Journal

of Financial

Economics

29 (1991) 287-313.

North-Holland

The staying power of leveraged buyouts*

Steven N. Kaplan
University of Chicago, Chicago, IL 60637, USA
Nationa! Bureau of Economic Research, Cambridge, ,tU 02138, tiSA
Received

February

1991, final version

received

May 1991

This paper documents

the organizational
status over time of 183 large leveraged
buyouts
completed
between 1979 and 1986. By August 1990. 62% of the LBOs are privately owned, 11%
are independent
public companies,
and 24% are owned by other public companies.
The
percentage
of LBOs returning
to public ownership
increases over time, with LBOs remaining
private for a median time of 6.82 years. The majority of LBOs, therefore,
are neither short-lived
nor permanent.
The moderate
fraction of LB0 assets owned by other companies
implies that
asset sales play a role, but are not the primary motivating force in LB0 transactions.

1. Introduction

In the 198Os, an unprecedented number of public corporations and their

divisions went private in leveraged buyout transactions (LBOs). LB0 activity
increased from $1.4 billion in 1979 to $77 billion in 198K1 In spite of the
large transaction volume, LBOs remain poorly understood. In particular,
there is a healthy debate about the longevity of LB0 organizations and their
distinguishing characteristics. The answer to this debate has implications for
the reasons LBOs occur and for sources of value in LB0 transactions.
*Michael Beane collected much of the data for this study as part of his MBA honors thesis.
Sang Han provided additional
research assistance.
George Baker (the referee), Eugene Fama,
Robert
Gertner,
Michael Jensen (the editor).
Mark Mitchell,
Kevin IM. Murphy,
Mitchell
Petersen, Andrei Shleifer, Jeremy Stein, Robert Vishny, and seminar participants
at Stanford
University
and the University
of Chicago have provided
helpful comments.
The Center for
Research in Security Prices provided research support.
See Jensen

(1989). Figures

are in 1988 dollars.

See Kaplan (1989a, b) for a discussion

buyouts of public companies.

0304~405X/91/$03.50

D 1991-Elsevier

of and evidence

Science

Publishers

on sources

of value

B.V. (North-Holland)

in management

23s

S.iV. Kaplan, The staying power of LBOs

Jensen (19891 argues that LB0 organizations solve the free cash flow
problem faced by companies in low-growth industries by providing superior
incentives to managers and their monitors. A large portion of the wealth
increases associated with LB0 transactions are directly attributable to incentives LBOs provide to pay out free cash flow. These include (1) large
debt-service payments required by debt-to-total-capital
ratios exceeding 85%,
(2) large equity stakes held by managers, and, often, (3) the monitoring of an
LB0 sponsor - labeled an active investor by Jensen - who structures the
transaction, owns the majority of the companys equity, and controls the
board of directors. Although public companies (and their managers) are
capable of obtaining the same benefits, they rarely have the incentive to do
so. In low-growth businesses, therefore, Jensen argues the public corporation
is inferior as an organizational form to the LBO.
Under Jensens view that LB0 incentives are a critical source of value,
LBOs should remain private for an unspecified, but significant period. When
and if they subsequently go public, Jensens view suggests that debt levels will
remain high and equity will still be largely held by managers and active
investors.
Jensens view is not the only one that predicts long-lived LB0 organizations. Kaplan (19S9a) and Schipper and Smith (198s) present evidence that
the tax deductibility of interest on buyout debt is potentially an important
source of value in management buyouts. The value of this deductibility
depends on how long the high LB0 debt load is maintained. Tax benefits of
debt are more valuable if the debt is permanent than if it is paid off within
two or three years. The maintenance of high debt levels over a long period,
therefore, would also be consistent with an important role for tax benefits.
The tax theory, however, makes no predictions about equity ownership.
Rappaport (1990) disagrees with Jensen that the LB0 organization is
superior to the public corporation. He argues that the discipline of debt and
concentrated
ownership impose costs of inflexibility to competition and
change. In addition, the typical active investor invests funds provided by
outside investors who expect to be repaid in five to ten years. For these two
reasons, Rappaport argues, buyouts are inherently transitory organizations.
Rappaport might have added that as their equity stakes increase in value,
managers bear an increasing amount of undiversified risk. Over time, one
way managers can reduce or diversify this risk is to return the company to
public ownership.
Although he does not say so explicitly, Rappaports position is consistent
with a buyout as shock therapy. According to this view, buyout incentives
lead managers to focus the organization on cash flow generation, to forego
unprofitable investment opportunities, and to sell unproductive assets. Many
of these changes, however, are one-time events. Once they have been
implemented, the benefits of the buyout incentives are necessarily smaller. At

S.N. Kaplan, The staying power of LBOs

289

some point, the costs of inflexibility, illiquidity, and risk-bearing exceed the
continuing benefits of buyout incentives and the company returns to public
ownership.3
Under Rappaports view, LBOs, needing the flexibility offered by access to
public markets as well as liquidity for managers and outside investors, will
return to public ownership relatively quickly through both initia1 public
offerings (IPOs) and sales to public companies. If managerial risk-bearing
concerns increase with buyout size, larger buyouts should return to public
ownership more often.
The results in Bhagat, Shleifer, and Vishny (1990) suggest that sales to
related public companies may be particularly important. They find that 72%
of the assets of large hostile takeovers are owned by corporations with similar
assets - related or strategic buyers - within three years of the takeover. The
analogous percentage is 43% for the six hostile LBOs in their sample. Only
20% of the assets in hostile takeovers are owned by raiders or buyout-type
organizations within two to three years. They conclude that incentive intensive organizations [i.e., LBOs] are not very important in the long run for
hostile takeovers. A similarly large role for (and rapid implementation of)
sales to related buyers in LBOs would not imply that ownership incentives
are unimportant, but would suggest different sources of buyout gains from
those implied by the Jensen, tax, or Rappaport explanations. Gains from
sales to related buyers could come from joint operating efficiencies between
buyers and sellers, market power, or buyers willingness to overpay for
divisions.
In this paper, I present evidence concerning the different views of LB0
organizations. I examine the post-buyout organizational form of 183 large
LBOs completed between 1979 and 1986 from the time of their completion
through August 1990. As of August 1990, 62% of the LB0 companies are
privately owned, 14% are publicly owned and still independent, and 24%
have been purchased by publicly owned U.S. or foreign companies. The 62%
reflects the current organizational status of the LB0 company, ignoring the
current organizational form of assets sold (and purchased) after the buyout.
The percentage of LB0 assets privately owned changes only slightly, decreasing to 59%, when assets sold (but not assets purchased) are considered.
The 62% and 59% figures are path-independent
percentages in that they
measure the current organizational status of the LB0 assets, ignoring the
path taken to reach that status. LBOs private as of August 1990 include those
that return to private ownership after first having returned to public ownership. Almost 45% of the 170 LBOs whose status can be identified through
3LBOs might also be expected to be short-lived
if buyouts provide a mechanism
for managers
to exploit information
advantages
over public shareholders.
Such information
advantages
would
be less acute in divisional buyouts.

