The Effects of Governance On Relational and Formal Contracts: Theory and Evidence From Groundwater Irrigation Markets by Mr. Jeffrey Michler

Introduction
Irrigation Contracts
Data & Results
Conclusions
The Effects of Governance on Relational and

Formal Contracts:
Theory and Evidence from Groundwater Irrigation Markets
Jeffrey D. Michler
Steven Y. Wu
Department of Agricultural Economics

Purdue University
IRRI SSD Seminar

Los Banos, Philippines
5 November 2014
Michler (Purdue)
Contracts for Groundwater Irrigation
1/27 IRRI 2014
1 / 27
Introduction
Data & Results
Conclusions
Land Tenure Contracts
Hayami and Otsuka (1993) in The Economics of Contract Choice: An

Agrarian Perspective develop a general theory of land tenure
contracts. They look to explain...
Why there is a variety of land tenure contracts
Why some contracts are preferred to other contracts
The role of risk sharing in contract choice
Interlinkages between tenure contracts and credit
How contracts are enforced despite poor legal system
Michler (Purdue)
2/27 IRRI 2014
2 / 27
Introduction
Data & Results
Conclusions

Michler (Purdue)
2/27 IRRI 2014
2 / 27
Introduction
Data & Results
Conclusions

Michler (Purdue)
2/27 IRRI 2014
2 / 27
Introduction
Data & Results
Conclusions

Michler (Purdue)
2/27 IRRI 2014
2 / 27
Introduction
Data & Results
Conclusions

Michler (Purdue)
2/27 IRRI 2014
2 / 27
Introduction
Data & Results
Conclusions

Michler (Purdue)
2/27 IRRI 2014
2 / 27
Introduction
Data & Results
Conclusions
Groundwater Contracts
Mukherji (2004)
There has so far been no attempt at formulating a general theory of
groundwater markets. Thus, the current mode of functioning of
groundwater markets still leaves a lot of unanswered questions such
as why do several modes of water contracts coexist under seemingly
similar conditions and why do they respond differently to similar sets of
incentives and disincentives?
Michler (Purdue)
3/27 IRRI 2014
3 / 27
Introduction
Data & Results
Conclusions
Motivation
Observation
Markets for groundwater irrigation are informal and unregulated
Research Questions
1. Why do farmers choose to use a particular type of contract?
2. What is the role of different types of governance in determining
contract choice?
Michler (Purdue)
4/27 IRRI 2014
4 / 27
Introduction
Data & Results
Conclusions
Motivation
Observation
Markets for groundwater irrigation are informal and unregulated
Research Questions
contract choice?
Michler (Purdue)
4/27 IRRI 2014
4 / 27
Introduction
Data & Results
Conclusions
Implications
Contract choice involves balancing an inter-temporal trade-off:

Short-term: exploit groundwater to maximize production and
achieve or maintain food security.

Long-term: conserve groundwater to insure against risk posed by
future extreme environmental events.

How do we design interventions or new market mechanisms to ensure
groundwater conservation while continuing large Boro rice yields?
Michler (Purdue)
5/27 IRRI 2014
5 / 27
Introduction
Data & Results
Conclusions
Implications
Contract choice involves balancing an inter-temporal trade-off:

Short-term: exploit groundwater to maximize production and
achieve or maintain food security.

Long-term: conserve groundwater to insure against risk posed by
future extreme environmental events.

Michler (Purdue)
5/27 IRRI 2014
5 / 27
Introduction
Data & Results
Conclusions
Methodology
Relational Contract Theory

We adapt a theoretical model of human behavior and institutional
governance developed by Dixit (2004).

Allows us to think of contracting as a non-cooperative game with
contracts distinguished by who has the last move in the game.

