Institutions and Organizational Structure: The Case of State-Owned Corporate Pyramids

Joseph P.H. Fan Organization and Value 1

Institutions and Organizational Structure:

The Case of State-Owned Corporate Pyramids

Joseph P.H. FanT.J. Wong

Tianyu Zhang

The Chinese University of Hong Kong

Emergence of Chinese State Owned Corporate Pyramids

Since the 1990s China witnessed a wave of enterprise restructuring in which state assets were spun off from parent state-owned enterprises (SOEs) and injected into their newly set up subsidiaries.

The organizational transformation is typically through several rounds of joint ventures established by SOEs and their subsidiaries with other government and SOE partners. A subsidiary may further spin off assets into another newly created subsidiary, and so on.

The system of state-owned companies forming China’s business groups is called “Multi-layer Legal Person System (多级法人制 )” (Wu, 2005).


Chinese Pyramids 4

SOE Under Old System

Central Govt

SOE

Chinese Pyramids 5

SOE Under New SystemCentral Govt

Local Govt / SAMB

Parent SOE / SAMC

Managers

Board of directorsListed SOE

Market competitions Product and jurisdictions

Outside Shareholders

Banks

CSRC,Judicial Courts

External auditors

Chinese Pyramids 6

Types of Control of SOEs in China

SAMB SAMB

Listed Firm Parent SOE/SAMC

Listed Firm

SAMB

Parent SOE

Intermediate Companies

Listed Firm


NINGBO VEKEN ELITE GROUP CO.,

LTD(600152)

NINGBO VEKEN

GROUP CO., LTD

NingBo Textile Holding

(Group) Co. LTD.

Ningbo Industry Investment

Co. LTD.

29.7%

17.5% 6.13%

93.19% 2.33%

Local Government

100% 100%

Pyramidal Organizational Structure in China

(Multi-layer Legal Person System)


Pyramidal Organizational Structure

X X

V=50%Y

V=20%Z

V=C=10%Z

V=C=50%Y

V = 20%, weakest link in the chain,C = 10%. Pyramid structure allows leveraging up in control


Why pyramiding? Highly popular around the world Why the owner builds pyramid is a puzzle Few theoretical and empirical works Morck, Wolfenzon, Young (2004); Khanna and

Yafeh (2005) Tax (Morck, 2005) Facilitating business group formation Relieving financial constraints while preserving

control (Almeida and Wolfenzon , 2006) Expropriation due to separating control from

ownership


Why Chinese build pyramids?

No prior research Ownership transfer restriction of SOEs precludes equity

financing and expropriation as primary motives Separation between ownership and control of SOEs is

small Cash flow over voting rights of common stocks

N Mean Median Std. Deviation Min Max

Government-controlled Firms

750 0.96 1.00 0.12 0.11 1.00

Privately-controlled Firms

62 0.54 0.52 0.23 0.14 1.00


Why pyramiding?Our new hypothesis

Credibly decentralizing decision rights within organization From a controlling owner (headquarter) to a local

(divisional) manager Intermediate layers associated with high information

cost, making the owner difficult to intervene the manager’s daily decision process

Compared with a horizontal control structure, a pyramidal structure allows local managers more autonomy in decision making


Theories of Firm Decentralization

Williamson (1985): decentralization provides manager high-power incentive

Jensen and Meckling (1992): the gain of decentralization arises from co-locating knowledge and decision rights

Credibility of decentralization depends on how costly is ex post intervention The IO literature emphasize information cost as a credible

commitment (Cremer, 1995; Aghion and Tirole, 1997; Baker, Gibbson, Murphy, 1999)

The extent of decentralization is constrained by the cost of losing control (agency problem)

Evidence is sparse: Baker (1992), Rajan and Wulf (2006)

Division of Power in SOEs

The government, lack of managerial expertise and information, will typically empower managers

Gain from the autonomy might be expropriated by managers – agency problem

However, empowering managers means lower government intervention – political cost

Optimal division of power should be the point where marginal agency cost equals marginal political cost



Ownership Transfer Restriction and Pyramiding of SOEs

The ownership restriction of SOEs prevents outright sales (such as M&As, MBOs) as a means to transfer decision rights that is common in markets of diffused ownership (Burkart, Gromb, Panuzi, 1997)