290

S.N. Kaplan, The staying power

of LEOs

the entire period return to public ownership at some point. The median
time - conditional on returning public - these companies remain private is
only 2.63 years. However, the unconditional estimate of the median time
private is 6.82 years.
The likelihood of returning to public ownership is largest and roughly
constant in the second to fifth years after the LBO, and then declines
somewhat, but is again roughly constant thereafter. This pattern can be
interpreted as consistent with two underlying types of buyouts - the shocktherapy type and a longer-term incentive type.
Overall, the evidence suggests that the typical buyout is neither short-lived
nor permanent. Consistent with Rappaport, the large fraction of LBOs
returning to public ownership suggests that the private LB0 is often a
transitory organizational form, bridging periods of public ownership. Alternatively, consistent with the Jensen model and the importance of the incentiveintensive organization, a substantial fraction of LB0 assets are private and
still highly leveraged many years after the LBOs. In addition, the LBOs that
are currently independent public companies appear to be hybrid organizations, retaining some of the characteristics of the LB0 organization - equity
ownership by insiders (managers and buyout investors) and debt ratios that
are higher than their pre-buyout levels. The maintenance of high debt levels
by private LBOs and by independent public companies is also consistent with
a role for tax benefits.
The results also suggest a moderate, but not primary, role for asset sales to
buyers in related industries. Almost 29% of the LB0 companies and 34% of
the original LB0 assets are owned by companies with other operating assets.
This is lower than the 72% reported by Bhagat et al. for only three years
after the hostile takeovers, and somewhat lower than the 43% reported three
years after the six hostile LBOs in their sample. Furthermore, the time
pattern of asset sales in my sample suggests that the typical buyout will
remain independent for more that 15 years. Unlike the Bhagat et al. findings
for hostile takeovers, my results for LBOs are consistent with long-term
incentives playing a role in explaining value increases.
Finally, I report cross-sectional differences in the current public/private
status of the sample LBOs: (1) larger transactions are only moderately more
likely to be publicly owned, suggesting a correspondingly moderate role for
risk-bearing considerations; (2) buyouts sponsored by well-known LB0 associations - the focus of Jensens article - are no more likely to stay private
than other buyouts; and (3) there is no significant difference in the current
public/private status of buyouts of divisions and public companies.
Most of the LBOs considered here are followed by a growing economy in
the first few post-buyout years. In contrast, many of the LBOs completed
after 1986 have to contend with a weakening economy and, possibly, less

S.N Kaplan,

The staying power of LBOs

291

favorable stock and bond markets. At the same time, Kaplan and Stein (1991)
present evidence that buyout prices and financial structures are more aggressive in the 1986-1988 period than earlier. The effect of the increase in deal
aggressiveness and the weakening economy in 1990 on the organizational
experience of LBOs is an open and interesting question.
The paper proceeds as follows. Section 2 describes the sample and the
data collection process. Section 3 presents a detailed analysis of the
private/public
organizational status of the LBOs over time. Section 4 describes other current characteristics of the LBOs. Section 5 examines the
cross-sectional determinants of organizational status, and section 6 concludes.
2. Sample and data
2.1. Sample
The sample includes those transactions identified as leveraged buyouts by
Securities Data Corporation (SD0 or Morgan Stanley & Company between
1979 and 1986. I exclude transactions completed after 1986 to ensure at least
3.67 post-buyout years for changes in organizational form to occur. To
increase the likelihood of identifying a companys current organizational
status, I include only those buyouts with a transaction value greater than
$100 million. This criterion potentially introduces a bias toward reversion to
public ownership, because the costs of risk-bearing to managers are likely to
be greater in larger buyouts.
These criteria generate a sample of 183 companies. The second column of
table 1 shows the number of transactions completed over time. These sample
buyouts were valued at $83.0 billion when they were completed. Over the
same period, W.T. Grimms Mergerstat Review reports $92.2 billion in
going-private and unit management buyout transactions. This sample, therefore, includes a large fraction of the dollar value of transactions completed
between 1979 and 1986.
2.2. Post-buyout information
I obtained post-buyout information on these companies from Lotus Datext
(public and private) databases, the NEXIS database, Wall Street Journal
articles from the year the LB0 was completed through August 1990, and,
when available, financial reports filed with the Securities and Exchange
Commission (SEC). Whenever possible, the corporate treasurer or controller
of each sample company was called to confirm the information. The postbuyout information includes, and the telephone interviews attempted to

292

S.N. Kaplan, The staying power of LBOs

Table 1
Current ownership status of parent company by year of completion for 183 leveraged buyouts
valued at more than $100 million and completed in the period 1979-1986.

Total
Year

LBOsa

1979-1980
1981
1982
1983
1984
1985
1986

1:
13
16
37
33
68

1979-1983
1984-1986

45
138

All deals

183

Status
not
knownb

LiquidatedC

Status
known

Percent
publicly
owned
August1990d

Percent
privately
owned
August 1990

11
13
14
31
30
66

40.0
45.5
46.2
64.3
29.0
40.0
31.8

60.0
54.5
53.8
35.7

43
127

51.2
33.1

48.8
66.9

170

37.6

62.4

71.0
60.0
68.2

Sample of leveraged buyouts identified as leveraged buyouts by either Securities Data

Corporation or Morgan Stanley.
bA buyouts status is not known if I could not contact the company or find any information
about it.
A buyout is considered liquidated if the company (1) has been completely sold in more than
one transaction to both public and private buyers, or (2) no longer exists.
dA buyout is considered a public entity if (I) it has issued equity to the public and is still a
public company as of August 30, 1990, or (2) the buyout company has been purchased by and is
still owned by any public company, either domestic or foreign.
A buyout is considered a private entity if the buyout company is still privately owned,either
by the buyout company or by a subsequent private buyer.

confirm, the following: (1) the date and dollar value of the original transaction, (2) whether assets had been sold since the LB0 and if so, their identity,
their dollar value, the acquirer, and the organizational form of the acquirer,
and (3) the current ownership status and organizational form of the company.
3. Post-buyout

status - Private or public assets?

3.1. Current organizational status - LB0 company

Table 1 presents the current status of the 183 buyout companies by year of
LBO. These companies are classified into one of four basic categories:
(1) unidentified, (2) liquidated, (3) still privately owned (including companies
in Chapter 111, or (4) publicly owned. Some post-buyout information is
available for 179 of the 183 companies. The remaining four could not be
identified; presumably, they have either changed their names, been sold, or

S.N Kaplan, Thr staying power of LBOs

293

gone bankrupt. At most, therefore, the analysis in the rest of the paper uses
these 179 companies. An additional nine companies have an ambiguous
organizational form because (1) they failed (and no longer exist) or (2) they
were sold in more than one transaction to both private and public buyers.
The remaining 170 companies have an August 1990 organizational form I
could identify. At a minimum, the analysis that follows uses these 170
companies.
As of August 1990, 62.4% (or 106) of the 170 LBOs of known status are
still privately owned. The remaining 37.6% (or 64) are publicly owned - 13.5%
(or 23) are independent public companies and 24.1% (or 41) are owned by
other public companies. The 62.4% result pertains to all of the LBOs in the
sample. LBOs completed at the end of 1986 have had as Iittle as 3.67 years to
change organizational forms since the buyout, compared with over 11 years
for LBOs completed in 1979. If LBOs are transitory organizations with
uncertain lives, one would expect fewer of the earlier LBOs to be privately
owned. The pattern in table 1 is roughly consistent with this. Only 48.8% of
LBOs completed by 1983 are still privately owned, compared with almost
67% of LBOs completed after 1983. But the pattern is by no means
monotonic: fewer than 36% of the LBOs completed in 1983 are still private,
compared with almost 55% of those completed in 1981.
3.2. Current organizational status - All assets
The previous results reflect the current organizational status of the LB0
company. They would be misleading if many of the buyout companies make
large divestitures and sell the divested assets to public companies. To address
this possibility, I calculate the fraction of a companys assets that are private
as of August 1990. For each sample LBO, I determine the current organizational form of all of the companys assets. If no assets have been sold, then all
of the companys assets are either public or private, depending on the
buyouts current organizational status. For LBOs that have sold assets, the
fraction of assets still private is calculated as the value-weighted average of
the public/private
status of both the retained and sold assets. In the few
cases in which all of the assets of the LB0 company have been sold, the sale
prices of the different assets are used as weights. In those cases in which
some assets are not sold (and, therefore, cannot be valued), accounting
numbers are applied as weights to all of the buyout-company assets. Operating incomes before interest, depreciation, and taxes of the different assets are
used as weights if they can be calculated. Book assets are used if operating
income results are not available, and, finally, sales, if book assets are not
available. In most cases, these accounting weights are based on the accounting results for the buyout company assets in the last pre-buyout fiscal year.

