MIT OpenCourseWare
http://ocw.mit.edu
14.771 Development Economics: Microeconomic issues and Policy Models
Fall 2008
For information about citing these materials or our Terms of Use, visit: http://ocw.mit.edu/terms.
14.771: Firms and Contracts Lecture 2
Ben Olken
October 2008
Olken ()
Firms/Contracts Lecture 2
10/08
1 / 30
Overview
Last lecture: problems in contract enforcement lead to other types of
contract enforcement mechanisms (e.g., reputations, networks)
This lecture: what are the implications of weak contract enforcement
for how …rms are structured?
Business groups.
Some problems with business groups (tunneling)
Family …rms
Olken ()
Firms/Contracts Lecture 2
10/08
2 / 30
Business Groups
Weak contract enforcement suggests that more is likely to be done
within the …rm
Olken ()
Firms/Contracts Lecture 2
10/08
3 / 30
Business Groups
Weak contract enforcement suggests that more is likely to be done
within the …rm
Other theories might also suggest integration across industries (i.e.,
unrelated production functions)
Olken ()
Firms/Contracts Lecture 2
10/08
3 / 30
Business Groups
Weak contract enforcement suggests that more is likely to be done
within the …rm
Other theories might also suggest integration across industries (i.e.,
unrelated production functions)
Access to …nance also means more may be done within the …rm in
places where …nance is less developed (Rajan and Zingales 1998)
Branding/reputations (discussed at the end of last lecture) suggests
reasons for …rms to integrate across sectors
Olken ()
Firms/Contracts Lecture 2
10/08
3 / 30
A diversi…ed business group
Diagram of the Slim Helu Group removed due to copyright restrictions.
See Perkins, Morck, and Yeung (2006).
Olken ()
Firms/Contracts Lecture 2
10/08
4 / 30
Conglomerate Size and Financial Development
Cross-country evidence mixed
Olken ()
Firms/Contracts Lecture 2
10/08
5 / 30
Conglomerate Size and Financial Development
Cross-country evidence mixed
Note that one shouldn’t necessarily count cross-country evidence too
much! One should view this as motivation for better micro studies.
Olken ()
Firms/Contracts Lecture 2
10/08
5 / 30
Conglomerate Size and Financial Development
Cross-country evidence mixed
Note that one shouldn’t necessarily count cross-country evidence too
much! One should view this as motivation for better micro studies.
Acemoglu, Johnson, and Mitton (2005) study vertical integration
worldwide
For each industry, they use US input-output tables to determine how
much input from each industry is required to produce a given type of
output
They then calculate using each …rm’s SIC codes what percent of the
…rm’s inputs are produced by industries in which the …rm operates
Olken ()
Firms/Contracts Lecture 2
10/08
5 / 30
Conglomerate Size and Financial Development
Cross-country evidence mixed
Note that one shouldn’t necessarily count cross-country evidence too
much! One should view this as motivation for better micro studies.
Acemoglu, Johnson, and Mitton (2005) study vertical integration
worldwide
For each industry, they use US input-output tables to determine how
much input from each industry is required to produce a given type of
output
They then calculate using each …rm’s SIC codes what percent of the
…rm’s inputs are produced by industries in which the …rm operates
Findings:
Vertical integration is greater in poorer countries, and in countries with
