|
||||||||||
|
Table: Empirical
studies on ownership structure and performance[1]
A B C D E F G H I J K L M N O P Q R S T U V W X Y Z |
||||||||||
|
Author(s) &Journal |
Sample
& Period[2]
|
Ownership |
Performance
variable(s) |
Other
variable(s): Controls & dependents[5] |
Statistical
methods |
Main
results |
Preferred
explanation |
|||
|
|
|
|
|
|
Abnormal returns at the announcement of stock splits increases with the level of insider ownership. |
An incentive alingment argument coubled with a signaling argument. |
||||
|
134 firms listed at NYSE. 1971 & 1974 & 1977 & 1980 & 1983. |
Combined stock ownership by present CEO and all former CEOs still on
the board in the ranges [0-1%], [1-5%], [5-20%], and [20-100%]. |
1) Tobin’s Q by market value of stock, preferred stock and debt to
market value of capital stock, inventories, and other assets. 2) Return on assets
by EBIT to replacement value of assets. |
1) Industry has been accounted for by subtracting average industry
differences in Tobin’s Q from each observation of Tobin’s Q. 2) Size by log
of replacement cost of assets. 3) R&D costs to size. 4) Advertising to
size. 5) Proportion of outside directors on the board in the ranges [0-40%],
[40-60%], and [60-100%]. 6) Tenure by CEO in the ranges [0-4], [5-9],
[10-14], and above 15 years. 7) Median tenure of inside or outside directors
on the board. 8) Dummy for family relation between any two board members. |
OLS regression pooling the data on all five periods. IV (instrumental
variable) regression using lagged values of stock ownership as instruments.
Use piecewise linear regression on CEO ownership, outside director proportion
of board, and CEO tenure. Use Hausman’s specification test. |
Performance increases significantly with CEO ownership in the [0-1%]
range and decreases significantly in the [1-5%] range. Significant controls:
R&D to size, and advertising to size. Otherwise the different regressions
present insignificant or contradicting evidence. The Hausman test rejects the
hypothesis that there is no simultaneity at the five percent level. |
A combination of entrenchment, incentive alignment, and insider-investment
argument. |
||||
|
Himmelberg, Hubbard, and
Palia [1999], Journal of Financial Economics |
Panel data from Compustat. Small and large |
1) % of common equity holdings by all top-level managers. This value,
say m, is transformed using log(m/(m-1)). 2) Average % of equity ownership per top-level managers. Use log of
this value. |
1) Tobin’s Q by market value of stock + estimated market value of preferred
stock + book value of total liabilities to book value of total assets. 2)
Return on assets. |
Owner determinants: 1) Standard deviation of stock return. 2) Firm
size by log of sales and squared size. 3) Capital expenditures to capital
stock. 4) R&D to capital. 5) Advertising to capital. 6) Free cash flow by
operating income to sales. 7) Capital to sales ratio and this ratio squared.
8) Time and industry dummies. Performance determinants: 1) Managerial ownership.