J.F.E.-

294

S.N. Kaplan, The staying power of LBOs

Table 2
Estimated current ownership status of average company assets by year of completion of 176
leveraged buyoutsa valued at more than $100 million and completed in the period 1979-1986.

Year

Number
LBOs

Percent assets
publicly owned
August 1990b

Percent assets
privately owned
August 1990

40.0

60.0
57.3
51.1
30.0
60.3
57.0
67.5

1979-1980
1981
1982
1983
1984
1985
1986

13
14
35
32
66

42.7
48.9
70.0
39.7
43.0
32.5

1979-1983
1984-1986

43
133

53.2
36.9

16.8
63.1

All deals

176

40.9

59.1

1;

aSample of leveraged buyouts identified as leveraged buyouts by either Securities Data

Corporation or Morgan Stanley. Seven of 183 LBOs are excluded because I could not contact
the company or find the required information about it.
bAssets of buyout company are considered public if they are owned by a publicly owned
company as of August 30, 1990. This column measures the average fraction of company assets
private for all LBOs completed in the given period.
Assets of a buyout company are considered private if they are still privately owned. either by
the buyout company or by a subsequent private buyer, as of August 30, 1990. This column
measures the average fraction of company assets private for all LBOs completed in the given
period.

Table 2 shows that the adjustment for asset sales decreases the fraction of
assets privately owned slightly - from 62.4% to 59.1%. The pattern over time
is similar to that for the current organizational form of the LBOs. The
number of LBOs that contribute to this table increases to 176 because the
assets of six of the nine companies with an ambiguous current organizational
form can be traced. It is worth adding that the 59.1% result is a lower bound
on the assets private because it does not adjust for post-buyout purchases of
public assets by the private companies.
3.3. OrganizationaE status by year after LB0
Instead of presenting the fraction of LBOs that are private by year of LB0
completion, table 3 presents the fraction of LBOs that are private by year
after the LBO. This is analogous to considering organizational status in event
Unfortunately, I did not collect information on asset purchases. Given the small effect asset
sales have on the percentage of privately owned assets, it seems likely that asset purchases would
increase the percentage of assets private by less than 5%.

295

S.N. Kaplan, The staying power of LBOs

Table 3
Ownership

Age of LB0
Year
Year
Year
Year
Year
Year
Year
Year
Year
Year

1
2
3
4
5
6
7
8
9
10

status of parent company

by age of leveraged
buyout for 179 leveraged
valued at more than SIOO million and completed in ther period 1979-1986.
Total LBOs status
known at year i
179
176
175
139
93
65
39
24
7
4

buyouts

Percentage
publicly ownedh

Percentage
privately ownedC

2.s
15.9
25.1
32.4
33.3
37.9
43.6
45.8
25.6
0.0

97.2
84.1
74.9
67.6
66.7
62.1
56.3
54.2
71.4
100.0

aSample
of leveraged
buyouts
identified
as leveraged
buyouts by either Securities
Data
Corporation
or Morgan Stanley. Four of 183 LBOs are excluded because I could not contact the
company or find any information
about it. Year i is the end of year i after the buyout.
bA buyout is considered
a public entity if (1) it has issued equity to the public and is still a
public company, or (2) the buyout company has been purchased
by and is still owned by any
public company, either domestic or foreign, i years after the buyout.
A buyout is considered
a private entity if the buyout company is still privately owned, either
by the buyout company or by a subsequent
private buyer i years after the buyout.

time. This table uses all 179 of the LBOs for which data are available. The
nine companies with an ambiguous current form are included for years in
which their forms are not ambiguous.
From year 1 to year 8 after the buyout, the fraction of LBOs that are
privately owned decreases monotonically from 97.2% to 54.2%. The fraction
increases in years 9 and 10, but the increases are based on a small number of
observations.
Tables l-3 show that LBOs return to public ownership at widely varying
times. Just over 25% are publicly owned three years after the buyout, rising
to almost 46% by eight years after the buyout. This pattern suggests that the
typical buyout is neither short-lived nor permanent.
3.4. Estimates of time spent pricate
The percentages given above are path-independent
percentages in that
they measure the current organizational status of the LB0 assets regardless
of how they reached that status. They do not distinguish between (1) LBOs
still private from the original LB0 and (2) LBOs that have gone private a
second time after returning to public ownership. Table 4 presents the time
pattern relative to the year of buyout completion by which LBOs return to

XV. Kaplan, The staying power of LBOs

296

Table 4
Percentage
leveraged

LBOS
private at
beginning of
Year i

Year
after
LB0
Year
Year
Year
Year
Year
Year
Year
Year
Year
Year
Year

of LBOs that
buyouts valued

1
2
3
4
5
6
7
8
9
10
11

170
164
142
124
80
48
32
20
12
5
4

return
to public ownership
by age of leveraged
buyout for 170
at more than $100 million and completed in the period 1979-1986.
LBOs
returning
to public
ownershipb

LBOs
censoredC

6
22
18
17
7
3
1
1
0
0
1

0
0
0
27
25
13
11
7
7
1
3

Cumulative
survival
rate (LBOs
privateJd

Cumulative
failure rate
(LBOs
publicjd

96.5
83.5
72.9
62.2
56.0
52.1
49.6
47.2
41.2
47.2
35.0

3.5
16.5
27.1
37.8
44.0
47.9
50.4
52.8
52.8
52.8
65.0

LBOs private at beginning of year i include those LBOs that (1) have not yet returned
to
some form of public ownership,
and (2) were completed
more than i - 1 years earlier. Includes
only the 170 LBOs whose organizational
status is known for all post-buyout
years.
A buyout is considered
a public entity if (1) it has issued equity to the public and is still a
public company, or (2) the buyout company has been purchased
by and is still owned by any
public company, either domestic or foreign, i years after the buyout.
LBOs censored are LBOs that (1) were completed between i - 1 and i years earlier. and (2)
are still private as of August 1990.
dThe cumulative
survival rate, S(f ), or product limit estimate equals:

S(r,) =

fr

&=I

(1 -d&/n,),

where d, is the number of LBOs that return public at I, and hk is the number of LBOs that
have (1) not yet returned public just prior to fj and (2) were completed at least t, years before.
The cumulative
failure rate, 1 - S(tj), is the estimated fraction of LBOs that have returned
to
public ownership
t, years after completing
an LBO.

public ownership. Almost 45% - 76 of 170 - of the LBOs whose status can
be identified through the entire period return to public ownership at some
point. Forty-two of these companies go public by issuing equity to public
shareholders. The remaining 34 are purchased by publicly owned companies,
both domestic and foreign. The median time - conditional on returning
public - these companies stayed private is only 2.63 years, and the conditional average is 2.85 years. This is similar to the median 2.42 and average
2.89 years spent under private ownership by the 76 firms examined by
Muscarella and Vetsuypens (1990).
The differences between the 45% result in this section and the smaller
percentage in the previous ones reflect the fact that 12 of the 76 companies
that return to public ownership subsequently go private again. Nine of these