greater cost of contract enforcement
Also greater in countries with greater entry cost
However, this is due almost entirely to industrial composition
So it’s not clear whether other factors cause these industries to be
more appropriate for developing countries, or vice-versa
Olken ()
Firms/Contracts Lecture 2
10/08
5 / 30
Acemoglu, Johnson, and Mitton Results
Actual vertical integration
BGD
7
CHN
NPL IRY
BOL
URY
TWN
HTI VNM
ECUITA
USA GTM MEX
CZE
COL
5
PAK
PAN
PHL
BEL
GBR
VRN
RRT
CHE CAN
ARG MYS
JAM
PER
ESP
IRSR
AUS
4
3
IND
LKA
6
PLW
NOR
FINAUT
DNK
0
.2
VUT
CRI
CIV
.4
.6
Contract enforcement cost
5
4
3
IND
LKA
PAK
PAN
TWN
PHL
ITA
USA
VNM
NLD
CZE MEX
COL
SVKCHI
GBR
BEL
DOM
HUN
CHE
JPN
POLMYS
SGP
PER
ESP MHL
AUS
ISR RL
CRI
NOR PLW
FIN AUT
DNK
0
.2
NPL
URY
ECU
SLB
GTM
VUT
.4
.6
HND
.8
6
5
4
0
.5
CHE
JPN
.1
1.5
Credit market development
2
BGD
1
LE
GUY
7
IND
5
4
3
CHN
MAC
CYM
PAK LKA PRY
IDA
NPL MRT
BMJ
BHS
URY
BOL BLZ
PAN
TWN
LUX
KNA
PHL
ITA
HTI
ANT
VNM ECU GTM
NLD USA
COL MEX CZE
SLV
SLB
BHK BRB
THA
DEU
HNB
DOM
GRD VEN
PRI SWG CHE
MMR
HUN
JAM
ARG
SUR
ESP SGP
DMAPER
ISR AUS
MCO
CRI
SMR
MNP
PYI
LCA
FRO NOR
PLW
BRN
CIV
DNK
DUM
ADO
6
6
7
Entry cost
Panel C: Vertical integration and entry costs
CHN
MAC
DLZ
PAN
BOL BHS
LUX
KNA
PHL
ITA
HTI EOU
USA
GTM ANM CZE
NLD
MEA
SLV
SVK
CHL
SLB
THA DEU GBR
HND BRB
BEL
YEN
CRD
PRT
YUTSWE
MMR
CAN
MYS
SUR PBR
TTO DMA ATG AUS ESP SGP
IRL
BR
CRI
LCA
NCR
FIN
CIV QAT
AUT
DNK
8
Vertical integration index
Vertical integration index
6
CHN
IND
LKA
Panel B: Vertical integration and credit market development
EGD
8
GUY
7
3
.8
Panel A: Vertical integration and contracting costs
7
BGD
8
Vertical integration index
Vertical integration index
8
8
9
Log GDP per capita
10
11
Panel D: Vertical integration and per-capita GDP
Figure by MIT OpenCourseWare.
Olken ()
Firms/Contracts Lecture 2
10/08
6 / 30
Acemoglu, Johnson, and Mitton Results
Vertical integration predicted by industry mix
-8
BGD
Vertical integration propensity
Vertical integration propensity
-8
-9
MRT
NPL
LKA CHN
IND
PAK
BOL
GEM
PRY
SLV NIC
PHL
CHL
URY VNM
COL
DOM
TWN
CIV
MEX HTI
JAM HND PAN
JPN THA
ARO
PER
HUN GBRITA
VEN PLW
USA
MYS
CRI
DNK
CAN NCD IRL
ISR
BEL
CHE
NOR
AUT
-10
-11
-12
-13
0
.2
VUT
.4
.6
Contract enforcement cost
Vertical integration propensity
Vertical integration propensity
-10
-13
NOR
0
.5
AUT
JPN
CHE
1
1.5
Credit market development
2
-8
BGD
-9
-11
-12
MRT
NPL
CHN
GUV
PAK
DOL MAC
GTMECU
PRYVNMPHL DMA
SLV
URY CHL
HTI COL
DOM
MEX
IND ELZ
PAN
MMR
SUR
TTO
THA
ARC
BRB
ITA
VEN PER
VCT GRD
GBR USA
HUN
LUX
MYS
CRI
CZE
SVK
AUS ESP
DEV DNK
SLB
IRL
RBD
CAN
POL SWE
SGP
BELISR
FIN
Panel B: Vertical integration propensity and credit market development
-8
-12
-10
-14
.8
Panel A: Vertical integration propensity and contracting costs
BGD
LKA
CHN
ECU
URY
PHL
VNM
CHL
DOL
TWN
DOM
MEX PAN
THAJPNJAM
ARG
PER
GBR PLW VEN HUN
CRI
USA CZE ESP
IRU
SLB
POLMHL
ISR
BEL
FIN CHE
NOR
AUT
0
.2
NPL
IND
PAK
GTM
.4
.6
Entry cost
VUT
HND
.8
Panel C: Vertical integration propensity and entry costs
1
BGD
MRT
IND LKA
CHN
PAKBOL
MAC
GUY
ECU GTM
PHL
VNM NICIDN
HTI
MEX MNP TWN
HNDBLZ
PAN
VUT CIV
CYMBMU
BHS
MMR
SURLCA THATTO
JPN
ITA
PERVEN
GBR USA
GUN
VCT
CRI MYS
SPK
SLB
POL
ISR BEL
BRN FIN CHE
NOR
GUMAUT
CRL
ADO
-10
NPL
-12
-14
6
7
8
9
Log GDP per capita
10
11
Panel D: Vertical integration propensity and per-capita GDP
Figure by MIT OpenCourseWare.
