<< . .

. 32
( : 66)



. . >>

Maher and Andersson (1999) refer to a number of articles that deal with
voting rights premia.22 Zingales (1995) ¬nds that while the voting rights
premium in the US is normally small, it rises sharply in situations where


21. In their chapter, they say as of December 31, 1994. However, I assume their use of moving
averages means that it is a span of time ending on that date.
22. I was unable to obtain those articles before this book went to press.




CHAPTER 7 Adjusting for Levels of Control and Marketability 219
control is contested, from which he infers that control shareholders re-
ceive private bene¬ts at the expense of minority shareholders.
The remaining evidence in this section is from other countries, where
concentrated ownership is the norm. Rydqvist (1987) ¬nds a 6.5% voting
rights premium for Sweden. Levy (1982) ¬nds a 45.5% VRP in Israel,
Horner (1988) ¬nds a 20% VRP for Switzerland, and Zingales (1994) ¬nds
an 82% VRP on the Milan Stock Exchange. The large voting premium in
Italy suggests high private bene¬ts of control, and Zingales (1994) and
Barca (1995) suggest that managers in Italy divert pro¬ts to themselves
at the expense of nonvoting shareholders. Zingales also measures the av-
erage proportion of private bene¬ts to be around 30% of the ¬rm value.
Zingales (1994) conjectures that the private bene¬ts of control in Italy are
so large because the legal system is ineffective in preventing exploitation
by controlling shareholders.

Bradley, Desai, and Kim (1988)
The authors document that successful tender offers increase the combined
value of the target and acquiring ¬rm by an average of 7.4% over the
period 1963“1984. In this article the 7.4% remains stable over the entire
22 years of analysis, which the Maquieira, Megginson, and Nail results
(in the next section) do, too. However, there was a constant movement
over time for the target shareholders to capture the lion™s share of syn-
ergies, with the acquirer faring worse over time. Also noteworthy is that
the authors present theoretical arguments why multiple-bidder contests
lead to larger payments to target stockholders.
The breakdown of the 7.4% overall synergy is very important to busi-
ness appraisers. The targets, who are, on average, 20% of the combined
entity (i.e., one-fourth of the size of the bidders), experienced an average
31.8% synergistic gain, as measured by cumulative abnormal returns
(CARs), and the bidders experienced a 1% synergistic gain. However, the
speci¬c results in different subperiods varied and will be signi¬cant in
my synthesis and analysis later in the chapter. The acquirers had CARs
of 4.1%, 1.3%, and 2.9% for July 1963“June 1968, July 1968“December
1980, and January 1981“December 1984, respectively. There is a clear
downward trend in the synergistic gains of the acquirers.
They also presented data showing that targets experience cumulative
abnormal returns (CARs) of 9.8% from ¬ve trading days before the an-
nouncement of the ¬rst bid to ¬ve trading days after the announcement.
A multiple bidding scenario increases the CARs by an absolute 13.0%,
which is consistent with the result from Schwert discussed above, al-
though not directly comparable in magnitude. Another interesting ¬nding
is that synergies were higher in multiple-bidder scenarios. As to the na-
ture of the synergies, the authors cite work (Eckbo 1983, 1985 and Stillman
1983) that indicates that the corporate acquisitions have no measurable
effect on the ¬rm™s degree of market power in the economy. This is con-
sistent with Maquieira, Megginson, and Nail™s results, discussed imme-
diately below, that the synergies are operating and not ¬nancial.

Maquieira, Megginson, and Nail (1998)
The authors examine wealth changes for all 1,283 publicly traded debt
and equity securities in 260 pure stock-for-stock mergers. They ¬nd non-