S.;V. Kaplun, The staying power of LBOs

297

12 are independent public companies that complete a second LBO, suggesting that public ownership was not optimal. The remaining three are owned by
other public companies that subsequently go private. The 12 companies
remain public a median of 2.46 years (and an average of 2.34 years) before
going private again.
If all the LBOs in the sample had returned to public ownership at a known
time, it would be possible to calculate the unconditional median (and
average) time private directly. As table 4 shows, however, the majority of
LBOs in this sample are private as of August 1990 and must be considered
censored observations. LBOs are censored in year i after the LB0 if they are
still private as of August 1990, but were completed between i - 1 and i years
before August 1990.
It is possible to use the information about returns to public ownership and
censored observations to estimate both the median time and the overall
distribution of time it takes an LB0 to return to public ownership. The
product limit or Kaplan-Meier estimate of the survivor function is

s(l,)=kfil(l-d!c/,,,),

(1)

where d, is the number of LBOs that return public at tk and nk is the

number of LBOs that have 1) not yet returned public just prior to fk and
2) were completed at least t, years before. These estimates can be considered maximum-likelihood estimates [see ch. 1 of Kalbfleisch and Prentice
(1980)]. Here, 1 - .5(t) is the cumulative distribution function of the probability of returning to public ownership.
Table 4 shows that the median survival time or time to public ownership is
between six and seven years after the LBO. The estimated median time is
6.82 years. If the last observation were not private and the oldest LB0
returned public in September 1990, the estimated mean would equal 6.80
years. Because the last observation is censored, this estimated mean is clearly
downward biased. The estimated standard error using the 6.80 year mean is
0.36.j
The 6.82 year median time private is longer than the period supposedly
targeted by LB0
associations
for cashing out their initial investments.
Jensen (19891 claims that the goal is three
to five years. Returning
to public ownership,
however, is not the only way for investors to cash
out. As Jensen and Rappaport
note, LBOs can be releveraged
in a second LBO. In addition to
collecting data on buyout public/private
status, I attempted
to keep track of releveragings.
By
including releveragings
as well as returns to public ownership.
I can estimate the median time for
an investor to cash out. Almost 60% of the buyouts have cashed out in some way. The product
limit estimate of this median is 4.02 years. almost exactly midway through the goal of three to
five years.

298

S.N. Kaplan,

The staying power of LBOs

3.5. Duration dependence

The data used to construct table 4 can also be used to estimate the relation
between the likelihood of returning public and time. This relation is known
as duration dependence. Duration dependence is positive if the probability
that an LB0 returns to public ownership during the ith period, conditional
on being private at the beginning of the ith period, increases over time. and
negative if the probability decreases. The conditional likelihood is commonly
referred to as the hazard.
An examination of duration dependence provides important information
on the process by which LBOs return to public ownership. A finding of
negative duration dependence would imply that LBOs remaining private for
some time become increasingly likely to remain private. This could result if
initial uncertainty about the benefits of private ownership are resolved over
time. Companies that find the benefits of private ownership large become
increasingly Less likely to return to public ownership. Negative duration
dependence would be consistent with permanence for some LBOs.
Alternatively, negative duration dependence could reflect the existence of
unobserved heterogeneity. For example, suppose there are two vpes of
LBOs - shock-therapy LBOs and longer-term LBOs - both with constant
hazards. The shock-therapy types might be those for which the LB0 serves as
a change agent to shake up an inefficient organization, which is then better
organized as a public company because of flexibility or risk-bearing considerations. The longer-term types might be buyouts that more closely follow
Jensens characterization. Initially, a random LBOs likelihood of returning
public (conditional on being private) will be a weighted average of the
number of the two types in the sample. Over time, however. the short-term
LBOs will return to public ownership and only longer-term LBOs will
remain. As a result, estimated duration dependence will be negative. The
longer-term LBOs, however, will continue to return to public ownership at
their lower, but constant rate.
Negative duration dependence, therefore, would be consistent with permanence for some LBOs, but could also indicate the presence of unobserved
heterogeneity. In contrast, positive or no duration dependence would imply
that LBOs return to public ownership at some constant or increasing rate.6
Estimating the hazard function over time can also provide information on
the timing of possible flexibility, liquidity, and risk-bearing pressures. Although the shock-therapy view suggests that these pressures will lead LBOs
to return to public ownership after making one-time changes, it makes no
predictions about how long the changes take and how quickly the pressures
build.
6See Kiefer (1988,

pp. 671-672)

for a clear discussion

of these

issues.

299

S.:V. Kaplan, The sta.vingpower of LBOs

Following Lancaster (1990>, I estimate a piecewise-continuous

hazard
model.This specification lets the data tell us how the hazard behaves as a
function of time. This involves maximizing the log-likelihood function:

where di = 1 if the buyout returns to public ownership r; years after the

buyout, di = 0 if the buyout is still private in August 1990, ti years after the
buyout, and

S(ti;B) =exp i

P(G)=&+I

if

k bjPj-(ti_cy)Py+1

j=O

bj=cj-cj_,,

cy <tiIcL.+,,

where f< ) is the distribution function and S( 1 the survivor function. This
specification assumes that the hazard, pi, is constant during each interval,
ci_, to ci. These intervals are generally taken to be annual. The hazard
represents the likelihood that a buyout will return to public ownership in the
ith year after the LBO, given that the LB0 was still private at the beginning
of the ith year.
In addition to potential duration effects, year effects can exist. For example, LBOs completed in 1983 may have been different from LBOs completed
in 1986. Accordingly, some specifications include a set of dummy variables
based on the year the LB0 was completed (except for 1986). These dummy
variables should be compared with LBOs completed in 1986.5
Table 5 presents the model estimates. Regression 1 indicates that the
estimated probability of returning public (conditional on being private) is
lowest in the first post-buyout year, at 3.6%. The estimated probability peaks
in the fourth year, at 15.4%, declines in the fifth through seventh years, and
rebounds thereafter. Given the standard errors, the pattern is roughly consistent with constant duration dependence in years 2 through 5, and then a
lower, but constant, duration dependence in years 6 and after.
The previous version of this paper used logit estimates
rather than the duration-model
hazard estimates.
I have made this change for two reasons.
First, the duration
model can
incorporate
partial-year
censoring while the logit model cannot. This is particularly
important
for
estimating
the hazards for more than five years after the buyout, where there are fewer
observations.
Second, the coefficients are easier to interpret.
In the estimation.
this amounts to redefining
each p, to equal p,exp(4x),
vector of completion-);ear
diimmy variables and 4 is a vector of five coefficients

where x is the
to be estimated.

XIV. Kaplan,

300

7%e staying power of LBOr

Table 5
Maximum-likelihood estimates2 of the (conditional) probability of returning to public ownership
as a function of the time since the LB0 and of the LB0 completion year for 170 leveraged
buyouts valued at more than $100 million and completed in the period 1939-1986.
Probability/hazard
(1)

1st yearb
2nd year
3rd year
4th year
5th year
6th year
7th year
8th year
9th year and later
2nd to 5th year
6th year and later
2nd year and later
1981 and earlier
1982
1983
1984
1985

estimates of returning to public ownership

(2)

Coeff.

SE.

0.036e
0.144d
0.134d
0.154d
0.112d
0.072f
0.040
0.062
0.063

0.015
0.031
0.032
0.038
0.042
0.041
0.041
0.061
0.067

Coeff.
0.032d
0.130d
0.123d
0.143d
0.102e
0.073
0.040
0.065
0.080

- 0.262
0.018
0.578
- 0.196
0.493

No. of obs.
Log-likelihood

(3)
SE.

Coeff.

0.013
0.040
0.041
0.042
0.047
0.048
0.044
0.070
0.095

0.505
0.494
0.409
0.388
0.336

170
- 235.37

170
- 238.65

(4)
S.E.

0.032d

0.013

0.130d
0.066f

0.031
0.034

- 0.288
- 0.017
0.544
- 0.227
0.464

0.474
0.471
0.399
0.374
0.326

170
- 235.85

Coeff.

SE.