Olken ()
Firms/Contracts Lecture 2
10/08
7 / 30
Khwaja, Mian and Qamar (2008)
What are the bene…ts of being in a network?
Olken ()
Firms/Contracts Lecture 2
10/08
8 / 30
Khwaja, Mian and Qamar (2008)
What are the bene…ts of being in a network?
Data on networks:
Authors have data on directors of all Pakistani companies, public and
private
De…ne two …rms as connected if they share a common director
De…ne two …rms are in the same network if they can be linked through
connected …rms
Since they have the entire universe of …rms, they can construct
networks for the whole economy
Olken ()
Firms/Contracts Lecture 2
10/08
8 / 30
Khwaja, Mian and Qamar (2008)
What are the bene…ts of being in a network?
Data on networks:
Authors have data on directors of all Pakistani companies, public and
private
De…ne two …rms as connected if they share a common director
De…ne two …rms are in the same network if they can be linked through
connected …rms
Since they have the entire universe of …rms, they can construct
networks for the whole economy
They …nd that there is one very large "super network"
Contains 5% of …rms
But 66% of bank credit!
Olken ()
Firms/Contracts Lecture 2
10/08
8 / 30
Khwaja, Mian and Qamar (2008)
What are the bene…ts of being in a network?
Data on networks:
Authors have data on directors of all Pakistani companies, public and
private
De…ne two …rms as connected if they share a common director
De…ne two …rms are in the same network if they can be linked through
connected …rms
Since they have the entire universe of …rms, they can construct
networks for the whole economy
They …nd that there is one very large "super network"
Contains 5% of …rms
But 66% of bank credit!
Empirical question:
What is the value of being in the super-network?
Olken ()
Firms/Contracts Lecture 2
10/08
8 / 30
Constructing a network
Courtesy of Asim Ijaz Khwaja, Atif Mian, and Abid Qamar. Used with permission.
Olken ()
Firms/Contracts Lecture 2
10/08
9 / 30
Networks in the economy
Courtesy of Asim Ijaz Khwaja, Atif Mian, and Abid Qamar. Used with permission.
Olken ()
Firms/Contracts Lecture 2
10/08
10 / 30
Empirical strategy
Compare super-network vs. non super-network …rms.
Olken ()
Firms/Contracts Lecture 2
10/08
11 / 30
Empirical strategy
Compare super-network vs. non super-network …rms.
Problem? Sign of bias?
Olken ()
Firms/Contracts Lecture 2
10/08
11 / 30
Empirical strategy
Compare super-network vs. non super-network …rms.
Problem? Sign of bias?
Do the same, but with …rm …xed e¤ects
Olken ()
Firms/Contracts Lecture 2
10/08
11 / 30
Empirical strategy
Compare super-network vs. non super-network …rms.
Problem? Sign of bias?
Do the same, but with …rm …xed e¤ects
Where does variation come from?
Problem? Sign of bias?
Olken ()
Firms/Contracts Lecture 2
10/08
11 / 30
Empirical strategy
Compare super-network vs. non super-network …rms.
Problem? Sign of bias?
Do the same, but with …rm …xed e¤ects
Where does variation come from?
Problem? Sign of bias?
Empirical idea: use incidental …rm entry and exits from the
super-network
Olken ()
Firms/Contracts Lecture 2
10/08
11 / 30
Empirical strategy
Compare super-network vs. non super-network …rms.
Problem? Sign of bias?
Do the same, but with …rm …xed e¤ects
Where does variation come from?
Problem? Sign of bias?
Empirical idea: use incidental …rm entry and exits from the
super-network
I.e., not whether your …rm entered or exited the super-network, but
whether another …rm in your network entered or exited the
supernetwork
Problem? Sign of bias?
Olken ()
Firms/Contracts Lecture 2
10/08
11 / 30
Empirical strategy
Courtesy of Asim Ijaz Khwaja, Atif Mian, and Abid Qamar. Used with permission.
10/08
Olken ()
Firms/Contracts Lecture 2
9 / 30
10/08
12 / 30
Results on Borrowing
Estimate
Yit = αi + αkt + αt + γ∆Yi ,t
1
+ β1 ENTRYit + β1 ENTRYit Directi + εi
Courtesy of Asim Ijaz Khwaja, Atif Mian, and Abid Qamar. Used with permission.
10
Olken ()
Firms/Contracts Lecture 2
Results on Probability of Default
Courtesy of Asim Ijaz Khwaja, Atif Mian, and Abid Qamar. Used with permission.
