Reforming an Institutional Culture of Corruption: A Model of Motivated Agents and Collective Reputation

Reforming an Institutional Culture of Corruption:
A Model of Motivated Agents and Collective Reputation
Justin Valasek
-WZB Berlin, CESifo-
SITE Conference: Fighting Corruption in Developing and
Transition Countries
August 31, 2015
Justin Valasek (WZB) Reforming Institutional Culture SITE: Fighting Corruption 1 / 20

Motivation
“There are three kinds of [nonmaterial] rewards: a sense of duty and
purpose, the status that derives from individual recognition and
personal power, and the associational beneﬁts that come from being
part of an organization...that is highly regarded by its members or by
society at large.”
∼ Wilson (1989) “Bureaucracy”

Motivation
Some agents motivated by the mission of a public institution.
Handy-Katz (1998), Francois (2000), Besley-Ghatak (2005).

Motivation
“Public Sector Eﬃciency Wage:”
Below-market wage → disproportionately select motivated.

Motivation
“Public Sector Eﬃciency Wage:”
Below-market wage → disproportionately select motivated.
(Contrasts with corruption “eﬃciency wage” (Besley-McLaren
(1993)).)

Motivation II
Public sector often more prosocial in developed nations, often
with wages below private sector.
Gregg et al. (2011), Dur and Zoutenbier (2012), etc.

Motivation II
India, Indonesia: Preference for public sector → less prosocial.
Hanna and Wang (2014), Banuri and Keefer (2014).

Motivation II
Mexico: Higher wage has (+) eﬀect on motivation.
Dal Bo et al. (2013).

Motivation II
Mexico: Higher wage has (+) eﬀect on motivation.
Dal Bo et al. (2013).
This paper adds: “the associational beneﬁts that come from
being part of an organization...that is highly regarded by its
members or by society at large.”

Outline of Approach
Analyze worker selection into institutions/ﬁrms A, B.

Outline of Approach
Motivated workers more productive in mission-oriented
institution A (no corruption). [Besley-Ghatak]

Outline of Approach
Motivated workers value collective reputation:
(+) if high reputation, (−) if low reputation.

Outline of Approach
1 Prosocial signaling (B´enebou-Tirole 2006).
2 Identity (Akerlof and Kranton 2000, 2005).
3 Value-homophily (Lazarsfeld and Merton 1954).

Outline of Approach
(1) Show that model can organize the empirical data:
developed/developing.

Outline of Approach
(1) Show that model can organize the empirical data:
developed/developing.
(2) Implications of the model for reform (wage, not laws).

Results Preview
Positive Question: Characterization of equilibria [Static].
Multiple equilibria (high/low motivation).

Results Preview
Comparative statics motivation-wage:
High-motivation reputation: ↑ wage crowds out motivation.

Results Preview
Low-motivation reputation: ↑ wage crowds in motivation.

Results Preview
Normative Question: Can wage be used as policy tool to
transition from high to low corruption?

Results Preview
Normative Question: Can wage be used as policy tool to
transition from high to low corruption?
Non-monotonic dynamic transition:
1 Increase wage – attract motivated ⇒
2 Gradual decrease wage – “push” non-motivated types out.*

Literature Review
Monetary incentives aﬀect non-monetary incentives:
Titmuss (1970), Gneezy-Rustichini (2000), B´enebou-Tirole (2003,
2006), Besley-Ghatak (2005).
Intrinsic incentives in the workplace:
Workplace norms: Dewatripont et al. (1999), Akerlof and
Kranton (2005), Dixit (2002), Murdock (2002), Sliwka (2007),
Ellingsen and Johannesson (2008), Prendergast (2008); Bandiera et
al. (2010).
Worker motivation and selection: Handy-Katz (1998),
Francois (2000), Casellia-Morelli (2004), Besley-Ghatak (2005),
Prendergast (2007), Delfgaauw-Dur (2008), Macchiavello (2008),
Auriol-Brilion (2014), Aldashev et al. (2015).
Dynamic reform of institutions:
Tirole (1996), Bisin-Verdier (2001), Bidner-Francois (2013),
Acemoglu-Jackson (2014, 2015).

Framework: Agents
Two institutions: A and B.
A: collective reputation and motivation relevant.
B: outside option.
Continuum of workers of measure 1, index set I.
λ motivated; (ai = 1).
(1 − λ) non-motivated; (ai = 0).
Employed in A (pi = 1), or in B (pi = 0).
Workers have heterogeneous private-sector ability.
Ability in B, xi ∼ U[x, x].
Independence, f(xi|ai) = f(xi).
Ability in A homogenous.

Framework: Payoﬀs Institutions
Institution B: πj = xi − wi.
Market fully competitive: wB
i = xi.
Institution A, demand measure ν < 1:
πj = πp
+ βai − wi
ai unobservable.
Aggregate output observed → C = I piai/ I pi observed.
wA
i = wA.