2) A vector of variables equal to the owner determinants mentioned above. |
OLS and IV regressions. Use panel data to obtain a fixed-effects
estimator to avoid inconsistent estimates because of endogeneity of
managerial ownership caused by unobserved common determinants of ownership
and performance as opposed to reverse causality. Test for roof-shaped
relation by using squared ownership and piecewise linear regression. Use
dummies to control for bias for missing values regarding R&D, Advertising
and return volatility. Use a Hausman test to test for endogeneity. |
Find evidence of endogeneity of managerial ownership caused by unobserved heterogeneity
as opposed to reverse causality. This is supported by Hausman specification
test. Performance regression: Find some evidence of a roof-shaped
relation. But after controlling for firm characteristics and firm fixed effects
they find no relation between managerial ownership and performance even for
sub-samples of large and small firms. Significant controls: Size, squared
size, return variation, income to sales, capital exp to capital. Ownership
regression: Do not test the effect of performance. Significant
‘controls’: Size, capital to sales ratio, returns variability (only for
sample of small firms). Almost similar results for both ownership measures. |
Natural selection argument combined with a mutual neutralization argument
that focus on incentive alignment and determinants of the scope for managerial
discretion. |
|||
|
Holderness, Kroszner,
and Sheehan [1999], Journal of Finance |
Full sample: 1419 listed US firms in 1935 and 4202 listed Limited sample: 120 largest |
1) % and $ ownership by the firms officers and directors both directly
and indirectly. 2) % and $ ownership by CEO. 3) Combined shareholding by officers and
directors in the ranges: [0-5%], [5-25%], and [25-100%]. 1935 data by the Security Exchange Commission. 1995 data by Compact
Disclosure and ExecuComp. |
Tobin’s Q by market value of stock, and book value of debt to book
value of assets. |
1) Size by total assets. 2) Leverage by debt to assets. 3) Industry
by one digit SIC codes. |
Descriptive statistics and OLS regressions. Test for roof-shaped
relation by using piecewise linear regression. |
Full sample descriptive statistics: Mean (median) ownership by managers and directors increased from 13% (7%) in 1935 to 21% (14%) in 1995. In $ millions it increased from $18 ($3) to $73 ($16). For any given firm size ownership is higher in 1995 than in 1935. However, using the value-weighted mean, ownership increases insignificantly from 4.2% in 1935 to 5.9% in 1995, and median ownership of the 10% largest firms has fallen from 2.1% to 1.5%. Full sample OLS regressions: Profitability is significantly increasing for management ownership in the [0-5%] range and significantly decreasing in the [5-25%] range in the 1935 sample. For the 1995 sample Tobin's Q is significantly increasing for management ownership in the [0-5%] range. Significant controls: Size. Limited sample: Mean (median) ownership by CEO is 1.25% (0.09%) in 1935 and 1.25% (0.06%) in 1995. In $ millions the mean (median) ownership increased by $23.6 ($1.5) in 1935 to $386.5 ($11.9) in 1995. |
Stulz’s [1988] combined takeover premium and entrenchment argument. |
|||
|
Holderness and Sheehan [1985], Journal of Financial Economic |
|
Presence of large-block equity holder or not. |
Abnormal returns |
|
Event study |
Significant and positive performance on announcement of outsider’s
acquisition of a large equity position, but only persistent if takeover or
other corporate restructure follows. |
The incentive argument. |
|||
|
Holderness and Sheehan [1988], Journal of Financial Economics |
101 majority held and 101 diffusely held large NYSE and AMEX listed
firms. 1979-84. |
95%> MH >50,1%, ownership by any single individual or entity
(other corporation, or fund). DH<20% ownership by any shareholder. |
1) Tobin’s Q by firm’s market value to replacement cost of plants and
inventories. 2) Return on equity. |
Each of the 101 majority held firms are paired with a diffusely held
firm with approximately same 1) Size by total assets. 2) Two-digit SIC code. |
Standard t-tests are applied. The sample is further classified in
order to control for identity of MH control: MH by individuals and OC by
entities. |
Finds no significant difference in performance between majority held
(MH) and diffusely held (DH) firms. |
The entrenchment argument. |
|||
|
183 listed 1948-60. |
MC all non-OC firms. OC if ³50% of vote
carrying shares are held by individual or if 20-50% of the votes are held by
an individual, or if at least 20% of the votes are held by largest 20 vote
holders subject to certain constraints. |
Pre-tax profit / Net worth. |
1) Growth of net assets. 2) Variance of profitability. 3) Skewness of
profitability. 4) Dividend distribution by dividends / dividents plus
retained earnings. |
Discriminant analysis and generalized Mahalanobis distance analysis.