PART 3 Adjusting for Control and Marketability
220
conglomerate mergers create ¬nancial synergies. They de¬ne conglom-
erate mergers as those mergers in which the ¬rst two digits of the SIC
code of the acquirer and the target are different. They determine the SIC
code by examining the primary line of business listing for each company
in the relevant edition of the Moody™s manual. This data source differs
from Roach (1998), described earlier, and the SIC code scheme is different,
which may explain their different results.
To compute the synergy from the mergers, the authors used data
from two months before the merger to predict what would have been the
value of the two companies (and their individual classes of equity and
debt) as separate entities two months after the merger. They then added
the two separate company values together to form a ˜˜predicted value™™
of the merged entity. From this, they subtracted the actual market valu-
ation of the merged entity at two months after the merger, and they call
this difference the valuation prediction error (VPE), as well as the mea-
sure of synergy.
The mean and median VPEs for common and preferred stock were
8.58% and 8.55% for nonconglomerate mergers”and statistically signi¬-
cant at the 1% level”while they were 3.28% and 1.98% for conglomerate
mergers”and statistically insigni¬cant. For all classes of securities, which
also include convertible and nonconvertible preferred stock and bonds,
the mean and median net synergistic gains were 6.91% and 6.79% for
nonconglomerate mergers”and statistically signi¬cant at the 1% level”
while they were 3.91% and 1.25% for conglomerate mergers”and statis-
tically insigni¬cant. The positive VPEs in nonconglomerate mergers occur
in a statistically signi¬cant 66.4% of the mergers, while a statistically in-
signi¬cant 56.3% of the conglomerate mergers yield positive VPEs.
The breakdown between acquirers and targets is signi¬cant. In non-
conglomerate mergers, the acquirers had mean and median VPEs of
6.14% and 4.64%, while the targets were at 38.08% and 24.33%. In con-
glomerate mergers, the acquirers were the only losers, with mean and
median VPEs of 4.79% and 7.36%.
Maquieira, Megginson, and Nail also mention similar synergy ¬gures
provided by Lang et al. (1991), Eckbo (1992), and Berkovitch and Naray-
anan (1993). Another very signi¬cant conclusion of their analysis is that
the stock-for-stock merger synergies are operating synergies, not ¬nancial.
The authors also report the time to complete each merger, which are
interesting data and provide a benchmark for the delay-to-sale compo-
nent of my economic components model, described later in the chapter.
The time to complete the mergers ranged from a low of 11 months to a
high of 31 months”roughly 1 to 2 1/2 years. This underestimates the
time to complete a merger, as it starts from the announcement date rather
than the date at which the parties ¬rst thought of the idea.

Other Corporate Control Research
This section is brief. Its purpose is to present summary ¬ndings of other
researchers that will ultimately add to the discussion of what business
appraisers need to know about corporate control.
Franks and Harris (1989) analyze 1,445 takeovers in British stock
markets using the London Share Price Database. Their ¬ndings are very



CHAPTER 7 Adjusting for Levels of Control and Marketability 221
similar to those in the United States discussed in the previous sections,
i.e., that targets capture the majority of the gains from acquisition.
Cumulative abnormal returns (CARs)23 to the target shareholders in
Month 0 for single bids for which there were no revisions and no contest
were 20.6% and CARs for contested bids were 29.1%, for a differential of
8.5%. This is fairly similar, although somewhat lower than Schwert. The
CARs for Months 4 to 1 are much higher: 27.4% for the single bids
and 46.6% for contested bids, for a differential of 19.2%, which is higher
than Schwert™s result. There is an interesting intermediate category of
revised, but uncontested, bids, which the authors say probably re¬‚ects
results when the buyers are worried that another ¬rm might compete
with their initial lower bid. The CARs for this category are 28.7% in
Month 0 and 40.5% for Months 4 to 1. These might provide interesting
benchmarks for different levels of competition, both actual and potential.
CARs to bidders are very low, which echo the US results.
In a cross-sectional analysis of total wealth gains, multiple bidders
increase the control premium by an absolute 8.44%.24 This is also similar
to Schwert™s result, although slightly lower.
Harris (1994) provides an explanation of why any ¬rm would want
to be the bidder rather than the target. If target shareholders are the big
winners and bidders barely break even, then why bother being a bidder?
Why not wait for the other ¬rm to be the bidder and be the target instead?
The answer is that while the target™s shareholders are the winners, the
target™s management are losers. Harris cites another author who cites a
Wall Street Journal article that reported 65% of a sample of 515 target CEOs
left their ¬rms shortly after the acquisitions (less so in mergers). The re-
ward for the bidder is that management gets to keep their jobs. The re-
ward for target management is that the bidder pays a high price for their
stock, which gives many of them plenty of time to take life graciously
while looking for their next job.