Framework: Payoﬀs Individuals
Non-motivated workers: U(wi, pi) = wi.
Motivated workers, conditional on working for A:
1 May value mission of A.
2 Value the collective reputation of A.
Ua(wi, pi, C) = wi + v(C)pi,
v (·) > 0, v (·) < 0.
Generic ﬁrm v(λ) = 0; Mission-oriented institution v(λ) > 0.

Framework: Employment
Employment in B always available → wi = xi “outside option.”
Employment in A capped at ν.
Worker i chooses ˆpi ∈ {0, 1}.
pi,t
= 0 if ˆpi = 0
= 1 w.p. q if ˆpi = 1,
q = min 1,
ν
I ˆpi
.

Framework: Equilibrium + Preliminaries
Deﬁnition (Equilibrium, Nash)
An equilibrium is deﬁned given wA
, by {ˆpi} such that {ˆpi}
maximizes worker utilities, given C = I
ˆpiai/ I
ˆpi.

ˆpiai/ I
ˆpi.
Lemma (Cutoff Equilibrium)
Given wA
, equilibrium is characterized by {xa
, xb
}, where motivated
workers choose public sector iff xi ≤ xa
, non-motivated iff xi ≤ xb
.

ˆpiai/ I
ˆpi.
Lemma (Cutoff Equilibrium)
Given wA
, equilibrium is characterized by {xa
, xb
}, where motivated
workers choose public sector iff xi ≤ xa
, non-motivated iff xi ≤ xb
.
Definition (Reputation Motivated/Non-Motivated)
Reputation is Motivated if C > λ and Non-Motivated if C ≤ λ.

Existence of High/Low Equilibria
Proposition (Existence of Market-Clearing Equilibria)
1 Generic Firm (v(λ) = 0):
Low-motivation equilibrium exists.
High-motivation equilibrium exists if ν < λ(x + v(1)).
2 Mission-Oriented (v(λ) > 0):
High-motivation equilibrium exists.
Low-motivation equilibrium exists if ν < (λ − 1)(x + v(0)).
Both High and Low Reputation equilibria exist when:
1 ν small.
2 −v(0), v(1) large.

Example: Mission-Oriented (v(λ) > 0)
Ua pi 0
Ua pi 1
xb
wA
0.2 0.4 0.6 0.8 1.0
xa
0.0
0.2
0.4
0.6
0.8
1.0
Ua wi, xa
, C xa

Dynamic Model: Transitioning between Equilibria
Dynamic Model: inﬁnitely repeated version of static model.
Paper: Measure δ ∈ (0, 1] workers randomly “replaced” per
period, binding tenure.
Presentation: All workers replaced each period (δ = 1).

Friction: Motivated workers lagged reputation:
Ua,t(wi, Ct−1) = wi,t + v(Ct−1)pi,t

Friction: Motivated workers lagged reputation:
Ua,t(wi, Ct−1) = wi,t + v(Ct−1)pi,t
Question: Given initial low reputation, does a wage path
exist that transitions to high reputation steady state equilibrium
{wA∗
, C∗
}?

Ct as a function of wA
t
Lemma (Crowding out/in motivation)
1 If low reputation (v(Ct−1) ≤ 0), then Ct increasing in wA
t .
2 If high reputation (v(Ct−1) > 0), then Ct decreasing in wA
t .

t
t .
t .
Linear distance xa
t , xb
t ﬁxed by v(Ct−1).

t
t .
t .
Linear distance xa
t , xb
(1) Low wage:
(1) High wage:

t
t .
t .
Linear distance xa
t , xb
(2) Low wage:
(2) High wage:

Existence & Characterization Transition
Proposition (Existence of Transition)
Given δ = 1, a wage path that transitions from low reputation,
v(C0) < 0, to high reputation steady-state, {wA∗
, C∗
}, exists iﬀ
v(λ) > 0.

Transition path {wA
0 , C0} ⇒ {wp∗
, C∗
}
v(λ) > 0; starting from C0 < λ.

Transition path {wA
0 , C0} ⇒ {wp∗
, C∗
}
Period 1: Set high wage (wA
1 + v(C0) = x).

Transition path {wA
0 , C0} ⇒ {wp∗
, C∗
}
1 + v(C0) = x).
All workers apply: C1 = λ.

Transition path {wA
0 , C0} ⇒ {wp∗
, C∗
}
1 + v(C0) = x).
Period 2: Set market-clearing wage (wA
2 solves I
ˆpi = ν).

Transition path {wA
0 , C0} ⇒ {wp∗
, C∗
}
1 + v(C0) = x).
2 solves I
ˆpi = ν).
Reputation payoﬀ positive (v(C1 = λ) > 0) ⇒ xa > xb.

Transition path {wA
0 , C0} ⇒ {wp∗
, C∗
}
1 + v(C0) = x).
2 solves I
ˆpi = ν).
C2 > λ.