Explains how contracts remain effective despite a lack of outside
enforcement.
Michler (Purdue)
6/27 IRRI 2014
6 / 27
Introduction
Data & Results
Conclusions
Types of Water Contracts

Fixed Charge - One-time fee paid by buyer at start of season for
water throughout season

Labor Charge - Buyer provides labor throughout season to seller
in exchange for water

Piece Rate - Buyer pays fee for each application throughout
season
Output Share - One-time fee paid at end of season and
dependent on crop yield

Input Share - Seller provides inputs to buyer in exchange for
larger output share

Fixed Rent - One-time fee paid by seller at start of season for use
of land
Michler (Purdue)
7/27 IRRI 2014
7 / 27
Introduction
Data & Results
Conclusions



season

larger output share

of land
Michler (Purdue)
7/27 IRRI 2014
7 / 27
Introduction
Data & Results
Conclusions



season

larger output share

of land
Michler (Purdue)
7/27 IRRI 2014
7 / 27
Introduction
Data & Results
Conclusions



season

larger output share

of land
Michler (Purdue)
7/27 IRRI 2014
7 / 27
Introduction
Data & Results
Conclusions



season

larger output share

of land
Michler (Purdue)
7/27 IRRI 2014
7 / 27
Introduction
Data & Results
Conclusions



season

larger output share

of land
Michler (Purdue)
7/27 IRRI 2014
7 / 27
Introduction
Data & Results
Conclusions



season

larger output share

of land
Michler (Purdue)
7/27 IRRI 2014
7 / 27
Introduction
Data & Results
Conclusions
The Water Selling Agent:

Cost and Signals
Cost function
c(a ) = (1/2)a 0 a
(1)
Pr(y = 1|a ) = y 0 a
(2)
x = x 0a +
(3)
Sufficient water
Crop output
Michler (Purdue)
8/27 IRRI 2014
8 / 27
Introduction
Data & Results
Conclusions

Cost and Signals
Cost function
c(a ) = (1/2)a 0 a
(1)
Pr(y = 1|a ) = y 0 a
(2)
x = x 0a +
(3)
Sufficient water
Crop output
Michler (Purdue)
8/27 IRRI 2014
8 / 27
Introduction
Data & Results
Conclusions

Cost and Signals
Cost function
c(a ) = (1/2)a 0 a
(1)
Pr(y = 1|a ) = y 0 a
(2)
x = x 0a +
(3)
Sufficient water
Crop output
Michler (Purdue)
8/27 IRRI 2014
8 / 27
Introduction
Data & Results
Conclusions

Compensation Package
x 0 a
| {z }
tra
ct
En
el
at
fo
rc
e
io
na
l
ab
le
on
on
tra
c
en
Pa
ym
xe
d
Fi
(1/2)a 0 a (4)
| {z }
os
t
y 0 a
| {z }
S
|{z}
t
U=
Fixed charge = 0, = 0, S > 0.

Piece rate 0 < < 1, = 0, S = 0.
Output share = 0, 0 < < 1, S = 0.
Michler (Purdue)
9/27 IRRI 2014
9 / 27
Introduction
Data & Results
Conclusions

x 0 a
| {z }
tra
ct
En
el
at
fo
rc
e
io
na
l
ab
le
on
on
tra
c
en
Pa
ym
xe
d
Fi
(1/2)a 0 a (4)
| {z }
os
t
y 0 a
| {z }
S
|{z}
t
U=

Piece rate 0 < < 1, = 0, S = 0.
Output share = 0, 0 < < 1, S = 0.
Michler (Purdue)
9/27 IRRI 2014
9 / 27
Introduction
Data & Results
Conclusions

x 0 a
| {z }
tra
ct
En
el
at
fo
rc
e
io
na
l
ab
le
on
on
tra
c
en
Pa
ym
xe
d
Fi
(1/2)a 0 a (4)
| {z }
os
t
y 0 a
| {z }
S
|{z}
t
U=

