Roy, Saktinil_Presentation

advertisement
Saktinil Roy
July 2nd, 2010
What does the current literature
tell us?
 Most
work analyzing the causes and
consequences of corruption assume a
principal-agent framework.
 A government official is an agent to the public
entrusted with performing tasks, but its actions
and the specific circumstances are not fully
observed or monitored.
 A dishonest government official either makes
illegitimate claims in return of its service or
engages in illegal deals with private agents
looking to evade the law (e.g. tax evasion or
violation of environmental regulations).
What does the current literature
tell us? (contd..)

Notice two key elements
 The bureaucrat is “dishonest” and
“opportunist”
 Only two parties interacting –
the government/supervisor and the
bureaucrat or the client and the bureaucrat
What does the current literature
tell us? (contd..)

Policy implications
 Monetary incentives to government
bureaucrats – the efficiency wage argument
(Becker and Stigler, 1978; Rose-Ackerman,
1978)
 Increase in economic and political
competition (notably, Shleifer and Vishny,
1993)
Problems with this approach

The assumption of agent dishonesty for all types of
corruption to occur is troubling in face of the twin facts:
 Corruption is widespread in many less
developed countries and also in some advanced
countries
 Real life examples where people are motivated
by non-monetary incentives are at least as
widespread

So, can we safely assume that it is always due
to imperfect monitoring of dishonest agents?
Problems with this approach
(contd..)

The assumption of “dishonest” agents is not
always quite consistent with recent findings
that indicate
 agents are emotionally attached to the idea of
“fairness” (Akerlof and Shiller, 2009)
 often people are happy to be part of an organization.
So, part of their satisfaction depends on how well
they work to fulfill the goals of the organization or
what others think how well they performed (Akerlof
and Kranton, 2005; Besley and Ghatak, 2005)
Problems with this approach
(contd..)
 Most
work also ignores how
corruption at the group level affects
corruption at the individual level – an
important criterion.
 In a different strand some work done in
this area (Tirole, 1996; Andvig and
Moene, 1990; Cadot 1987; Sah 1988)
Problems with this approach
(contd..)

Does not incorporate the influences of third
parties from outside the government
 However,
 real world instances suggest that government
officials are quite often under pressure from
private influential parties
 Individuals are, at times, caught by adverse
circumstances (e.g. Basu, 1986, 2000)
 So, can “honest” bureaucrats be
“corruptible”?
This is my point of departure
Problems with this approach
(contd..)

Overall, the literature tends to ignore that
 like social practices such as caste, rat race,
and racial discrimination, types of corruption
that exhibit a recurrent pattern are also social
phenomena – the causes and consequences
are pretty much implied by the society’s
institutional structure.
 corrupt incidents are not always acts of
bureaucrats who are “dishonest” by nature
or due to lack of monetary incentives.
The approach of the present
research


The literature has taken an extreme stance:
“dishonest” agents engaging in corruption.
I address the issue from the other extreme: “honest”
agents engaging in corruption due to third party
influence?
 So, corruption in the context of triadic relationships
 Triadic relationships in economics: Akerlof (1976), Basu
(1986)
 Cite real world examples
The idea: we get better insights into the many
realistic possibilities where the government is partly
benevolent and bureaucrats are partly honest.
The Model

Context
 Assistance programs for the rural poor in an LDC –
such as Swarnjayanti Gram Swarozgar Yojana in India
(IRDP until 1999)
 Selection of beneficiaries depending on skills and
background.
 The poor performance of these programs is partly due
to “poor targeting” – often assistance provided to
people who deserve less or nothing.
 According to NABARD: as high as 70 percent in the state
of Andhra Pradesh and 47 percent in the state of Gujrat in
India
 Fenichel and Smith (1992) discuss failure of IRDP in
Zambia due to poor targeting.
The Model (contd..)
 Bureaucratic corruption is often thought to be responsible for
“poor targeting.”
 However, several accounts on these programs suggest that
power relations in traditional rural societies – particularly as
they pertain to the role of the rural elite – has to do with poor
targeting.
 Succinctly described by the World Bank (World Development
Report, 2000/01): even if sufficient fiscal autonomy is vested
to a local government, political manipulation and rent seeking
by local elites (quite often landlords) might turn that privilege
against the poor.
Question:
How this is possible as an equilibrium outcome?
The Model (contd..)