Transition path {wA
0 , C0} ⇒ {wp∗
, C∗
}
1 + v(C0) = x).
2 solves I
ˆpi = ν).
C2 > λ.
Period 3: Set market-clearing wage.

Transition path {wA
0 , C0} ⇒ {wp∗
, C∗
}
1 + v(C0) = x).
2 solves I
ˆpi = ν).
C2 > λ.
Since C2 > C1, wA
3 < wA
2 .

Transition path {wA
0 , C0} ⇒ {wp∗
, C∗
}
1 + v(C0) = x).
2 solves I
ˆpi = ν).
C2 > λ.
Since C2 > C1, wA
3 < wA
2 .
Ct decreasing in wage ⇒ C3 > C2.

Transition path {wA
0 , C0} ⇒ {wp∗
, C∗
}
1 + v(C0) = x).
2 solves I
ˆpi = ν).
C2 > λ.
Since C2 > C1, wA
3 < wA
2 .
Ct decreasing in wage ⇒ C3 > C2.
Period t: Set market-clearing wage.............{wA∗
, C∗
}.

Visual
wt
p
1 Ct
0 2 4 6 8 10
t

Example: No transition path C0 ⇒ {wA∗
, C∗
}
v(λ) = 0; starting from C0 < λ.
Period 1: Set high wage.
C1 ≤ λ.
Period 2: v(Ct−1) ≤ 0 ⇒ C2 is decreasing in wage.
C2 ≤ λ for any wA
2 .
v(λ) = 0, no {wA
} exists that transitions C0 ⇒ {wA∗
, C∗
}.

Optimal Transition: How big of a wage increase?
Proposition (Minimum Budget)
The following wage path minimizes the maximum budget required to
transition from any C0 such that v(Ct−1) ≤ 0 to {wA∗
, C∗
}:
wA
t =
x + v(0) for t s.t. v(Ct−1) ≤ 0
Market-Clearing for t ≥ min{t|v(Ct−1) > 0}
Set wA
1 high enough lowest-ability motivated worker joins A.

Minimum Budget
xb
0.2 0.4 0.6 0.8 1.0
xa
0.0
0.2
0.4
0.6
0.8
1.0
u wa , xa

Exploiting Correlation Ability and Motivation
Two public-sector ability types yn
∈ {yl
, yh
}: yh
greater
average motivation.
Value average reputation → transition if v(¯λ) ≥ 0.
Improve reputation through disproportionate hiring of h type.
Hire only yh (high wage) until C > ¯λ → market-clearing wage.

∈ {yl
, yh
}: yh
greater
average motivation.
Corruption reform through creating “elite” units:
1 Screen ability-type with high average motivation.

∈ {yl
, yh
}: yh
greater
average motivation.
Poor reputation → problem of adverse selection.

∈ {yl
, yh
}: yh
greater
average motivation.
Poor reputation → problem of adverse selection.
2 Couple screening with high wage to solve adverse-selection
problem.

Conclusion
Mechanism for transforming an institution’s reputation through
the selection: Transition is likely a long-term process.
Complementary to changing institutional norms (conformity).
Works even if wage change for new workers only.

Conclusion
Initial point matters: starting from low reputation, higher wages
crowd in motivation.
High Corruption: higher wages decrease corruption directly
(eﬃciency wage) and indirectly (selection).
Caveat: Mechanism to assign over-demanded jobs minimally
non-corrupt.

Conclusion
Initial point matters: starting from low reputation, higher wages
crowd in motivation.
High Corruption: higher wages decrease corruption directly
(eﬃciency wage) and indirectly (selection).
Caveat: Mechanism to assign over-demanded jobs minimally
non-corrupt.
How to endow a new elite institution with a good reputation:
high-motivation ability type + high wage to avoid adverse
selection.

Takeaways
Thanks!

Bounding the analysis: Robustness
Institutional norms: Complementary to changing institutional
norms.
Direct eﬀect wages on behavior: Orthogonal to changing direct
incentives for corrupt behavior.
Transition through decreasing wages: Market-clearing wages for
new workers.
Over-demanded jobs auctioned oﬀ: need minimal amount of
random allocation.

Reforming an Institutional Culture of Corruption: A Model of Motivated Agents and Collective Reputation

Recommended

Recommended

More Related Content

What's hot

What's hot (20)

Viewers also liked

Viewers also liked (9)

Similar to Reforming an Institutional Culture of Corruption: A Model of Motivated Agents and Collective Reputation

Similar to Reforming an Institutional Culture of Corruption: A Model of Motivated Agents and Collective Reputation (20)

More from Stockholm Institute of Transition Economics

More from Stockholm Institute of Transition Economics (20)

Recently uploaded

Recently uploaded (20)

Reforming an Institutional Culture of Corruption: A Model of Motivated Agents and Collective Reputation