Pyramiding is a way of decentralizing business decision rights, or establishing real authority (Aghion and Tirole, 1997) as oppose to formal authority, short of actual transfer of ownership Compared with a policy order, decentralization through

pyramid is more credible to firm management Credible because of high information and bureaucratic

costs if the government intervenes ex post


This research focuses on local government controlled firms upon

their going publicYear Total

1 2 3 >=4

Before 1993 16 13 6 1 36 67.92%

1993 54 34 5 0 93 75.00%

1994 42 32 5 0 79 71.17%

1995 7 8 2 0 17 70.83%

1996 33 82 18 0 133 65.84%

1997 24 94 14 2 134 65.05%

1998 8 59 7 0 74 69.81%

1999 6 43 10 2 61 62.24%

2000 7 50 6 2 65 47.45%

2001 3 37 9 1 50 63.29%

Total 200 452 82 8 742 65.09%

Number of pyramidal Layers As a percenterage of total IPO firms


China – One Country, Many Systems

Protection for property rights

HighMiddleLow Missing


Predictions of the Extent of Pyramidal Layers Controlling a Publicly Listed SOE

A government owner’s (whose transfer rights is restricted) pyramiding (decentralization) decision is affected by his conflict with the firm in objective and the degrees of managerial agency problem Regional unemployment problem (-) Fiscal health (+) Government long-term incentive (R&D and education

expenditure) (+) The degree of conflicts of interest reduces when the

market and the laws provide strong disciplines Regional market development (+) Regional legal environment (+) Regional property rights protection (+) Government deregulation (+)


Empirical Design

Regression (probit) analysis to find the determinants of pyramids

Regression analysis to find whether firm performance/productivity is related to the pyramidal control structure


Determinants of pyramids (probit regressions)

Independent Variables (1) (2) (3) (4) (5) (6) (7) (8)Institutions 0.108

(2.89)***Marketization 0.094

(2.37)**Legal Environment 0.201

(3.76)***Property Rights 0.127

(3.05)***Deregulation 0.252

(2.53)**Incentive 0.095

(2.19)**Unemploy rat -0.123

(3.65)***Fiscal Surplus 0.539

(2.35)**R&D 0.109

(3.01)***Total Assets 0.187 0.142 0.177 0.173 0.183 0.184 0.199 0.142

(2.97)*** (2.50)** (3.04)*** (2.67)*** (3.17)*** (3.18)*** (3.31)*** (2.38)**Leverage 0.018 0.015 0.010 0.009 0.019 0.018 0.020 0.014

(0.52) (0.40) (0.28) (0.24) (0.55) (0.50) (0.55) (0.37)M/B 0.043 0.037 0.040 0.038 0.044 0.040 0.049 0.034

(1.34) (1.13) (1.26) (1.19) (1.33) (1.19) (1.53) (1.04)Regulated Industry 0.110 0.073 0.099 0.094 0.116 0.138 0.129 0.087

(0.71) (0.45) (0.58) (0.64) (0.68) (0.86) (0.78) (0.54)N 742 742 742 739 742 742 742 739Pseudo R-square 0.084 0.097 0.090 0.093 0.084 0.088 0.084 0.099


Independent Variable(1) (2) (3) (1) (2) (3)

Commercial Port 1.967 1.135(3.95)*** (1.81)*

Leased Territory 2.103 1.320(4.09)*** (2.01)*

Institutions 0.100 0.076 0.086(2.52)** (2.25)** (2.63)***

Incentive 0.146 0.142 0.143(3.47)*** (3.32)*** (3.39)***

Total Assets 0.160 0.164 0.160(2.93)*** (2.83)*** (2.86)***

Leverage 0.020 0.021 0.020(0.55) (0.58) (0.57)

M/B 0.038 0.038 0.037(1.14) (1.13) (1.11)

Regulated Industry 0.108 0.117 0.112(0.62) (0.69) (0.65)

Constant -1.595 -1.439 -1.678(4.79)*** (4.77)*** (5.27)***

Observations 29 29 29 742 742 742Adjusted (Pseudo) R-square 0.34 0.36 0.41 0.092 0.091 0.092