Piece rate 0 < < 1, = 0, S = 0.
Output share = 0, 0 < < 1, S = 0.
Michler (Purdue)
9/27 IRRI 2014
9 / 27
Introduction
Data & Results
Conclusions

x 0 a
| {z }
tra
ct
En
el
at
fo
rc
e
io
na
l
ab
le
on
on
tra
c
en
Pa
ym
xe
d
Fi
(1/2)a 0 a (4)
| {z }
os
t
y 0 a
| {z }
S
|{z}
t
U=

Piece rate 0 < < 1, = 0, S = 0.
Output share = 0, 0 < < 1, S = 0.
Michler (Purdue)
9/27 IRRI 2014
9 / 27
Introduction
Data & Results
Conclusions

Agents Maximization Problem
Incentive Compatibility Constraint assumes agents maximization
problem is
max U = S + x 0 a + y 0 a (1/2)a 0 a
(5)
Maximized Utility
U = S + 1/2( 2 + 2k + 2 ) u0
(6)
Assume agents Participation Constraint satisfied with equality
Michler (Purdue)
10/27IRRI 2014
10 / 27
Introduction
Data & Results
Conclusions

problem is
max U = S + x 0 a + y 0 a (1/2)a 0 a
(5)
Maximized Utility
U = S + 1/2( 2 + 2k + 2 ) u0
(6)
Michler (Purdue)
10/27IRRI 2014
10 / 27
Introduction
Data & Results
Conclusions

problem is
max U = S + x 0 a + y 0 a (1/2)a 0 a
(5)
Maximized Utility
U = S + 1/2( 2 + 2k + 2 ) u0
(6)
Michler (Purdue)
10/27IRRI 2014
10 / 27
Introduction
Data & Results
Conclusions

The Value of K
k = x 0y
(7)
Geometrically, k is cosine of angle between x and y

By Cauchy-Schwarz k 2 1
Economically, k is correlation between marginal effects of a on x
and y
Michler (Purdue)
11/27IRRI 2014
11 / 27
Introduction
Data & Results
Conclusions
The Water Buying Principal

Principals Maximization Problem
Maximization Problem
max V (, ) = (k + ) 1/2( 2 + 2k + 2 )
(8)
(,)
s.t. V (, ) r V EC 0
|
{z
}
(9)
Self Enforcement Constraint
r - short term gain from defecting on Relational Contract

V EC - value from defaulting to Enforceable Contract
Michler (Purdue)
12/27IRRI 2014
12 / 27
Introduction
Data & Results
Conclusions

max V (, ) = (k + ) 1/2( 2 + 2k + 2 )
(8)
(,)
s.t. V (, ) r V EC 0
|
{z
}
(9)

Michler (Purdue)
12/27IRRI 2014
12 / 27
Introduction
Data & Results
Conclusions

max V (, ) = (k + ) 1/2( 2 + 2k + 2 )
(8)
(,)
s.t. V (, ) r V EC 0
|
{z
}
(9)

Michler (Purdue)
12/27IRRI 2014
12 / 27
Introduction
Data & Results
Conclusions
Solutions
= rk /(1 k 2 )
(10a)
= (1 k 2 r )/(1 k 2 )
(10b)
Relational Contract: = 0, = 0, = 1
Fixed charge
Piece rate
Enforceable Contract: = 1, = k , = 0
Output share
Michler (Purdue)
13/27IRRI 2014
13 / 27
Introduction
Data & Results
Conclusions
Solutions
= rk /(1 k 2 )
(10a)
= (1 k 2 r )/(1 k 2 )
(10b)
Relational Contract: = 0, = 0, = 1
Fixed charge
Piece rate
Enforceable Contract: = 1, = k , = 0
Output share
Michler (Purdue)
13/27IRRI 2014
13 / 27
Introduction
Data & Results
Conclusions
Comparative Statics

When discount rates (r ) are low, RCs are more likely than ECs.
When performance signals (k ) are inaccurate, RCs are more
likely than ECs.