0.034d

0.014

0.129
- 0.529
-0.164
0.437
-0.315
0.455

0.031
0.432
0.442
0.377
0.366
0.326

170
- 237.12

The log-likelihood function is

L(P)=

Cd,lnf,(l,,p)+

C(l--d,)lnS(t,,/-J),

where d, = 1 if buyout returns to public ownership and 0 if censored, with distribution function
f,(f,>P) =P,+J(t,.P)>
and survivor function
S(ti,P)=exP

Cy<fiZZCCy+,,

-j~~j-(t-cy)Oy+l}3

Nt,)

= P,+1

if

cY < ri <cy_,.

bThe ith year coefficients estimate the hazard or likelihood of returning to public ownership in
the ith year after the buyout conditional on being private at the beginning of the year. The nth
year and later coefficients estimate the annual hazard in the n year or later after the LBO. The
second to fifth year coefficient estimates the annual hazard in the second to fifth year after the
LBO.
The variables 1982 to 1985 equal one if the LB0 is completed in that year and zero
otherwise. The 1981 and earlier variable equals one for LBOs completed in 1981 and earlier.
dSignificant at the 1% level.
Significant at the 5% level.
Significant at the 10% level.

S.,V. Kaplan, The staying power of LBOs

301

Regression 2 uses the same variables as regression 1, but controls for the
year the LB0 is completed. The basic pattern of the hazards is similar to that
of regression 1, but the coefficients in the second to fifth post-buyout years
decrease, while the coefficients for the sixth to ninth post-buyout years
increase. The likely explanation for the compression of these coefficients is
based on the negative coefficient (-0.262) on the 1981 and earlier variable.
The coefficient suggests that 1981 and earlier deals are 23% less likely in any
post-buyout year to return to public ownership than are 1986 buyouts.
Because a large fraction of the sixth to ninth post-buyout-year observations
are for LBOs completed in 1981 and earlier, not controlling for the LB0
completion year lowers the coefficients on the sixth to ninth post-buyout-year
variables relative to the earlier year variables. Consistent with table 1,
regression 2 also indicates that LBOs completed in 1983 are somewhat more
likely to return to public ownership in a given year than are LBOs completed
in other years.
The coefficients in regressions 1 and 2 suggest that there are two different
periods of constant duration dependence. To test explicitly for this, regression 3 replaces the eight post-buyout variables for the second year through
the ninth year and later with two variables: one for the second through fifth
year and one for the sixth year and later. The likelihood-ratio test statistic of
0.56 is not significant at conventional levels (chi-square distribution with six
degrees of freedom, p-value = 0.99). This breakdown suggests three different
likelihoods of returning to public ownership. The likelihood is small - at
3.2% - in the first post-buyout year, high - at 13.0% - in years 2 through 5,
and intermediate - at 6.6% - but constant, thereafter. As discussed above,
this pattern is potentially consistent with two underlying types of buyouts - a
shock-therapy type and a long-term type - if the shock-therapy LBOs have
largely returned to public ownership by the end of the fifth post-buyout year.
It is important to add that the hazards for the longer-term types do not
exhibit negative duration dependence. Although the probability that the
longer-term types return to public ownership is smaller, it does not decrease
over time, suggesting that these longer-term types will eventually go public
again.
The standard errors on the second through fifth year and sixth year and
later coefficients, however, are sufficiently large that the coefficients may not
be statistically different. Regression 4 tests this by replacing the two postbuyout-year variables with one variable that equals one if the observation is
for the second post-buyout year and after. The likelihood-ratio test statistic
of 2.54 is not significant at conventional levels (chi-square distribution with
one degree of freedom, p-value = 0.12). If one were to argue on purely
statistical grounds, therefore, the data fail to reject the hypothesis that after
the first year, the LB0 companies are equally likely to return to public
ownership. Economically, however, the difference between 13% and 6.6%

302

S.N. Kaplan,

The slaying power of LBOs

seems meaningful. The resolution of this discrepancy must await more

observations.
Overall, the results in this section suggest that LBOs are neither short-lived
nor permanent organizational forms. Neither the estimates in regression 3
with two LB0 types nor those in regression 4 with the constant hazard rate
find evidence for negative duration dependence.

4. Current LB0 characteristics

The analysis to this point distinguishes between privately and publicly
owned assets. It does not describe whether LB0 companies and assets are
still independently owned or have been purchased by other companies. In
addition, the private/public
dichotomy does not necessarily imply large
differences in post-buyout leverage or equity ownership. This section presents
evidence on post-buyout independence, leverage, and equity ownership.
4.1. Independent

or strategic assets?

As mentioned earlier, whether buyout companies remain independent or

are purchased by other companies has a strong bearing on the source of
value in LBOs. Although it would not imply that ownership incentives are
unimportant, a large role for and rapid implementation of asset sales in
LBOs would suggest different sources of buyout gains from those implied by
the Jensen, tax, or Rappaport explanations. Gains from asset sales to related
buyers could come from joint operating efficiencies between buyers and
sellers, market power, or buyers willingness to overpay for divisions, and not
from long-term changes in incentives.
Table 6 classifies the companies in my sample by their current private or
public status, and, within that status, by whether they are independent or
owned by another company. Of the 170 LBOs I can classify, 49 or 28.8% are
owned by companies with other operating assets. Because some of the
purchasers might be in unrelated businesses, 28.8% is an upper bound on the
percentage of LB0 companies that can be owned by related buyers. Most of
the companies that purchase LBOs - 41 of 49 - are public.
Although not presented in a table, a duration dependence analysis similar
to that reported for public/private status finds negative duration dependence
more than four years after the buyout. The estimates imply that the median
company will stay independently-owned
for more than 15 years.

9LB0 companies that are releveraged by a new LB0 investor group are considered to be
independently owned.

303

S..V, Kaplan, The slaying power of LBOs

Table 6
Number

Year

and ownership

Total
LBOs
status
known

status as of August 1990 of 170 leveraged buyouts

$100 million and completed in the period 1979-1986.
Publicly owned
Independent

August

1990b Privately

Owned by
other
ocompany

Total

Independent

owned August

valued

at more than

1990 often&

Owned by
other
company

Total

owned by
other
company

1979-1980
1981
1982
1983
1984
1985
1986

5
11
13
14
31
30
66

2
1
2
3
3
3
9

0
4
4
6
6
9
12

2
;:
9
9
12
21

2
6
6
4
20
I6
44

1
0
1
1
2
2
1

3
6
7
5
22
18
45

20.0
36.4
38.5
50.0
25.8
36.7
19.7

1979-1983
1984-1986

43
127

8
15

14
27

22
42

18
80

3
5

21
8.5

39.5
25.2

All deals

170

23

41

64

9s

106

28.8

%ample
of leveraged
buyouts identified
as leveraged
buyouts by either Securities
Data
Corporation
or Morgan Stanley. Includes only the 170 LBOs whose organizational
status is
known for all post-buyout
years. A buyouts status is not known if I could not contact the
company or find any information
about it.
bA buyout is considered
a public entity if (1) it has issued equity to the public and is still a
public company as of August 30, 1990, or (2) the buyout company has been purchased
by and is
still owned by any public company, either domestic or foreign.
A buyout is considered
a private entity if the buyout company is still privately owned, either
by the buyout company or by a subsequent
private buyer.

I also consider (but do not report in a table) the independent or nonindependent status of all original LB0 assets, including those that are divested. I
find that 33.6% are owned by other companies. Because I do not measure
post-LB0 asset purchases by the LB0 companies, the 33.6% overstates the
extent of sell-off activity.
These results imply a moderate role for asset sales to strategic buyers in
LBOs. Only 28.8% of the LB0 companies and 33.6% of the original LB0
assets are owned by companies with other operating assets at least 3.67 years
after the buyout. These are lower than the 72% reported by Bhagat et al. for
three years after hostile takeovers, and the 43% reported for the six LBOs in
their sample. Furthermore, the time pattern of independence suggests the
average life of independence is long.
It is possible that asset sales are more important for LBOs motivated by
hostile pressure. Accordingly, I divide the sample into hostile and friendly
LBOs. LBOs are considered hostile if: (1) the LB0 receives or is a response
to a hostile takeover bid; (2) the LB0 announcement follows the purchase of
at least 5% of the LB0 company equity by a hostile party in the prior six
months; or (3) the LB0 is a division of a company that satisfies the hostile

304

S.N. Kaplan, The rtayingpowr

of LBOs

definitions in (1) and (2). For LBOs classified as hostile, 26.2% of the
companies and 35.1% of the assets are owned by companies with other
operating assets; for LBOs classified as friendly, the percentages are 28.6%
and 32.8%. Neither of the differences between hostile and friendly LBOs is
significant. The extent of asset sales to strategic buyers. therefore, is not
related to the actual presence of hostile pressure. (It is possible, but not
testable, that this result would change if it were possible to measure unobserved, but perceived, hostile pressure.)
Overall, these results are consistent with the view that long-term incentives
are important in explaining the gains in LBOs.