Menyah and Paudyal
This research provides a method of quantifying the bid“ask spread (BAS).
Later in the chapter we review some of the work on DLOM by Larry
Kasper, which involves using an econometric equation to determine the
BAS to add to the CAPM-determined discount rate before DLOM. Those
interested in using Kasper™s method may want to understand this re-
search. Otherwise, this work is not used in my own models and can be
skipped.
The authors study stocks on the London Stock Exchange and ¬nd
the security prices, volume of transactions, risk associated with security
returns, and the degree of competition among market makers explain 91%
of the cross-sectional variations in bid“ask spreads (BAS) (Menyah and
Paudyaul 1996).



23. The authors actually use the term total abnormal returns.
24. See the x-coef¬cient for variable 2 in their Table 9.




PART 3 Adjusting for Control and Marketability
222
The average inside spread25 for liquid stocks was 0.83% before the
October 1987 stock market crash. It increased to 2% but has since declined
to 0.71% by the end of 1993. The average inside spread for less liquid
stocks declined from 10% to 6% over the same period. Transactions over
£2000 have lower BASs.26
The academic literature has identi¬ed three components to the BAS:
order processing, inventory adjustment, and adverse information. The au-
thors quote Stoll (Stoll 1978a, b), who says that because dealers must
service their customers, they cannot maintain an optimal portfolio suit-
able to their risk“return strategy. Therefore, total risk, not just systematic
risk”as measured by beta”is the relevant variable in determining the
BAS.
Their regression equation is: ln BAS 0.097 0.592 ln Price
0.649 ln 0.369 ln # Market Makers 0.209 ln Volume. All coef¬cients
were signi¬cant at the 5% level, except the y-intercept. R 2 91%.
The BAS equation shows that in the public markets, an increase in
volatility increases the BAS”and thus DLOM”while an increase in the
number of market makers decreases BAS and DLOM. With privately held
¬rms, there is no market maker, i.e., a dealer who is willing to buy and
sell. Business brokers and investment bankers never take possession of
the ¬rm. This is an substantial intellectual problem for one wishing to
use this model. Nevertheless, it may provide some useful benchmarks.


My Synthesis and Analysis
Trying to make sense of the oceans of research and opinions is like trying
to put together a giant picture puzzle. For a long time it was dif¬cult to
see where some of the pieces ¬t, and some did not seem to ¬t at all.
However, some coherence is beginning to form.

Decomposing the Acquisition Premium
Let™s begin by decomposing the acquisition premium into its potential
components:
1. Performance improvements.
2. Synergies.
3. Control premium, i.e., the pure value of control.
Performance improvements are the additional expected cash ¬‚ows
from the target when the bidder runs the target more ef¬ciently. In val-
uation profession parlance, synergies mean the additional value that
comes from combining the target with the bidder”let™s call that pure
synergy. Speci¬cally, it is that portion of the pure synergy in the control


25. The best bid and offer prices at which market makers are prepared to deal in speci¬ed
quantities are quoted on the yellow strip of the Stock Exchange Automated Quotation
(SEAQ) screens.
26. Commission rates are also lower since 1986. In 1991 the commission rate for small trades was
2% of transaction value, while trades over £1 million incurred commissions of 0.15% and
declined further in 1993 to 0.13%.




CHAPTER 7 Adjusting for Levels of Control and Marketability 223
premium for which the bidder pays. We never see the portion that the
bidder keeps in the Mergerstat acquisition premiums.27 However, in the
academic literature cited, synergy is used to mean the increase in value
of the combined entity regardless of the source. It is the combination of
the added value from performance improvements and pure synergy.28
Ideally, if we could quantify each component, we would want to
apply only the pure control premium, item 3, to a marketable minority
interest fair market value (FMV) to arrive at a marketable control FMV.
We would not want to apply the average amount of performance im-
provements in the market to a subject company, as it is more appropriate
to quantify the speci¬c expected performance improvements for the sub-
ject company and add those to the forecast cash ¬‚ows. This follows Mer-
cer (1998) and McCarter and Glass (1995). Finally, synergies normally
belong in investment value, not fair market value”unless the market is
dominated by strategic buyers and the subject company is a serious can-
didate for the M&A market.