contract choice?
Weak punishment/governance makes RCs more likely than ECs
Michler (Purdue)
14/27IRRI 2014
14 / 27
Introduction
Data & Results
Conclusions
Comparative Statics

likely than ECs.

contract choice?
Michler (Purdue)
14/27IRRI 2014
14 / 27
Introduction
Data & Results
Conclusions
Comparative Statics

likely than ECs.

contract choice?
Michler (Purdue)
14/27IRRI 2014
14 / 27
Introduction
Data & Results
Conclusions
Survey Data
Conducted with support from
U.S. Borlaug Fellowship, IRRI &

Bangladesh Ag. University
96 villages selected randomly
from 48 Boro rice growing

districts
10 households randomly
selected from each village

Collected at end of Boro rice
season (May - July) 2013
Michler (Purdue)
15/27IRRI 2014
15 / 27
Introduction
Data & Results
Conclusions
Personal Discount Rates
We use simple inter-temporal

choice experiment to elicit
individual interest rates
We convert to discount rates

for regression
r=
1
1+i
Michler (Purdue)
16/27IRRI 2014
16 / 27
Introduction
Data & Results
Conclusions
Performance Signal
We use soil quality as our

performance signal
(x,y )
k = VarCov
(x)Var (y )
As soil quality , Var (y ) and
Michler (Purdue)
17/27IRRI 2014
17 / 27
Introduction
Data & Results
Conclusions
Third-Party Enforcement
Michler (Purdue)
18/27IRRI 2014
18 / 27
Introduction
Data & Results
Conclusions
Punishment Types
Given third-party arbitration

exists, we observe three types
of punishment
Use indicators for each type,

with no third-party arbitration
as base case
Michler (Purdue)
19/27IRRI 2014
19 / 27
Introduction
Data & Results
Conclusions
Econometric Model
Pr(RC = 1|X) = (0 + 1 ri + 2 ki + 3 VjEC + Zi 1 + Wi 2 )

ri - discount rate
ki - performance signal
VjEC - severity of village level punishment
Zi - vector of buyer characteristics
Wi - vector of well characteristics
We instrument ri with distance between buyers and sellers house
Michler (Purdue)
20/27IRRI 2014
20 / 27
Introduction
Data & Results
Conclusions
Econometric Model
Pr(RC = 1|X) = (0 + 1 ri + 2 ki + 3 VjEC + Zi 1 + Wi 2 )

ri - discount rate
ki - performance signal
VjEC - severity of village level punishment
Zi - vector of buyer characteristics
Wi - vector of well characteristics
We instrument ri with distance between buyers and sellers house
Michler (Purdue)
20/27IRRI 2014
20 / 27
Introduction
Data & Results
Conclusions
Determinants of Contract Choice

All Water Buyers
Base: Enforceable Contract
LPM
Probit
IV
Discount Rate (r )
Discount Rate (0-12 months)
9.048
(5.182)
0.031
(0.113)
0.481
(0.941)
0.028
(0.023)
0.001
(0.029)
0.322
(0.230)
0.638
(0.258)
0.455
(0.172)
0.722
(0.225)
0.133
(0.118)
0.074
(0.038)
0.155
(0.071)
0.307
(0.719)
0.702
(0.335)
1.485
(0.321)
0.249
(0.538)
0.505
(0.373)
1.244
(0.483)
Performance Signal (k )
Standard Soil Quality
Good Soil Quality
Punishment (V EC )
Private Punishment
Economic Punishment
Social Punishment
Michler (Purdue)
21/27IRRI 2014
21 / 27
Introduction
Data & Results
Conclusions
Effect of Governance by Contract Type

