, banks ...),
each of them requesting a payment in exchange for the speedy
approval of a business license.
 This leads to the “overgrazing” of the corrupt marketplace.
Disorganized Corruption
ADI 2010/11
 The problem is comparable to road bandits.
 Optimal income for road bandits requires a limit on their activity, so
as to avoid that too many travelers are scared off.
 Taking, for example, only 20% of the traveler’s money may be best.
 But competition among road bandits destroys this calculus: Any
money not confiscated will be taken by others.
 In this spirit, investors prefer to deal with only one extortionist.
Parker and Hart, June 30 1999
Disorganized Corruption
ADI 2010/11
Exercise:
Assume that there is a total number of n symmetric departments (i,
i=1, 2, ..., n) requesting a percentage (bi) of the investment (X). The
costs for dealing with this payment, providing the corrupt service and
obfuscating are cbi, 0<c<1. Investments (X) decrease with the total
amount of bribes,
X=X(b1+b2+...+bn), dX/dbi<0.
Each department’s profit (Pi) is given by
Pi=biX-cbiX.
Let us assume that the departments are controlled by a strong party (a
monopoly), which maximizes total income of all departments. In this
case, they are all forced to behave symmetrically (b1=b2=...=bn),
resulting in
Pi=(1-c)biX(nbi).
Disorganized Corruption
ADI 2010/11
Profit is maximized in the monopoly case if:
d Pi
dX
 (1  c) X  (1  c)bi n
0
d bi
dbi
After the breakdown of the party the departments start to maximize
their income individually, each considering the other departments
not to react to their own level of pay-off. Profit maximization is
carried out individually by the departments, resulting in:
d Pi
dX
 (1  c) X  (1  c)bi
0
d bi
dbi
The second term is no longer multiplied by n. Therefore, a higher bi
is needed to satisfy the equation. This is what is known as an
“overgrazing” effect. The optimum level of total bribes collected
(which was determined in the monopoly case) is missed, because
individual bribes are too high.
The Economist, May 3rd 2007: “Rules of the road”
ADI 2010/11
In many parts of the world, greasing the palms of corrupt officials is a fact of life. For lorry drivers
in poor countries, for example, it can be even more commonplace than paying tolls or taxes. The
extortionists are often armed police or soldiers, manning roadblocks, who make little attempt to
pretend what they do is legal. Yet, although such people may operate outside the law of the land,
they do obey the law of supply and demand. That, at least, is the finding of a recent study (“The
Simple Economics of Extortion: Evidence from Trucking in Aceh”, by Benjamin Olken and
Patrick Barron) of illicit transactions in Indonesia. In a rare attempt at documenting bribery, it
shows how crooked officials act as independent monopolists, maximising their profits and
employing sophisticated pricing schemes in ways that have an economic logic, as well as a
criminal one.
The authors studied bribes paid by lorry drivers along two main roads in Aceh, an
Indonesian province where separatist guerrillas had long been active. Over nine months in 2005
and 2006, data gatherers accompanied the truckers on 304 trips to and from Aceh, recording more
than 6,000 illegal pay-offs at military roadblocks, police checkpoints and weigh stations. During
the 637-kilometre (396-mile) trip from Medan to Meulaboh, for example, drivers typically passed
through 27 checkpoints and forked out a total of $23 in bribes, representing roughly 13% of the
cost of their trip—more, even, than the total wages of those in the truck.
In the early stages of each trip, transactions were typically conducted without
negotiation—at each stop, drivers simply handed over a few thousand rupiah ($0.50 to $1.00) or a
couple of packs of cigarettes. As the trucks neared their destination, however, checkpoint officials
demanded increasingly larger sums. At each stop, drivers found themselves with a progressively
stronger incentive to avoid hassle and safeguard their cargo, which gave the extortionists greater
power over them.
Midway through the study, after the Indonesian
ADI 2010/11
government had signed a cease-fire with the rebels, it
began a phased withdrawal of 30,000 troops, leading to
the number of checkpoints falling by half. This gave the
researchers the chance to see how the extortion market
would adjust. As you would expect, the amount lost to
bribery decreased—but only by 36%.
Fewer stops meant fewer bribes, but this was
offset by a rise in the amounts demanded at the remaining
checkpoints, whose operators, with entrepreneurial zeal,
seized the chance to capture part of the newly liberated surplus. In fact, they behaved just like
monopolists, setting their prices so as to maximise their own revenue, without considering the
response of the fellow at the next checkpoint, or whether their activities would deter truckers.