First Stage Second Stage

Two-stage model


Control Pyramids and Managerial Professionalism

More decentralized firms, more managerial professionalism

More decentralized firms, higher productivity

We do not claim causality


Pyramidal Structures and Management Professionalism

CEO and ChairmanProfessionalism Political Connection

(1) (2) (3) (4)

Pyramid 0.554 0.532 0.085 0.072

(2.84)*** (2.97)*** (0.49) (0.44)

Aggregate Institutions

-0.066 -0.080

(1.01) (0.95)

Aggregate Incentive 0.173 0.071

(1.13) (0.83)

N 625 622 625 622

Pseudo R square 0.097 0.100 0.062 0.063


Pyramidal Structures and Management Professionalism

Board of Directors Professionalism Political Connection

(1) (2) (3) (4)

Pyramid 0.077 0.080 0.014 0.016

(4.48)*** (4.33)*** (0.81) (0.96)

Aggregate Institutions

-0.007 -0.005

(1.33) (0.80)

Aggregate Incentive 0.002 0.004

(0.15) (0.68)

N 645 642 645 642

Adjusted R square 0.19 0.20 0.11 0.11


Pyramidal Structures and Employment Efficiency

Employee Employee

(1) (2)

Pyramid -1.409 -1.161

(2.88)*** (2.93)***

Aggregate Institutions -0.662

(4.48)***

Aggregate Incentive -0.034

(0.17)

N 617 614

Adjusted R-squared 0.19 0.20


Structures and Total Factor Productivity

Output Output

(1) (2)

Pyramids 0.088 0.075

(1.72)* (1.75)*

Aggregate Institutions 0.048

(1.40)

Aggregate Incentive 0.069

(2.13)**

Labour 0.197 0.202

(5.46)*** (6.47)***

Capital 0.513 0.516

(8.57)*** (9.34)***

Observations 1922 1913



Pyramidal Structures and Tobin’s Q

Tobin's Q Tobin's Q

(1) (2)

Pyramids 0.086 0.058

(1.76)* (1.09)

Aggregate Institutions 0.040

(1.80)*

Aggregate Incentive 0.025

(0.99)

Observations 562 559


Pyramidal Structures and Accounting Returns


ROS ROS ROA ROA

(1) (2) (3) (4)Layers 0.018 0.021 0.005 0.006

(1.73)* (2.27)** (1.47) (1.77)*Aggregate Institutions -0.006 -0.002

(2.44)** (0.67)Aggregate Incentives -0.008 -0.002

0.007 0.007 -0.030 -0.030Financial Leverage (0.85) (0.83) (11.97)*** (11.75)***

0.005 0.009 0.012 0.013Firm Size (0.68) (1.22) (6.98)*** (7.27)***

0.011 0.012 0.007 0.008Growth (4.37)*** (4.43)*** (6.49)*** (7.24)***

0.148 0.151 0.016 0.017Regulated (8.90)*** (9.01)*** (3.00)*** (3.16)***

(1.22) (1.98)**Constant -0.010 -0.090 -0.177 -0.200

(0.06) (0.57) (4.93)*** (5.13)***Observations 2224 2215 2224 2215Adjusted R-squared 0.32 0.33 0.43 0.44

Survey of Bureaucrats in State Asset Management Systemthe response from 15 government bureaucrats in state assets management system with regards to control of firms under different structure.

1 - if the government has more control of firm A’s particular decision than firm B’s.

0 - if the government has the same level of control over firm A and B’s decision.

-1 - if the government has more control of firm B’s decision than firm A’s.


Government

Firm A

Government

Parent SOE

Firm B

Survey among bureaucrats in state assets management system


Question N Mean Median Std.Dev. Min Max

Long-term strategic decisionsSet production plans 15 0.53*** 1*** 0.52 0 1Make investment decisions 15 0.93*** 1*** 0.26 0 1Appoint executives 11 0.64** 1** 0.81 -1 1Set dividend policy 15 0.67*** 1*** 0.72 -1 1Closure/sell of subsidiary 15 0.67*** 1*** 0.72 -1 1Pay equity compensation to executives 15 0.73*** 1*** 0.70 -1 1Borrow loans from banks 15 0.27 1 0.88 -1 1Raise additional capital from stock market 15 0.80*** 1*** 0.41 0 1M&A activities 15 0.67*** 1*** 0.62 -1 1