4.2. Is high leverage maintained?

One of the distinguishing characteristics of an LB0 is high leverage.
Kaplan (1989b) reports a median debt to total capital ratio of 87.8% at
buyout completion for management buyouts announced between 1979 and
1985. This contrasts with a debt to total capital ratio of only 18.8% before the
buyout. LBOs that remain private need not retain their high leverage.
Similarly, LBOs that return to public ownership do not necessarily eliminate
their debt. This section considers the post-buyout capital structure, when
available, of the sample LBOs at the end of fiscal year 1989 (the last fiscal
year-end with such data).
I measure post-buyout capital structure or leverage in three ways. The first
measure is the book value of total debt (short-term and long-term) as a
fraction of the book value of total capital where total capital is the sum of
total debt, preferred stock, and common equity. The second measure is the
book value of total debt as a fraction of the LB0 transaction value reported
by SDC or Morgan Stanley. The third measure is the ratio of interest expense
to operating income before interest, depreciation, and taxes in fiscal year
1989. To serve as a benchmark, I compare the first and the third measures
for the LBOs with those for all nonfinancial public companies on COMPUSTAT with total assets of at least $100 million.
Capital structure data for the 1989 fiscal year are publicly available for 22
of the 23 LBOs that are independent public companies in August 1990. Such
data are available for only 33 of the 98 privately owned LBOs. Although the
relatively small fraction of data for privately owned LBOs leaves open the
possibility of ex post selection bias, the direction of that bias is not clear.
Panel A of table 7 shows that the 33 privately owned and independent
LBOs maintain high debt levels after the buyout. These companies have a
median percentage of total debt to total capital of 97.8% at the end of fiscal
yeas 1989. Total debt for these companies is 91% of the LB0 transaction
value. These values are similar to those found by Kaplan (1989b) at the time

S.&Y Kaplan. The staying power of LBOs

305

Table 7
Median debt to total capital, debt to initial transaction
value, interest to operating
income, and
equity ownership
by insiders at the end of the 1989 fiscal year by current organizational
status
for 90 leveraged
buyoutsa
valued at more than $100 million and completed
in the period
1979-1986 and for comparison
sample of public companies.

A. All independent
LBOs private
LBOs public

Total debt
to initial
deal valueC

Interest
expense to
operating
incorned

Inside
equity
ownership
fractione

33
22

0.978
0.718

0.910
0.607

0.719
0.276

N.A.
0.390

4
31

0.923
0.481

N.A.
N.A.

0.704
0.248

N.A.
0.066

0.435

N.A.

0.235

0.050

LBOsf

B. All purchased
LBOs
Private LB0 purchasers
Public LB0 purchasers
C. Comparison
companiesg

Number
of LBOs

Total debt to
total capital
(book valuejb

public

Sample
of leveraged
buyouts identified
as leveraged
buyouts by either Securities
Data
Corporation
(SDC) or Morgan Stanley.
bTotal debt is the book value of total debt for fiscal year 1989. Total capital (book value)
equals the book value of total debt, preferred
stock, and common equity for fiscal year 1989.
Initial deal value is the LB0 transaction
value given by SDC or Morgan Stanley.
Operating
income is before interest, depreciation,
and taxes.
eInsider equity ownership is the fraction of common stock owned by managers, directors, and
buyout sponsors of the independent
public LBOs and by managers and directors of the public
LB0 purchasers.
Shares of different classes of common stock are treated equally regardless of
voting rights.
LBOs are independent
if they have not been purchased
by another company with operating
assets. Assets of a buyout company are considered
public if they are owned by a publicly owned
company as of August 30, 1990, and private if they are owned by a privately owned company.
Debt to total capital and interest expense to operating
income are calculated
for fiscal year
1989 for all nonfinancial
public companies
on COMPUSTAT
with total assets exceeding $100
million. Insider equity ownership
is calculated
by 1McConnell and Servaes (1990) for 1,093 firms
tracked by the Value Line Investment
Survey in 1986.

LBOs are completed. The debt to total capital ratio is much greater than the
median ratio of 0.435 in fiscal year 1989 for the sample of nonfinancial public
companies. The median ratio of interest expense to ,operating income of
0.719 is also high relative to the median ratio of 0.235 for the nonfinancial
public companies. The 0.719 ratio, however, is somewhat less than the 0.833
ratio of projected post-buyout interest expense to pre-buyout operating
income found by Kaplan and Stein (1991) for 124 larger management buyouts
in the 1980s.
The 22 publicly owned and independent LBOs maintain substantial debt,
but less than the privately owned LBOs. The median percentage of tota debt
to total (book) capital is 71.8%, and total debt is 60.7% of the LB0
transaction value. The debt levels for the public LBOs are higher than the

306

S.N. Kaplan, The staying power

of LBOs

pre-buyout levels of 18.8% reported by Kaplan (1989b). Interest expense to

operating income, however, falls more sharply than for the private LBOs.
The ratio of 0.276 slightly exceeds the median 0.235 for the nonfinancial
public companies, but is much lower than the 0.833 found by Kaplan and
Stein for management buyouts in the 1980s at the time of the buyout.
The seeming discrepancy between the coverage and debt-ratio results is
probably caused by two factors. First, book values of equity do not necessarily
reflect market values. This is particularly true following a large stock market
increase like that experienced in much of the 1980s. Second, interest expense
to operating income will decline as short-term interest rates fall - as they did
from the early to the later 1980s - and as operating income increases. Such
operating-income increases will not immediately translate into large increases
in book equity.
As the previous section indicates, almost 29% of the LB0 companies have
been purchased by other companies. Leverage data for the 1989 fiscal year
are publicly available for 35 of these purchasers. Panel B of table 7 shows
that the median ratios of debt to total capital and interest expense to
operating income for the four private LB0 purchasers are similar to those of
the private and independent LBOs. The ratios for the 31 public LB0
purchasers are simiIar to those for the nonfinancia1 public companies presented in panel C.
The results in this section suggest that privately owned LBOs, both
independent and purchased, maintain debt levels similar to the levels when
the LB0 was compIeted. In contrast, publicly owned LBOs, both independent and purchased, maintain debt levels lower than the initial LB0 levels,
but somewhat higher than both pre-buyout levels and median pubIic-company levels.
4.3. Is high equity ownership maintained?
Another distinguishing characteristic of LBOs is concentrated equity ownership. Management and the LB0 promoter typically own or control 100% of
the post-buyout equity. According to the Jensen view, the concentrated
equity ownership provides strong incentives for managers and the LB0
promoter to maximize shareholder value. By definition, LBOs that remain
privately owned retain their concentrated equity ownership structure. It is an
empirical question whether LBOs that return to public ownership, either as
independent companies or as purchases of other companies, are stili characterized by concentrated equity ownership.
Stock ownership data are available for 18 of the 22 independent public
LBOs for fiscal year 1989. As panel A of table 7 indicates, insiders (buyout
sponsors and managers) hold a median of 39.0% (and an average of 41.6%)
of post-IPO equity in these 18 companies. McConnell and Servaes (1990)