The result was a textbook example of double marginalisation—when a chain of independent
monopolists charges more, but receives less overall than a single monopoly.
These results have intriguing implications for stamping out corruption. In theory, a
centralised system of bribery is better for both truckers and extortionists than a string of corrupt
officials acting independently. Taking out a kingpin may yield splashy headlines, but it can
actually increase the total amount paid in bribes if it destroys co-ordination among his
henchmen. A better bet, say the authors, is to reduce the number of troops and police officers.
That would, indeed, be a novel way of fighting crime.
Disorganized Corruption
ADI 2010/11
 Is disorganized corruption better or worse than organized corruption?
 This question can be posed more provocative: Should we help
extortionists to better organize their illegal business?
 Yes: The model predicts that disorganized corruption is worse.
 No: Often there exist legal alternatives to the payment of bribes, even
for victims of extortion. Disorganized extortion is more costly and
induces some actors to prefer the legal alternative. Once corruption is
organized, it becomes deeply embedded in a country and widely
accepted as the only way for doing business. The model fails in
recognizing such legal alternatives such as complaints mechanisms.
Disorganized Corruption
ADI 2010/11
This can be illustrated by the following game in extensive form. B
requests to a bribe to share a joint surplus of 12. E chooses between
paying the bribe or complaining and thus resisting the payment, which
yields him a bonus payment of 2 (and a positive social externality of 10,
denoted by the logo of “doctors without borders”). Nature then tosses a
coin that determines whether the bribe market is organized or
disorganized. The latter implies that E and B must share their surplus
again with others (the effect of E withdrawing from the market is
disregarded here). Complaining is the dominant strategy.
But if the risk of a disorganized bribery is deleted, paying the
bribe is dominant. The “bad” outcome of disorganized bribery thus has
welfare enhancing repercussions.
(0; 5)
Complain
B
Request
Bribe
Bribery
disorganized
(2; 2)
Bribery
organized
(6; 6)
N
E
Pay Bribe
The Self-Seeking Government
ADI 2010/11
 How should we define corruption when the government (the principal)
is maximizing only his own interests and disregards public interests?
 The term corruption is misplaced when applied merely to a
disobedient agent. Instead, the principal’s own self-seeking behavior
may be termed corrupt.
 ”misuse” is not clearly related to the trespassing of rules, because
rules are themselves the result of self-seeking:
– Markets are distorted where this profits the principal.
– Public resources are allocated to further the principal’s economic
interests.
– The principal is above any rules and immune to any accusations.
The Self-Seeking Government
ADI 2010/11
 Harden [1993: 186-7 and 208-13] reports about the devastating
environmental effects of the Turkwel Gorge Dam in Kenya and how it
was motivated by overpricing and kickbacks. But feasibility studies did
not have to be falsified, they were just forbidden to be carried out in the
first place. When the Financial Times published information from a
whistleblower, copies of the newspaper were seized at Nairobi Airport.
 Inefficiencies that can be detected are there by design.
 In contrast to a single agent, the principal (government) can follow its
corrupt goals in a much more systematic way.
 The principal does not have to circumvent laws and regulations but
can design them to serve his own interests.
The Self-Seeking Government
ADI 2010/11
 The principal (government) thus takes advantage of his monopoly
position.
 A common conclusion is that prices are likely to increase. Market
restrictions and price increases already occur with corrupt agents. But
principals can impose such restrictions in a much more forceful way by
making them part of law.
 While we may feel desperate about such a ruler, welfare losses are
difficult to prove. Imagine a “perfect” kleptocrat.
 Welfare losses reduce the kleptocrat’s potential to take away income
from the citizenry; a kleptocrat therefore dislikes welfare losses.
 A kleptocrat attempts to organize a corrupt system to operate like a
tax.
The Self-Seeking Government
 He contains low-level corruption among the bureaucracy.
 He prevents individuals from ”overgrazing” the market.
 He dislikes low quality in public procurement.
 He dislikes white-elephant projects.
 He attempts to take more from those who can give more.
 So, why worry about this type of corruption?
ADI 2010/11
The Self-Seeking Government
ADI 2010/11
Welfare losses due to monopoly price setting
Return from
investment
Consumer surplus
with maximum price
Supply
Dead Weight Loss
Producer surplus
with maximum price
Demand
0
sectors
The Self-Seeking Government
ADI 2010/11
The kleptocrat can levy the burden equally by price discrimination
Return from
investment
Kleptocrats surplus
Supply
Demand
0
sectors
The Self-Seeking Government
ADI 2010/11
Wizard of ID,
Parker and Hart,
September 17, 2000
The Self-Seeking Government
ADI 2010/11
 Some observers argue that a kleptocrat, also labeled as a predatory
regime, tends to squeeze his citizens without pity.