Short-term operating decisionsPurchase raw material 14 0.14 0 0.53 -1 1Purchase electricity and other energy input 14 0.29** 0** 0.47 0 1Set product prices 14 0.14 0 0.53 -1 1Choose customers 14 0.21* 0* 0.43 0 1Choose suppliers 14 0.29** 0* 0.47 0 1Set bonus level 14 0.29 0 0.61 -1 1Set salary level 14 0.14 0 0.77 -1 1Set bonus differentials 14 0.36* 0* 0.63 -1 1Recruit employees 14 0.29** 0** 0.47 0 1Dismiss employees 14 0.29** 0** 0.47 0 1

Survey of Senior managers

12 responses from SOEs and 19 from non-SOEs.

Does government set limit on a particular firm decision

1-No limit; 2-Some limit; 3-Many limits; 4-Completely limit


Survey among senior management


State-owned Enterprises

Non SOEs(N=19)

One layer (N=3)Two Layers

(N=4)More than two layers

(N=5)

Long-term strategic decisions Mean Median Mean Median Mean Median Mean MedianSet production plans 2.00 2.00 1.50 1.50 1.60 1.00 1.16 1.00Make investment decisions 3.00 3.00 2.50 3.00 2.20 2.00 1.32 1.00Appoint executives 3.67 4.00 2.75 3.00 2.80 3.00 1.05 1.00Set dividend policy 2.00 2.00 2.50 2.50 1.80 1.00 1.21 1.00Closure/sell of subsidiary 3.33 3.00 2.25 2.50 2.20 2.00 1.37 1.00Pay equity compensation to executives 3.00 3.00 2.75 3.00 2.60 3.00 1.05 1.00Borrow loans from banks 2.33 2.00 1.50 1.00 1.80 1.00 1.42 1.00Raise additional capital from stock market 3.00 4.00 2.50 2.50 2.20 2.00 2.21 2.00M&A activities 3.00 3.00 2.50 2.50 2.40 2.00 1.68 1.00Mean of long-term strategic decisions 2.81 2.89 2.31 2.39 2.18 1.78 1.39 1.33

Short-term operating decisionsPurchase raw material 1.00 1.00 1.25 1.00 1.20 1.00 1.00 1.00Purchase electricity and other energy input 1.50 1.50 1.25 1.00 1.20 1.00 1.32 1.00Set product prices 2.67 3.00 1.75 1.00 1.80 1.00 1.05 1.00Choose customers 2.00 2.00 1.00 1.00 1.20 1.00 1.00 1.00Choose suppliers 1.67 1.00 2.00 1.50 1.20 1.00 1.11 1.00Set bonus level 1.67 2.00 2.00 2.00 1.60 1.00 1.05 1.00Set salary differentials 2.00 2.00 2.25 2.50 1.80 2.00 1.58 1.00Set bonus differentials 1.67 1.00 1.75 1.50 1.40 1.00 1.06 1.00Recruit employees 1.67 2.00 1.25 1.00 1.40 1.00 1.11 1.00Dismiss employees 2.00 2.00 1.75 1.50 1.40 1.00 1.32 1.00Mean short-term decisions 1.85 2.25 1.63 1.45 1.42 1.10 1.16 1.10


Summary of Findings

The extent of a government-controlled pyramid is positively related to local government’s decentralization incentive

Positive relation between the pyramidal control structure and firm performance/efficiency

Positive relation between pyramidal control and managerial professionalism and productivity

Survey of bureaucrats and managers confirms our hypothesis


Conclusions and Future Research Implications

The pyramidal ownership structures of China’s state enterprises are fundamentally related to state ownership transfer restriction

It is the right to transfer, not cash flow or voting right, that is of first order importance in shaping owners’/managers’ incentives

We need more research on the effects of these (and other) property rights constraints on firm behaviors

How general are the results? Any state owned pyramids outside China? Can the decentralization explanation of pyramiding

behavior be generalized to private firms?

Institutions and Organizational Structure: The Case of State-Owned Corporate Pyramids

Documents