Inferences from the Academic Articles
The Bradley, Desai, and Kim results are very revealing. On average over
22 years, the acquirers actually gained, with a CAR of 1%. This means
that the acquirers are not paying for control! They are paying for expected
cash ¬‚ows. There is no information in this article to tell us how to break
down the CAR to the target between performance improvements and
synergies. However, the Maquieira, Megginson, and Nail article provides
some information to enable us to do that. The VPEs for the nonconglom-
erate acquirers were about 11% higher than they were for the conglom-
erate acquirers, which suggests that synergies account for the entire pre-
mium. Bidders are approximately four times the size of the targets in
their study.29 Multiplying the VPE differential of 11% 4 44% attrib-
utable to synergies. Since acquisition premiums are rarely even that large,
this is evidence that acquisition premiums are being paid exclusively for
pure synergies and not for performance improvements.
The Roach article suggests the opposite”that the majority of the
increase should be from performance improvements, since there was no
pattern to the acquisition premiums by the difference in SIC codes. The
two articles used different schemes for determining a potential synergistic
merger. Maquieira, Megginson, and Nail only require that the two ¬rms
be in the same two-digit SIC code. A merger of SIC codes 3600 and 3699
would be nonconglomerate, while in Roach™s work they would be a dif-
ference greater than 50 and presumably the equivalent of conglomerate,
although he did not use that terminology. On the other hand, a merger


27. That also would be true about the portion of performance improvements that the bidder keeps.
28. We could add valuation corrections for underpricing errors to the previous list. To the extent
that bidders spot undervalued ¬rms and pay a premium for some or all of that
undervaluation, that portion does not belong in the control premium applicable to private
companies, as we already presume that the valuation before discounts and premiums was
done correctly.
29. The book value of total assets of the bidding ¬rms made up 81.2% of total assets of both ¬rms
combined, and the targets made up 18.8% of total assets.




PART 3 Adjusting for Control and Marketability
224
of SIC codes 3599 and 3600 would be a conglomerate merger according
to Maquieira, Megginson, and Nail, but a difference of one in the SIC
code in Roach™s work. It is logical that one can achieve synergies from
combining two ¬rms in similar but different businesses and that the two-
digit SIC code scheme is better for that purpose. For our analysis, we will
assume that the academic article is the more correct approach.30
Strong evidence supports this conclusion that acquirers are paying
for synergies and not control in Table 7-4, which compiles Mergerstat
acquisition premiums for control and minority interests. The average dif-
ference of control and minority purchase acquisition premiums was
0.16% (K7), and the average minority-to-control ratio was 99.11% (K8).31
If acquisition premiums were really measuring the value of control, then
minority interest acquisitions should have had a signi¬cantly lower pre-
mium. Instead, this is strong evidence that acquisition premiums are mea-
suring synergies, not control.
Back to Bradley, Desai, and Kim, in the 1981“1984 subperiod, the
acquirers did suffer a loss of 2.9%, as measured in CARs. This could mean
that the acquiring ¬rms were willing to suffer a net loss of 2.9% of market
capitalization for the privilege of control over the target. Since targets
were, on average, about one-fourth the size of the bidders, this translates
to 11.6% of target value. However, 4 years out of 22 does not seem enough
to assert strongly that this is a reliable control premium”let alone the
control premium.32
The Maquieira, Megginson, and Nail article provides similar results.
The only negative VPE was for the acquirers in conglomerate mergers,
who had a mean and median VPE of 4.79% and 7.36%. Multiplying
that mean VPE by four for the bidder-to-target size ratio leads to a pos-