Note: coe¢ cients multiplied by 100
Olken ()
Firms/Contracts Lecture 2
10/0
10/08
14 / 30
Mechanisms
Courtesy of Asim Ijaz Khwaja, Atif Mian, and Abid Qamar. Used with permission.
10/0
Olken ()
Firms/Contracts Lecture 2
10/08
15 / 30
Bertrand, Mehta, and Mullainathan (2002)
What is the downside of being in a network?
Olken ()
Firms/Contracts Lecture 2
10/08
16 / 30
Bertrand, Mehta, and Mullainathan (2002)
What is the downside of being in a network?
Control rights over a …rm and cash ‡ow rights over …rm’s pro…ts are
not identical:
Control rights are awarded to whoever has a majority
Cash ‡ow rights are awarded in proportion to ownership
Olken ()
Firms/Contracts Lecture 2
10/08
16 / 30
Bertrand, Mehta, and Mullainathan (2002)
What is the downside of being in a network?
Control rights over a …rm and cash ‡ow rights over …rm’s pro…ts are
not identical:
Control rights are awarded to whoever has a majority
Cash ‡ow rights are awarded in proportion to ownership
With pyramid ownership structures, these can be totally separated:
Principal owns 51% of company i
Company i owns 51% of company i + 1
As i ! ∞ principal retains complete control but has 0 cash ‡ow rights
Olken ()
Firms/Contracts Lecture 2
10/08
16 / 30
Tunneling
With these types of corporate structures, those with control rights
have incentives to expropriate minority shareholders
Olken ()
Firms/Contracts Lecture 2
10/08
17 / 30
Tunneling
With these types of corporate structures, those with control rights
have incentives to expropriate minority shareholders
How?
Give loans to other …rms in groups at arti…cially high/low interest rates
Sell assets to other …rms in the group at arti…cially high/low prices
Etc
Olken ()
Firms/Contracts Lecture 2
10/08
17 / 30
Tunneling
With these types of corporate structures, those with control rights
have incentives to expropriate minority shareholders
How?
Give loans to other …rms in groups at arti…cially high/low interest rates
Sell assets to other …rms in the group at arti…cially high/low prices
Etc
Why do we care?
If minority shareholders will be expropriated, means that business
groups will have trouble attracting equity …nance for their entities
This o¤sets the potential bene…ts of business groups discussed above
Olken ()
Firms/Contracts Lecture 2
10/08
17 / 30
Tunneling
With these types of corporate structures, those with control rights
have incentives to expropriate minority shareholders
How?
Give loans to other …rms in groups at arti…cially high/low interest rates
Sell assets to other …rms in the group at arti…cially high/low prices
Etc
Why do we care?
If minority shareholders will be expropriated, means that business
groups will have trouble attracting equity …nance for their entities
This o¤sets the potential bene…ts of business groups discussed above
Point of this paper is to detect tunneling
Olken ()
Firms/Contracts Lecture 2
10/08
17 / 30
Methodology
Idea: Consider external shock to predicted pro…ts, and examine how
actual pro…ts respond to predicted pro…ts
Predictions:
Actual pro…ts should respond less to predicted pro…ts if …rm is in a
group
Response is smaller the lower the cash ‡ow rights of the controlling …rm
Controlling …rm’s pro…ts should be more responsive to the bottom
…rm’s shocks than their cash ‡ow rights would imply
Response is greater if they have low cash ‡ow rights
(this I don’t see–seems to ignore actual pro…ts)
Asymmetry: bottom …rms pro…ts are not sensitive to top …rm’s shocks
This distinguishes tunneling from insurance
Olken ()
Firms/Contracts Lecture 2
10/08
18 / 30
Data
Outcome: Pro…ts Before Interest Depreciation and Taxes (PBIDT)
Shocks: Average asset-weighted industry returns (excluding your …rm)
Why exclude your …rm?
Cash ‡ow rights:
Measure direct cash ‡ow rights with several proxy variables:
Cash ‡ow rights of directors (likely to be from the controlling group)
Cash ‡ow rights of "other shareholders" (not directors, …nancial
instituions, government bodies, corporate bodies, nor top …fty
shareholders)
No measure of indirect cash ‡ow rights (i.e., cash ‡ow through
intermediate …rms)
Does this matter?
Olken ()
Firms/Contracts Lecture 2
10/08
19 / 30
Regressions and Results
Question 1: sensitivity to own shocks
π kt = a + b (predkt ) + c (cashk
predkt ) + dXkt + αk + αt + εkt
Courtesy of MIT Press. Used with permission.