I examine this with a very simple model
 Three agents:
 The government official (G):
 Honest – gets zero utility from bribe
 Positive utility from reputation of an able bureaucrat
 A potential beneficiary (L)
 A member of the rural elite (Y) who buys labor
from L
 Based on documented information, G
determines how much assistance L and
other beneficiaries should receive.
The Model (contd..)


Unlimited surplus labor – the wage rate is fixed
at the subsistence level.
Under normal conditions, Y maximizes utility by
choosing the amount of work to be bought from
L – but L should receive at least its reservation
utility.
 Gives a take-it-or-leave-it offer to L
Question:
Can Y intervene in the assistance provision
process and get a larger amount of assistance
for L, part of which it enjoys itself?
The Model (contd..)

The sequence of moves
Y decides whether or not to intervene in the
assistance provision process
 G offers an amount of assistance to L
 L accepts or rejects G’s offer
 Y decides whether or not to buy labor from L – if
it decides to buy then it gives a take-it-or-leave-it
offer .
 L accepts or rejects Y’s offer
The Fair Outcomes
Full information: Y does not intervene, L
receives the legitimate amount of
assistance, and Y buys labor from L.
 Asymmetry of information about L’s
eligibility: Y does not intervene, G provides
L with an amount of assistance based on a
probability distribution, and Y buys labor
from L.
Note: these can be realized as parts of
subgame perfect Nash equilibria when Y
adopts the strategy of “no interference.”

Corruption under Influence and
Coercion

Corruption occurs when G deviates from the public
mission by providing a greater amount of assistance
to L than what it deserves.

Asymmetry of Information:
 G seeks Y’s help to learn more about L’s eligibility
 Y deliberately overstates L’s eligibility
 G believes partly and alters the probability distribution about
L’s eligibility for all possible amounts of assistance.
 As a result, G might provide L with a greater amount of
assistance than what it deserves
 Part of the additional illegitimate amount is extracted by Y via
the offer of employment
 L accepts this arrangement as long as it is better off with the
employment from Y than with the whole assistance amount
but no employment.
Corruption under Influence and
Coercion (contd..)

Overt threat under full information:
 Y can put forward an overt threat of harassment
to G and ask to provide an illegitimate, greater
amount of assistance to L.
 It will do so if the cost involved in harassing Y is
less than the extra benefit from part of the
additional illegitimate assistance amount that it will
extract from L.
 G will provide a greater illegitimate amount of
assistance to L if the loss of utility due to
harassment is greater than the loss of utility due
to bad reputation of a corrupt bureaucrat
Corruption under Influence and
Coercion (contd..)

Covert threat under full information
 Y can adopt and announce the following strategy.
“G should provide L with a greater amount of
assistance than what it deserves. If G does not
obey me and L still accepts G’s offer then I will not
hire L.”
There is no direct threat to G, the threat is only to
L. But it amounts to an indirect threat to G.
Corruption under Influence and
Coercion (contd..)

For a specified amount the above strategy is
“credible” and will lead to a subgame perfect
Nash equilibrium if the following hold.
 The loss of utility due to bad reputation of a corrupt
bureaucrat is less than the loss of utility due to bad
reputation resulting from part failure of the assistance
program.
 L is better-off with an employment from Y than with the
legitimate assistance amount and no employment from
Y.
 Y can find a substitute for L without any additional cost
– which is possible under unlimited surplus labor.
 The off-the-equilibrium path along which the threat will
be carried out can be realized as a Nash equilibrium.
Corruption under Influence and
Coercion (contd..)
Note that in ALL three cases G will be
labeled as a “corrupt” bureaucrat and
suffer a loss of reputation, although it
does NOT accept any bribe.
Possible similar incidents in other
contexts, such as tax collection, granting
business license, pollution control, or
arms purchase contracts.
Policy Implications
 Increased monetary incentives to bureaucrats is a
solution that is incomplete at best – could even
misdirect economy’s resources.
 Increased economic and political competition in the
public sphere as well as in private sphere.
 One way of achieving the above goal is to provide
greater entitlement to the vulnerable section of the
society. But, note that the status quo of concentration
of power and wealth is an impediment to realizing this.
 A vicious circle in effect.
 Greater law enforcement, better transparency of
governmental procedures, and greater participation of
people in the decision making process.
Download