Optimal Cash Tender Offers*

Jacques A. Schnabel
School of Business and Economics. WiIfrid Luurier University, Waterloa Ontario,
Canada N2L 3C5,
Ebrahim Roumi
Division of Administration, University of New Brunswick, PO. Box 5050,
Saint John, New Brumwick, Canada E2L 4L5

ABSTRACT
This paper develops a contingent claims analysis, a simple decision model for determining the opti-
mal price in a cash tender offer. The implied behavior of the optimal tender offer premium over the
market price of the target common stock is also investigated. Those readers interested in applying the
model m a y obtain an interactive computer program written in FORTRAN and implementable on a
personal computer from the authors on request.
Subject Amrrr. Corpomte Finma?, F i n a n d Phnnkg and Modehg, and Options W i n g Mod&.

INTRODUCTION
In recent years, mergers and acquisitions have been widely discussed in the
popular financial press. In the academic literature, empirical studies on the mar-
ket valuation effects of such financial events have been voluminous. Jensen [ 5 ]
and Jensen and Ruback [6] provide comprehensive summaries of these studies.
Normative models in this area, however, have been few and far between. In the
usual textbook treatment of this topic, a merger is viewed as nothing more than
a large capital budgeting decision; for example, refer to the discussion in such pop-
ular textbooks as Van Horne [ll] and Schall and Haley [8]. A singular exception
to the foregoing is the model for determining share exchange ratios in a merger
developed by Larson and Gonedes [7].
This paper attempts to remedy this omission in the literature. A simple norm-
ative model is developed for determining the optimal cash tender offer price. The
analytical apparatus of contingent claims analysis is exploited in the development
of this model.
MODEL DEVELDPMENT
A cash tender offer represents a cash bid by one company (henceforth, the
bidder) for the common stock of another company (henceforth, the target). In this
type of corporate acquisition consummation is achieved not by an exchange of shares
between the target and the bidder as in the Larson-Gonedes model, but rather by a
cash purchase of the target’s common stock by the bidder. The stockholders in the
target firm can accept or reject the offer any time during a certain length of time
(usually two weeks). If the offer is accepted, they tender or sell their shares to the
bidder at the stipulated tender offer price. Troubh [lo] has provided a fairly complete
description of the institutional details surrounding this type of corporate acquisition.
What motivates the bidder to acquire the target? After reviewing the empir-
ical literature of mergers and acquisitions, Copeland and Weston [3] arrived at the
following conclusion:
In a tender offer the Likelihood is greater that the bidder sees an opportunity for
improving the management of the target firm or that its shares are undervalued
for one of a number of possible reasons. [3, p. 6051

*Readers interested in obtaining the interactive computer program which implements the model
developed here should contact Dr. Roumi.

272
19891 Schnabel and Rourni 273

In this paper the motivation on the part of the bidder is assumed.

Define V, as the market value of the target’s common stock at time t. Con-
sistent with the literature in finance, V, is assumed to follow the diffusion process of

d V,/ V, = jdt+ &,

where p is the instantaneous expected rate of return, u is the standard deviation

of rate of return on the target’s shares, dt is a small increment of time, and dz
is the unit normal random variable. The model developed here (specifically, the
option pricing results invoked) assumes one of two actions. Either the target pays
no dividends during the tender offer’s period of availability or, if dividends are
payable during the tender offer period, the tender offer price is appropriately
adjusted downward. The former assumption is reasonable because the frequency
of dividend payment is usually once every three to six months and the period of
availability of the tender offer is typically two weeks. If the latter assumption is
invoked, the tender offer is said to be dividend protected. ?foubh [lo] has observed
that tender offers usually stipulate this dividend protection.
Consistent with the observation noted above [3], it is assumed that the bidder
can enhance the target’s market value by a certain proportion a if the acquisition
is successful, that is, the target’s stockholders decide to tender their shares. Although
the estimation of a! is outside the scope of this paper, the difficulties encountered
in assessing this value should not be ignored. Presumably, the bidder has studied
the management of the target in detail (e.g., its business strategy and financial pol-
icy) and has concluded that some positive value of a is achievable with suitable
changes in the target’s management. Clearly, the foregoing considerations call for
sensitivity analysis of the tender offer price with respect to a.This topic is addressed
later in the paper.
Define T as the length of time the tender offer is availableto the target’s stock-
holders. Although in theory the target’s common stock can be tendered at any time
up to and including T, in practice (as in [lo]) the bidder accepts tendered stock
only at T the expiration date of the offer. In addition, the time period involved
in a tender offer is usually quite short, approximately two weeks. Thus, to sim-
plify the development the assumption is made that the target’s stockholders’ option
to tender is exercisable only at time T. In the vernacular of finance, the target’s
option is “European” (exercisable only at ?J rather than “American” (exercis-
able at any time between now, time 0, and 7).
Given the foregoing, all uncertainty regarding the success or failure of the
tender offer will be resolved at Z Consistent with the notion of efficient financial
markets, if the tender offer is successful the bidder will experience a wealth incre-
ment in the value of the acquired target of a! times the value of the target company
at time T. The value of the company at T equals the value of its common stock,
V,, plus the value of its bonds, B. The uncertainties experienced during the tender
offer period are assumed to affect the firm’s equity but not the firm’s debt. Thus,
although V, is assumed to be an uncertain quantity, B is assumed to be a known
constant. One situation which would render the foregoing assumption valid is when
the target’s debt is riskless, that is, the target’s outstanding debt is small compared
to the target’s assets and the earnings generated by those assets. mically, target
firms are not aggressive about their use of debt financing. A popular reason for
acquiring a target is its unused debt capacity which the bidder can exploit to gen-
erate debt-related tax shields. Thus, the assumption of a nonstochastic B (aside
from simplifying the presentation) is reasonable in the present context.
214 Decision Sciences [Vol. 20