Olken ()
Firms/Contracts Lecture 2
10/08
20 / 30
Regressions and Results
Question 2: sensitivity to amount of director equity
Courtesy of MIT Press. Used with permission.
Olken ()
Firms/Contracts Lecture 2
10/08
21 / 30
Regressions and Results
Question 5: Is there asymetry, i.e., do pro…ts move towards the ’top’
…rm in the chain?
Courtesy of MIT Press. Used with permission.
Olken ()
Firms/Contracts Lecture 2
10/08
22 / 30
Family …rms
Many …rms are by family members of the original founder.
A priori, this seems ine¢ cient: why would we think that managerial
talent is hereditary? Shouldn’t the market …nd a better manager?
Why might this be?
Olken ()
Firms/Contracts Lecture 2
10/08
23 / 30
Burkart, Panunzi, and Shleifer (2003)
Tunneling!
Olken ()
Firms/Contracts Lecture 2
10/08
24 / 30
Burkart, Panunzi, and Shleifer (2003)
Tunneling!
Assume no superior manager has resources to buy …rm outright
Olken ()
Firms/Contracts Lecture 2
10/08
24 / 30
Burkart, Panunzi, and Shleifer (2003)
Tunneling!
Assume no superior manager has resources to buy …rm outright
Then:
If shareholder protections are strong, then you can sell all your stock in
the company, and it is run with diversi…ed ownership.
If shareholder protections are intermediate, you sell some stock but
continue to be a large shareholder, and monitor the professional
manager to limit expropriation.
If shareholder protections are very weak, so even a manager can
expropriate a large shareholder, you retain control within the family.
Olken ()
Firms/Contracts Lecture 2
10/08
24 / 30
Perez-Gonzalez (2006)
What is the impact of inherited management on …rm performance? Is
it actually negative (as above model suggests)?
Olken ()
Firms/Contracts Lecture 2
10/08
25 / 30
Perez-Gonzalez (2006)
What is the impact of inherited management on …rm performance? Is
it actually negative (as above model suggests)?
Idea:
Look at …rms that were initially controlled by a family, and where there
was a CEO succession
Compare stock returns for those …rms that announce family member
will be new CEO with those that announce external new CEO
Similarly compare change in actual pro…ts before and after new CEO
takes over
Olken ()
Firms/Contracts Lecture 2
10/08
25 / 30
Stock-market event studies
Stock market event studies:
Basic idea: e¢ cient markets hypothesis implies that the full long run
value of new information on a …rm is incorporated in the stock price
immediately
So the change in a stock’s price right around the time of new
information tells you the value of that new information
Olken ()
Firms/Contracts Lecture 2
10/08
26 / 30
Stock-market event studies
Stock market event studies:
Basic idea: e¢ cient markets hypothesis implies that the full long run
value of new information on a …rm is incorporated in the stock price
immediately
So the change in a stock’s price right around the time of new
information tells you the value of that new information
Development examples:
Fisman (2001) studies e¤ect of Suharto’s health on connected …rms to
determine the value of political connections
Guidolin and La Ferrara (2007) studies impact of con‡ict shocks on
mineral …rms to detect illegal diamond trade
Olken ()
Firms/Contracts Lecture 2
10/08
26 / 30
Stock-market event studies
Estimation:
Estimate a market model to …nd "abnormal returns" for a …rm, i.e.
take the residuals from
rf = α + βrm + εf
De…ne a window around the event e.
Then estimate average abnormal returns during the event window e
and test the null that they are equal to 0.
What do we learn from these models? When might they be
reasonable? When might they not be reasonable?
Olken ()
Firms/Contracts Lecture 2
10/08
27 / 30
Results
Courtesy of the American Economic Association. Used with permission.
Olken ()
Firms/Contracts Lecture 2
10/08
28 / 30
Results
Also examines changes in accounting pro…ts
Courtesy of the American Economic
Association. Used with permission.
Olken ()
Firms/Contracts Lecture 2
10/08
29 / 30
Concluding thoughts
Firms are important engines of economic growth
Problems with contracting and credit lead to unusual corporate
stuctures, with some bene…ts but also some costs
But I think there’s much more about …rms that hasn’t been explored
much.
Some things I think are interesting:
Business clusters
Branding
Endogenous adoption of technology
Internal …rm capital markets
Political capture
Firm behavior
Olken ()
Firms/Contracts Lecture 2
10/08
30 / 30