The sample is classified in order to control for industry. |
No significant difference between MC and OC when industry bias is
accounted for. |
The incentive argument. |
||||
|
343 of the largest 1960-72. |
Same as Palmer [1973]. However, in the final test he applies this
classification: MC £10% single block
of common stock. OC ³10%. |
Return on stocks calculated as average stock returns observed from
1962 to 72 assuming that dividends are reinvested. |
1) The effectiveness of the market for corporate control by market
value to book value. 2) Monopoly power by barriers to entry. 3) Size of firm
by revenues. |
Standard t-tests are applied. The sample is classified in order to
control for ‘efficiency of market for corporate control’, monopoly and size. |
OC firms are only significantly more profitable than MC firms with
regard to MC firms who lack an efficient market for corporate control. |
The incentive alignment argument coupled with a ‘market for corporate
control’ argument. |
||||
|
Holthausen and Larcker [1996], Journal of Financial Economics |
|
|
|
|
|
Operating performance declines following a reversed leveraged buyout
and so does insider ownership. |
|
|||
|
|
|
|
|
|
Evidence of a roof-shaped relation. |
|
||||
|
103 of 200 largest French companies. |
1) Majority control ³ 50% Minority control- between 5 and 50%. Internal control < 5%. 2) Case by case classification of firms into family and non-family
control. |
1) Net cash flow / Book value of equity and reserves. 2) Return on equity. |
1) Size by log of added-value. 2) Industry by dummies. |
OLS regression. Checks for simultaneous effect of size and ownership
by including interaction terms. |
For none of the ownership definitions are there any significant differences
in performance. Significant controls: Size, industry, and family*size. |
A combined size and incentive argument. |
||||
|
|
|
|
|
|
Find that managers make abnormal returns when trading in their firm's
stock. |
The insider-investment argument. |
||||
|
|
|
|
|
|
Operating profitability declines after going public and so does insider ownership. |
|
||||
|
551 adoptions of antitakeover charter amendments among US firms. 1979-85. |
1) Insider ownership by officers and directors. 2) Institutional ownership.
|
30-days' cumulative abnormal return, CAR, by the firm over the interval
(AD-20, AD+10) where AD is the announcement date. |
None controls. Other dependents: The probability of adopting a 'fair price' amendment. This amendment
is the least entrenching of five types of amendments. |
Event-study. OLS and logit
regressions. |
CAR decreases significantly with the adoption of anti-takeover amendments.
The more entrenching the type of amendment the larger the decrease in CAR.
OLS regression finds no significant relation between ownership and CAR. Logit
regression finds that the probability of 'fair price' adoption increases
significantly with institutional ownership. |
The entrenchment argument. Note that managers can entrench themselves
using anti-takeover provisions instead of stock ownership. |
||||
|
94 1976-87. |
Insider ownership by officers and directors. |
Four-days' cumulative abnormal return, CAR, by the firm over the interval
(AD-1, AD+2) where AD is the announcement date. |
1) Net-of-market increase in stock price over previous year. 2) Dummy
for NYSE announcing the probating dual class shares in 1984. |
Event-study. OLS-regressions. |
CAR decreases significantly at the announcement of the introduction
of dual-class stocks with differential voting rights. CAR decreases with insider
ownership. Significant controls: Net of market increase. |
The entrenchment argument. |
||||
|
120 largest 1938 & 1974 & 1984. |
% ownership by highest paid executive, typically the CEO. |
|
|
|
CEO mean (median) ownership declined from 1.7% (0.3%) in 1938 to 1.5%
(0.05%) in 1974 to 1% (0.03%) in 1984. MBOs have increased from $1,2 billion
in 1979 to $77 billion in 1987. Franchising accounts for 12% of GDP in 1986,
page 245. |
|
||||
|
53 non-anticipated deaths of senior executives in publicly listed US
firms. 1971-1982. |
Dummy for executive being a corporate founder. |
Three-days' standardized cumulative abnormal return, CAR, by the firm
over the interval (AD0, AD+2) where AD is the announcement date. |
1) Executive position by the dead executive’s compensation / compensation
of the highest paid officer excluding the dead one. 2) Executive ability by a
composite measure of past performance. 3) Executive cost. 4) Executive
benefit. |
Event-study. OLS regression in the final model. |
No general effect of executive death. However, significant positive
returns if executive is the founder. Significant controls: Lower returns for
higher executive position. |
Use an entrenchment like argument to explain the result on founder
death. The result on position is explained by loss of important ability. |
||||
|
- Copyright 1997-2008, ViamInvest. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Legal notice. |
||||||||||