T A B L E 7-4

Mergerstat Mean Premiums: Control versus Minority Purchases


A B C D E F G H I J K L

4 1990 1991 1992 1993 1994 1995 1996 1997 1998 Average
5 Control interest [1] 42.3% 35.4% 41.3% 38.7% 40.7% 44.1% 37.1% 35.9% 40.7%
6 Minority interest [1] 39.6% 32.6% 38.3% 38.3% 54.5% 61.7% 29.4% 22.4% 39.5%
7 Difference 2.7% 2.8% 3.0% 0.4% 13.8% 17.6% 7.7% 13.5% 1.2% 0.16%
8 Minority/control interest 93.6% 92.1% 92.7% 99.0% 133.9% 139.9% 79.2% 62.4% 97.1% 99.11%
← Total
9 SYD weights 2.8% 5.6% 8.3% 11.1% 13.9% 16.7% 19.4% 22.2% 25.0% 100.00%
10 SYD difference 0.1% 0.2% 0.3% 0.0% 1.9% 2.9% 1.5% 3.0% 0.3% 0.17%
11 SYD minority/control 2.6% 5.1% 7.7% 11.0% 18.6% 23.3% 15.4% 13.9% 24.3% 97.63%
13 # Transactions
14 Control interest 154 125 127 151 237 313 358 480 506
15 Minority interest 21 12 15 22 23 11 16 7 6

[1] Mergerstat 1999, Table 1-17, Page 25 (Mergerstat 1994, Figure 43, Page 100 for 1990-1993). Mergerstat is a division of Houlihan Lokey Howard & Zukin.




30. This is not to denigrate Roach™s work, which was very creative and is still signi¬cant evidence.
I made the same mistake in my own research.
31. In the data provided to me by Mergerstat, the average size of minority purchase was 38%.
32. Even though the regression coef¬cient was signi¬cant at the 0.01 level.




CHAPTER 7 Adjusting for Levels of Control and Marketability 225
sible control premium of 19.2% of the target™s pre-announcement value.
Perhaps this is a subset of the market that is paying something for pure
control, but it is not representative of the market as a whole.
Thus, it appears that our tentative conclusion, that the 7.4% differ-
ence between acquisition premiums in ordinary acquisitions and going
private premiums represents synergies, appears to be incorrect. However,
it is still possible that going private premiums might be the true control
premium. We will come back to this question.

The Disappearing Control Premium
Let™s consider the acquisition process. Conventional wisdom is that Com-
pany A buys control of Company B and pays, say, a 40% premium for B.
Therefore, B is worth 40% more on a control basis than on a marketable
minority basis. However, what happens after the acquisition? B no longer
exists as an entity. It is absorbed into A, which itself is a public ¬rm
owned by a large number of minority shareholders.
How can one justify the 40% premium to the shareholders of A?
Won™t the minority shareholders of A lose? If it is true that A is paying
purely for the control of B, then yes, they will lose, because the minority
shareholders of A pay for control that they ultimately do not receive.
Paying for control means that the buyer is willing to accept a lower rate
of return in order to be in control of the seller, and the Bradley, Desai,
and Kim results do not support that contention.33 After the acquisition,
who is in control of B? The management of A is in control, not the share-
holders of A. For there to be a pure value of control, it must go to man-
agement, who may enhance their salaries and perquisites for running a
larger organization. It makes sense that if ¬rms are paying for control
anywhere, it is in conglomerates. That only goes so far, though, before
the shareholders revolt or another ¬rm comes along and makes a hostile
takeover, booting out the inef¬cient management team (or a team who
looks after its own interests at the expense of the shareholders).
This seems to suggest that Mercer (1998) and McCarter and Glass
(1995) are correct. There is no value to control in itself. The appraiser
should simply try to quantify the performance improvements that one
can implement in the subject company, if they are relevant to the purpose
of the valuation, and proceed with the discounted cash ¬‚ow or guideline
company valuation. The difference in the marketable minority value and
the ˜˜control value™™ comes from the changes in cash ¬‚ows, not from a
control premium.

The Control Premium Reappears
Does this mean there is no such thing as a value to control? No. It™s just
that we cannot ¬nd it directly in the U.S. M&A market or in the public
markets, with the possible exception of the conglomerate mergers in the
Maquieira, Megginson, and Nail article. The reason is that there has to
be one individual or a small group of individuals34 actually in control


33. Again, with the possible exception of the 1981“1984 period.
34. Henceforth, for ease of exposition, reference to one individual in this context will also include
the possibility of meaning a small group of individuals.


PART 3 Adjusting for Control and Marketability
226
who derive psychic bene¬ts from it for there to be a pure control pre-
mium”and there is no such thing in the United States”almost.

<< . .

. 32
( : 66)



. . >>