S.IV Kaplan, The sruying power of LBOs

307

examine over 1,000 nonfinancial companies tracked by the Value Line

Investment Survey in 1986 and find a median insider ownership of 5.0% (and
an average of 11.8%). The much larger insider ownership percentages of
post-IPO equity suggest that significant incentives to maximize shareholder
value are still present in LBOs that return public but remain independent.
Stock ownership data are also available for the 31 public puchasers of LB0
companies for fiscal year 1989. Insiders own a median of 6.6% (and an
average of 17.1%) of the equity in these companies. These ownership
percentages are slightly higher than, but not significantly different from, the
corresponding percentages in McConnell and Servaes. Public LB0 purchasers, therefore, do not appear to be characterized by particularly concentrated equity ownership.
5. Cross-sectional

determinants

of public/private

status

The public/private
status results presented above control only for LB0
completion date and LB0 age. This section examines the relation between
public/private status and several other variables that may affect that status.
Univariate results are followed by multivariate results that also control for
the LB0 completion date.
5.1. The importance of size

As the market value of equity owned by undiversified LB0 equity owners

increases, the risk-bearing costs of these holdings also increase. The higher
these costs, the more likely should be the LB0 companys return to public
ownership. Similarly, an LB0 company may require access to public equity
markets to finance future investment after desired organizational changes are
implemented under the LB0 organization. If risk-bearing costs and the need
ultimately to gain access to public equity markets increase with the value of
the LB0 transaction, larger LBOs should be more likely to be publicly
owned.
Panel A of table 8 divides the sample approximately into size quartiles and
presents the current private or public status of LBOs in the different
quartiles. The percentages of privately owned buyouts from the smallest to
largest quartile are 68.2%, 64.3%, 54.5%, and 62.5%. Although fewer of the
larger buyouts are privately owned, the patterns are not monotonic, nor are
the differences significant at conventional levels. The chi-square statistic that
all quartiles have equal percentages is 1.85 with three degrees of freedom
(p-value = 0.61); the chi-square statistic that the smaller two quartiles have
the same percentage as the larger two quartiles is 1.46 with one degree of
freedom (p-value = 0.23). Surprisingly, this evidence implies at best a moderate role for the size of the LBO. One possible explanation is that risk-bearing

308

S.N. Kaplan, The staying power of LBOr

Table 8

Public or private status of (1) buyouts by transaction

size, (2) buyouts arranged
by LB0
partnerships,
merchant
banks, and all others, and (3) divisional and public company buyouts as
of August 1990 for 183 leveraged buyouts valued at more than St00 million and completed in the
period 1979- 1986.

Total
A. Smallest quartile, less than $150 M
Second quartile, $150 M-$244 M
Third quartile, $245 M-$459 M
Largest quartile, at least $460 M
B. LB0 partnership=
Merchant bankb
All others
C. Division
Public company
All buyouts

Status
unknown or
liquidated
as of
August
1990

Status
known
as of
August
1990

Percentage
publicly
owned
as of
August
1990

Percentage
privately
owned
as of
August
1990

47
46
45
45

3
4
1
5

44
42
44
40

31.8
35.7
45.5
37.5

68.2
64.3
54.5
62.5

50
24
109

4
0
9

46
24
100

45.7
29.2
36.0

54.3
70.8
64.0

9.5
88

7
6

88
82

38.6
34. I

61.4
65.9

183

13

170

37.6

62.4

aLBO partnership
LBOs are LBOs arranged
by Adler Shaykin, Clayton Dubilier, Forstmann
Little, Gibbons
Green, Hicks & Haas, Kelso, Kohlberg
Kravis Roberts,
Riordan
Freeman,
Thomas Lee, Warburg Pincus, and Wesray.
bMerchant
bank LBOs are LBOs arranged
by Allen & Company,
Bankers Trust, Citicorp,
Donaldson
Lufkin Jenrette, First Boston, Merrill Lynch, and Morgan Stanley.

costs do not vary much once a buyout has reached $100 million in size. To
find an effect, one might have to look at smaller buyouts as well.
5.2. The importance of actLIe incestors

Although Jensens (1989) arguments are relevant for all LBOs, he focuses
on active investors and LB0 associations. The primary examples of these
organizations are LB0 partnerships and the merchant banking divisions of
investment banks and commercial banks. Accordingly, I classify the sample
LBOs as involving one of three types of buyout investors: LB0 partnership,
merchant banking division, and other. LB0 partnerships include Adler
Shaykin, Clayton Dubilier, Forstmann Little, Gibbons Green, Hicks & Haas,
Kelso, Kohlberg Kravis Roberts, Riordan Freeman, Thomas Lee, Warburg
Pincus, and Wesray. These organizations are devoted exclusively to arranging
LBOs. Merchant banking divisions of investment banks include Allen &
Company, Bankers Trust, Citicorp, Donaldson Lufkin Jenrette, First Boston,
Merrill Lynch, and Morgan Stanley. All other sample LBOs are classified as
other. These include LBOs organized entirely by management and by less

S.N. Kaplan, The staying power of LBOs

309

well-known LB0 partnerships. This classification, therefore, also distinguishes between the less well-known and the larger, better-known LB0
sponsors.
Panel B of table 8 shows the current private/public
status of LB0
companies by type of LB0 investor. A lower percentage of the LBOs
arranged by LB0 partnerships, 54.3%, are currently private while the percentage of LBOs arranged by merchant banks that are currently private,
70.8%, is slightly higher than the 64.0% for all others. None of the differences between any two of these groups, however, are significant at conventional levels. LBOs organized by well-known LB0 associations, therefore, are
no more likely to be permanent than are other LBOs.

5.3. DiGsions Cersus public companies

The sample includes approximately equal numbers of LBOs of divisions
and of public companies. Although it is possible that LBOs of the two groups
are driven by the same underlying causes, there are reasons that divisional
LBOs might be different. First, information differences between divisional
managers and parent-company managers are arguably smaller than those
between public-company managers and the stock market. The managers in
LBOs of public companies may be able to use their private information to
purchase the company at a price below its true value. If so, the gains in such
LBOs would come from information advantages rather than from any superiority of organizational form. If such differences are important, public companies should return to public ownership more often and more quickly than
divisional buyouts.
Alternatively, one could argue that the willingness of divisional managers
to do an LB0 indicates that they believe the parent corporation was
managing the business incorrectly. This is consistent with Kaplan and
Weisbach (forthcoming), who find that one common reason for divestitures is
that the divested division is not part of the parent companys core business.
Although the parent may understand this, it may be unable or unwilling to
sell the division to another company or to the public if the division has not
performed well recently. With a buyout, the parent gets a higher price, the
managers make the necessary changes, and then the buyout company returns
to public ownership. According to this view, divisional LBOs would be more
likely to return to public ownership than LBOs of public companies.
Panel C of table 8 presents the current private/public
status of both
divisional and public company LBOs. Slightly more public-company
LBOs - 65.9% - are still privately owned than divisional LBOs - 61.4%.
The difference, however, is not significant and does not indicate a strong
difference between divisional and public-company buyouts.

310

&V. Kaplan, The staying power of LBOs

Table 9

Ordinary-least-squares
and logit regressions of the probability that a leveraged buyout is publicly
owned as of August 1990 as a function of LB0 size, presence of LB0 partnership.
division or
public-company
status, and year of LB0 completion
for 170 leveraged buyouts valued at more
than $100 million and completed
in the period 1979-1986.
(Dependent
variable equals one if
LB0 company is publicly owned at the end of August 1990.)
(1)
Ordinary-leastsquares regression

Constant
$150 M to S244 M
$245 M to $459 Ma
At least $460 M
LB0 partnership
DivisionC
1981 and earlierd
1982
1983
1984
1985

Coeff.