 Indeed, a kleptocrat is able to gather large bribes. So why do we fail in
proving welfare losses?
 It is argued that the “stationary bandit” exercises power consistent
with the interests of society, (Olson 1993).
 A ruler has an ”encompassing interest” (unlike the predator).
 Murphy, Shleifer and Vishny [1993: 413] argue that the problems with
corruption are mitigated when corrupt rulers can collect bribes
efficiently. Perfecting corruption rather than fighting it is the avenue
suggested for reform.
The Self-Seeking Government
ADI 2010/11
 Example: One of the grandchildren of President Suharto in Indonesia
attempted to make a cut from taxes on beer which was collected by his
private company. But as a result tourism in Bali was suffering from a
shortage of beer and inflated prices, forcing President Suharto to
withdraw the tax.
 Counterexample: One illustration of a strong and corrupt government
is that of Mobutu in Zaire, an uncontested kleptocrat for decades. But
his regime crippled the economy and established a strong example that
public welfare bitterly suffers from such regimes.
The Self-Seeking Government
ADI 2010/11
 In reality various problems exist for kleptocrats.
 They assign property rights to comrades and contestants (rather than
to those making best use of resources) in exchange for loyalty.
 A flourishing economy may threaten their power.
 A short time horizon motivates them to “run with the loot while they
can”.
Wizard of ID, Parker and Hart, May 11, 2000
 Kleptocrats set a bad precedent for lower level bureaucrats.
 But, is an eternal dynasty of uncontested rulers the solution?
The Problem of the Kleptocrat
ADI 2010/11
 Investments require the committing of resources, hoping that their initial
investment will be honored.
 Such resources cannot easily be transferred or assigned to different
tasks. Railroads cannot be removed, power plants cannot be relocated to
different countries. Investors fear for the expropriation of their future
revenues. They carry out investments if political promises to honor their
revenues are credible.
 In a survey of business people in Karnataka, India, it was found that the
software industry was less affected by the high level of corruption among
the local administration. Compared to the construction and manufacturing
industries these units could easily shift assets outside the state because
this industry depends less on immovable assets. This lower dependency
seems to have reduced extortionate demands for bribes among public
officials. For other sectors credibility is essential because investments
cannot be redeployed.
The Problem of the Kleptocrat
ADI 2010/11
 Corrupt tax collectors can impose an excessive tax burden on
investors; corrupt regulators may threaten an arbitrary application of
the law; corrupt customs authorities may control trade and demand
their cut; corrupt politicians may threaten an unfavorable application or
drafting of the law. All these actors may be in a position to appropriate
the investor’s future revenues.
 We note in passing that investors fear corruption not only due to
kleptocrats or politicians who may extort them. Also private contractors
are frequently in a position to behave opportunistically and expropriate
an investor. Corrupt courts will not help an investor in protecting his
property but sell the judgment to the highest bidder.
The Problem of the Kleptocrat
ADI 2010/11
 Investors want to rely on binding laws, but corruption motivates rulers
to arbitrarily change the law and expropriate resources.
 Strong corrupt rulers are motivated only by self-enrichment and face
no restrictions.
 They cannot credibly commit to effective policies.
 Due to kleptocrats' failure to make binding commitments, they are not
trustworthy to investors.
 Investments will be stopped unless the ruler is effectively limited.
The Problem of the Kleptocrat
ADI 2010/11
 There are no easy approaches to
avoiding adverse welfare effects of
corruption.
 An organized type of corruption brings
about other forms of adverse welfare
effects than disorganized corruption.
 Changing the type of corruption is not a
convincing approach for reform.
The Problem of the Kleptocrat
ADI 2010/11
 Even kleptocratic rulers may have an
(apparently very limited) interest to
engage in anti-corruption.
 This is somewhat comparable to Mafia
bosses who attempt to legalize their
business in a strategy to avoid the
hazards for their offspring.
 Totalitarian rulers must seek ways to
commit themselves to their announced
policies, which opens the door to reform.
The Problem of the Kleptocrat
ADI 2010/11
 Myerson (2008) provides a good model for the credibility problem faced
by autocrats.
 Assume production, Y, in a country to be determined by capital, K, and
raw materials, n, according to Y=(K+n)0.5.