Village Enforcement
Base: Fixed Charge
Share
Piece Rate
No Village Enforcement
Share
Piece Rate
Discount Rate (r )
Discount Rate (0-12 months)
1.905
(3.308)
3.911
(2.104)
0.633
(3.799)
5.228
(2.955)
0.868
(0.500)
1.303
(0.773)
0.282
(0.454)
0.051
(0.646)
1.029
(0.866)
2.156
(0.931)
0.179
(0.507)
0.031
(0.439)
3.003
(1.533)
4.104
(1.534)
0.467
(0.716)
0.498
(0.588)
Accuracy Signal (k )
Standard Soil Quality
Good Soil Quality
Punishment (V EC )
Economic Punishment
Economic Social
Michler (Purdue)
22/27IRRI 2014
22 / 27
Introduction
Data & Results
Conclusions
Conclusions
High discount rate decreases probability of relational contracts.
Existence of an accurate performance signal decreases
probability of relational contracts.

contract choice?
Severe punishment of enforceable contracts decreases probability
of relational contracts.
Michler (Purdue)
23/27IRRI 2014
23 / 27
Introduction
Data & Results
Conclusions
Conclusions
High discount rate decreases probability of relational contracts.
Existence of an accurate performance signal decreases
probability of relational contracts.

contract choice?
Severe punishment of enforceable contracts decreases probability
of relational contracts.
Michler (Purdue)
23/27IRRI 2014
23 / 27
Introduction
Data & Results
Conclusions
General Policy Implications

Adopt a systematic approach to analyzing the market
Policies must consider the governance structure and institutional
framework
Attempts to strengthen governance may result in sub-optimal
outcomes
Michler (Purdue)
24/27IRRI 2014
24 / 27
Introduction
Data & Results
Conclusions

framework
outcomes
Michler (Purdue)
24/27IRRI 2014
24 / 27
Introduction
Data & Results
Conclusions

framework
outcomes
Michler (Purdue)
24/27IRRI 2014
24 / 27
Introduction
Data & Results
Conclusions

framework
outcomes
Michler (Purdue)
24/27IRRI 2014
24 / 27
Introduction
Data & Results
Conclusions
Specific Policy Implications
If goal is to incentive households to conserve water...

Output share wastes water most but is good for credit constrained
buyers
Fixed charge conserves water most but is bad for credit
constrained buyers
Piece rate balances these two issues
Michler (Purdue)
25/27IRRI 2014
25 / 27
Introduction
Data & Results
Conclusions

buyers
constrained buyers
Michler (Purdue)
25/27IRRI 2014
25 / 27
Introduction
Data & Results
Conclusions

buyers
constrained buyers
Michler (Purdue)
25/27IRRI 2014
25 / 27
Introduction
Data & Results
Conclusions
Questions Remain
Does price of piece rate contract account for "true" price of water?
Village governance may be able to address the common pool
resource problem
But, better governance makes piece rate less likely
How does one design productivity enhancing, water conserving,

incentive compatible contracts within a given governance structure?
Michler (Purdue)
26/27IRRI 2014
26 / 27
Introduction
Data & Results
Conclusions
Questions Remain
resource problem

Michler (Purdue)
26/27IRRI 2014
26 / 27
Introduction
Data & Results
Conclusions
Questions Remain
resource problem

Michler (Purdue)
26/27IRRI 2014
26 / 27
Introduction
Data & Results
Conclusions
Thank You
QUESTIONS?
Michler (Purdue)
COMMENTS?
SUGGESTIONS?
27/27IRRI 2014
27 / 27

The Effects of Governance On Relational and Formal Contracts: Theory and Evidence From Groundwater Irrigation Markets by Mr. Jeffrey Michler

Transféré par

Informations du document

Titre original

Copyright

Formats disponibles

Partager ce document

Partager ou intégrer le document

Options de partage

Avez-vous trouvé ce document utile ?

Ce contenu est-il inapproprié ?

Droits d'auteur :

Formats disponibles

The Effects of Governance On Relational and Formal Contracts: Theory and Evidence From Groundwater Irrigation Markets by Mr. Jeffrey Michler

Transféré par

Droits d'auteur :

Formats disponibles

Introduction

Data & Results