SE.

Coeff.

-0.15
0.08
0.18
0.16
0.06
0.06
0.11
0.21
0.36
- 0.03
0.09

0.11
0.11
0.11
0.12
0.09
0.08
0.14
0.16
0.15
0.11
0.11

- 1.53
0.38
0.83g
0.74
0.28
0.29
0.49
0.91
1.56
-0 12f
0:39

170
0.07

No. of obs.
R/log-likelihood

(2)
Logit regression
S.E.
0.54
0.48
0.50
0.53
0.38
0.39
0.62
0.67
0.66
0.50
0.48
170
- 106.9

The variables $150 M to $244 M, $245 M to $459 M, and at least $460 M equal one if the
LB0 transaction
value falls in those ranges, and zero otherwise. The intercept refers to buyout
transaction
values of less than $150 M. These values divide the buyouts into transaction-size
quartiles.
bThe LB0 partnership
variable equals one if the LB0 is arranged by an LB0 partnership
and
zero otherwise.
The division variable equals one if the LB0 company is a division of another company and
zero if it is a public company before the LBO.
dThe variables
1982 to 1985 equal one if the LB0 is completed
in that year and zero
othetwise. The 1981 and earlier variable equals one for LBOs completed
in 1981 and earlier.
Significant
at the 1% level.
Significant at the 5% level.
Significant
at the 10% level.

5.4. Multicariate

estimates

The results in the previous subsections show that the current public/private
status of LBOs does not vary significantly by transaction value, by whether
the LB0 is organized by an LB0 partnership or a merchant bank, and by
whether the LB0 is of a public company or of a division. All three of these
are univariate results. Table 9 presents regression results that control for
these three types of variables and for the year the LB0 was completed.
Regression 1 is an ordinary-least-squares
regression in which the dependent
variable equals one if the LB0 is publicly owned at the end of August 1990

LV. Kaplan, The staying power of LBOs

311

and zero otherwise. This linear probability model is presented for ease of
interpretation. Regression 2 gives analogous estimates for a logit regression.
The multivariate results generally parallel the univariate results, except
perhaps for the role of transaction value. The two largest transaction-value
quartiles are, respectively, 18% and 16% more likely to be publicly owned
than LBOs in the smallest transaction-value quartile. The 18% value for the
third quartile is significant at the 10% level. Although the relationship
between transaction value and public/private
status is somewhat stronger
here than in the univariate analysis, it is still modest and only marginally
significant.
Consistent with the univariate results, regression 1 implies that LBOs of
divisions are no more likely to be publicly owned than LBOs of public
companies; and LBOs organized by LB0 partnerships are no more likely to
be publicly owned than other LBOs. The completion-year dummies have a
nonmonotonic pattern, with only LBOs completed in 1983 being significantly
different from the others. The 1983 LBOs are 36% more likely to be publicly
owned than LBOs completed in 1956.
Although not reported in the table, I also have estimated regressions that
control for (1) the industry of the LB0 at the two-digit SIC code level and (2)
post-buyout operating performance. The industry analysis uses dummy variables for the 15 two-digit SIC code industries that have at least three sample
LBOs. Only two of the coefficients for the industry variables are significant.
LBOs in rubber products WC code 30) and food stores (SIC code 54) are
more likely to be publicly owned. The results for the other variables are
qualitatively unaffected by the industry controls. Post-buyout performance is
measured as the percentage changes in operating income and operating
income to sales from the last pre-buyout year to the first or second post-buyout
year. These data are available for only 56 of the sample buyouts. None of
these performance variables is significantly related to buyout public/private
status.

6. Conclusion
This paper documents the organizational status over time of large leveraged buyouts (LBOs) completed between 1979 and 1986. As of August 1990,
62% of these LBOs are privately owned, 14% are independent public
companies, and 24% are owned by other public companies. As time since the
LB0 increases, the percentage of LBOs that have returned to public ownership increases. The (unconditional) median time an LB0 remains private is
6.82 years. These results and the time pattern by which LBOs return to
public ownership suggest that LBOs do not remain private permanently.

312

S..& Kaplan.

The sraying power of LBOs

These results are consistent with Rappaports view that buyouts are transitory organizational forms.
At the same time, the median life of 6.82 years implies that although LB0
organizations are not permanent, they are not short-lived. Furthermore,
private ownership is not the only distinguishing characteristic of an LB0
organization. The paper also considers post-buyout leverage and equity
ownership. LBOs that remain privately owned maintain debt levels similar to
the levels when the LB0 was completed. LBOs that are currently public,
both independent and purchased, maintain debt levels lower than the initial
LB0 levels, but higher than pre-buyout levels and median public-company
levels. The independent public LBOs also maintain relatively concentrated
equity ownership. These results are consistent with Jensens view and the
importance of long-term incentives.
The paper also finds a moderate role for asset sales to strategic buyers in
LBOs. Fewer than 29% of the LB0 companies and 34% of the original LB0
assets are owned by companies with other operating assets. This is much
lower than the 72% reported for hostile takeovers and somewhat lower than
the 43% reported for six LBOs by Bhagat et al. over a shorter period of three
years. Furthermore. the time pattern of purchases of the buyouts by other
companies suggests that the median buyout remains independent for at least
15 years. LBOs do not appear to be primarily motivated, therefore, by the
opportunity to sell the company to buyers in related businesses.

References
Bhagat. Sanjay, Andrei Shleifer, and Robert Vishny. 1990, Hostile takeovers in the 1980s: The
return to corporate
specialization,
Brookings Papers on Economic Activity: Microeconomics,
l-72.
Jensen,
Michael,
1986. Agency costs of free cash flow, corporate
finance and takeovers,
American Economic Review 76, 323-329.
Jensen, Michael, 1988,Takeovers:
Their causes and consequences,
Journal of Economic
Perspectives 2, 21-48.
Jensen, Michael, 1989, The eclipse of the public corporation,
Harvard Business Review no. 5,
61-74.
Kalbfleisch,
John and Ross Prentice.
1980, The statistical analysis of failure time data (Wiley,
New York, NY).
Kaplan, Steven, 1989a, Management
buyouts: Evidence on taxes as a source of value, Journal of
Finance 44, 611-632.
Kaplan, Steven, 1989b, The effects of management
buyouts on operations
and value, Journal of
Financial Economics 24, 217-254.
Kaplan, Steven and Jeremy Stein. 1991, The evolution of buyout pricing and financial structure
in the 1980s. Working paper (University of Chicago, Chicago, IL).
Kaplan, Steven and Michael Weisbach, forthcoming,
The success of acquisitions:
Evidence from
divestitures,
Journal of Finance.
Kiefer, Nicholas,
1988, Economic
duration
data and hazard functions,
Journal of Economic
Literature
26, 616-679.
Lancaster,
Tony, 1990, The econometric
analysis of transition
data (Cambridge
University Press,
New York, NY).

S. N. Kaplan, The staying power of LBOs

313

Muscarella.
Christopher
and Michael Vetsuypens,
1990, Efficiency and organizational
structure:
A study of reverse LBOs. Journal of Finance 45. 1389-111-1.
McConnell,
John and Henri Servaes. 1990, Equity ownership
and corporate
value. Journal of
Financial Economics 27, 595-612.
Rappaport,
Alfred, 1990, The staying power of the public corporation.
Harvard Business Review
no. 1, 96-104.
Roundtable
discussion,
1990. The economic consequences
of high leverage and stock market
pressures on corporate
management,
Journal of Applied Corporate
Finance 3. no. 2, 6-37.
Schipper.
Katherine
and Abbie Smith, 1988. Corporate
income tax effects of management
buyouts. Working paper (University of Chicago, Chicago. IL).
Smith. Abbie. 1990, Corporate
ownership structure
and performance:
The case of management
buyouts. Journal of Financial Economics 27, 143-164.