 An autocrat may honor the investments and allow a return of rK. The
variable r is the country’s discount rate, which would just suffice to keep
capital in the country. The autocrat would thus tax (or take as bribes) only
the difference, which is (K+n)0.5-rK. The current value of his income would
be ((K+n)0.5-rK)/(r+b). The variable b is an additional discount rate due to
the autocrat’s exogenous risk of losing power.
 An autocrat may also expropriate the capital, which suggests that in
future periods capital is no longer provided by international investors
(K=0). The current value of his income would then be K+n0.5/(r+b).
 The autocrat’s moral hazard constraint is thus
((K+n)0.5-rK)/(r+b) ≥ K+n0.5/(r+b)
The Problem of the Kleptocrat
ADI 2010/11
 The ideal capital stock that maximizes ((K+n)0.5-rK)/(r+b) is obtained by
the first derivative, (0.5(K+n)-0.5-r)/(r+b) =0  (K+n)-0.5=2r. With r=0.05 we
obtain K=100-n. Let, for example, n=12 and we see that the moral hazard
constraint is violated. For any K>0 this results, suggesting that only K=0
(n=100) is feasible.
 The autocrat may now liberalize his regime, for example by designing
constitutional constraints or countervailing powers that limit his
capacity to expropriate. The extent of liberalism l, 0 ≤ l ≤1, reduces his
income from expropriation according to (1-l)[K+n0.5/(r+b)].
 There is also a random risk for an honest autocrat to be overthrown,
which is 0.05.l. His income from honesty thus reduces to
((K+n)0.5-rK)/(r+b+0.05l),
which he seeks to maximize by setting l subject to his moral hazard
constraint,
((K+n)0.5-rK)/(r+b+0.05l)≥(1-l)[K+n0.5/(r+b)]
The Problem of the Kleptocrat
ADI 2010/11
 This moral hazard constraint is depicted in the figure below, alongside
with the isopayoff-curve.
Liberalization and Capital with n=12
IsopayoffCurve
100
Capital K
80
Moral
Hazard
Constraint
60
40
20
0
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
Liberalization l
 Optimal liberalization is achieved with l=0.504, inviting for capital
investments K=52.4
The Problem of the Kleptocrat
ADI 2010/11
 In case of n=0 the autocrat’s optimal regime would have l=0 and
K=44.2.
Liberalization and Capital with n=0
100
IsopayoffCurve
Capital K
80
Moral
Hazard
Constraint
60
40
20
0
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
Liberalization l
 Intuitively, a lack of natural resources makes it more costly for the
ruler to lose his reputation for protecting capital, and so he can credibly
encourage substantial investments even without liberalizing. Examples
of such countries may be Singapore, Monaco or Liechtenstein.
The Problem of the Kleptocrat
ADI 2010/11
 In case of n=25 the autocrat’s optimal regime would have l=0 and K=0.
Liberalization and Capital with n=25
100
IsopayoffCurve
Capital K
80
Moral
Hazard
Constraint
60
40
20
0
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
Liberalization l
 Intuitively, a great wealth of natural resources makes the ruler
unwilling to accept the additional political risk from liberalization, even
though it means that nobody will invest in the nation without
liberalization. Examples may be Saudi Arabia and Venezuela.
Appendix
ADI 2010/11
Discussions
1) Why may a disorganized form of corruption be worse than an
organized form, for investors and public servants alike; in how far may
independent departments “overgraze the market” when requesting
bribes?
2) Describe the type of corruption that emerges when self-seeking is the
actual object of government!
3) How may in this case allocation be distorted?
4) Show that in a static model a “perfect kleptocrat” avoids welfare
losses!
5) Why is such a static analysis misleading?
6) What may motivate autocrats to abstain from taking bribes?
7) In how far is abundance of raw materials important for an antocrat’s
decision to restrict his actions by liberalizing?
ADI 2010/11
Exercise:
There is a total number of 4 symmetric departments (i, i=1, 2, 3, 4)
requesting a percentage (bi) of the investment (X). Costs for
providing and obfuscating are cbi with c=0.5. Investments decrease
with the total amount of bribes,
4


X  4800 1   bi 
 i 1 
a) The departments jointly maximize total income. Determine the
resulting investment!
b) The departments maximize their income individually, each
considering the other departments not to react to their own level of payoff. Determine the resulting level of investment!
c) A powerful kleptocrat may unite the departments, but one type of
welfare loss becomes particularly strong and unavoidable. Describe this
type of welfare loss!