Define P as the cash tender offer price. As formulated here, P applies to the
entire target firm’s common stock rather than the target’s price per share. The cash
tender offer grants target shareholders the right to sell their shares to the bidder
at price Z? Define rp, as the value at time t of this put option which the target stock-
holders possess. At time T, the value of this put option may be formulated as

if Pr VT,
+.=( P-vT
0 if P c V ,

Observe that the stock would be tendered only if the tender offer price exceeds
the market value of the stock. It would be irrational for the target stockholders
to tender their stock otherwise Figure 1 depicts the relationship between +T and V p
Define S,as the value at time t of the bidder’s contingent claim on the wealth
gain from superior management of the target. A contingent claim is an asset whose
value depends on the value of another asset. Here S,depends on 5. The value of

rvT+B)
the bidder’s claim at time T, S , may be formulated as follows:

if P Z VT,
(2)
ST= 0 if P<Q.

Figure 2 depicts the relationship between ST and V p

Define W, as the bidder’s wealth at time t which is affected by the cash tender
offer. The value of this variable at time T WT, may be formulated as

Figure 1: Value of target’s put against value of target at T

19891 Schnabel and Roumi 275

Figure 3 depicts the relationship between WT and 5, Observe that W , may be

reformulated as the difference between ST and $ p In terms of the three figures,
the graph in Figure 3 is merely the vertical difference between the graphs in Figures
1 and 2. A similar statement may be made regarding the valuation of the bidder’s
wealth for any time t between the present, time 0, and time T. Thus,

Figure 2 Value of bidder’s contingent claim against value of target at T.

ST

a(B+P)

aB

P “T

Figure 3: Value of bidder’s wealth against value of target at i?

276 Decision Sciences [Vol. 20

w,= s,- 4,. (4)

Equation (4) makes clear that the bidder’s wealth is the difference between
the values of two contingent claims. S, is the value of the contingent claim which
+,
pertains to the bidder, whereas is the value of the contingent claim which per-
tains to the target. The objective is to choose the cash tender offer price which
maximizes the value of the S, net of the value of 4t at time 0, the date when the
offer is initiated. Before proceeding with the optimization process, however, each
of these contingent claims has to be valued.
Consider first the valuation of the bidder’s contingent claim at time 0, So. The
following valuation model is obtained by applying the theorem discussed in Appen-
dix A.

SO=aVo[l-N(dl)]+ c ~ B e - ~ [ l - N ( d 2 ) ]

where

N( is the cumulative unit normal probability distribution, r is the instantaneous

0 )

riskless rate of return, and e is the base of the natural logarithms.

The valuation of the target’s contingent claim at time 0, +o, involves a direct
application of the Black-Scholes put option model [ 2 ] .Thus,

MODEL OPTIMIZATION AND IMPLEMENTATION

As described above, the optimization problem is

max Wo= So - 4o (5)

with respect to P. The derivatives of So, d , and d,, result in

a s d a P = -av@(d,)(ad,/ap)- aBe-fln(d&3d2/aP), (6)

ad,/aP= ad,/ap= - i / ( p u j 7 ) , (7)

where n ( * )is the unit normal probability density function. Substitute (7) into (6)
to achieve

Observe that equation (8) is unequivocably positive. Thus, when the tender offer
price is higher, the value of the bidder’s contingent claim is higher. The intuition
19891 Schnabel and Roumi 277

underlying this result is clear in Figure 2 and equation (2). When P is increased,
both the probability of a successful tender offer and the expected payoff from such
an offer are increased. We take the derivative of 9o(by applying some of the results
in [9] and [4]) to obtain

Observe that equation (9) is unequivocably positive. Thus, the target's put option
is also a positive function of the tender offer price. This conclusion is an appli-
cation of a well-known result from option pricing theory which states that put option
values are positive functions of their exercise prices.
The following first-order optimality condition is obtained for problem ( 5 ) by
invoking equations (8) and (9):

Equation (10) simplifies to the following implicit equation in P:

a v$(dl) + aBe-fln(dz,
P=
a./r e-fl[l -N ( ~ ) I

Because d , and d2 are functions of P, equation (11) does not yield an explicit for-
mula for P. The roots of equation (11) satisfy the required first-order optimality
condition. In Appendix B we show that the root of equation (11). which is greater
than Vo, yields a unique maximum of Wo. Because values of P less than Vo are
of no economic relevance, we focus on the root of equation (3) which exceeds Vo.
The Newton-Raphson method was used to obtain the desired root of equa-
tion (11). A polynomial approximation for the cumulative unit normal probability
distribution was employed. This approximation is described in Appendix C.

NUMERICAL ILLUSTRATION AND DISCUSSION

Tb illustrate the model, assume a riskless rate of 10 percent, a time 0 market
value of the target of $10 million, and no debt in the target's capital structure.
Since tender offers are usually available for a maximum of two weeks and T is
quoted in years, set Tequal to .04.We focus here on the impact of 01 and a because
these are the only variables in the model that require subjective estimation; the
other variables are observables. Assume initially that a=.l5. Table 1 shows that
the tender offer premium over market price (TOP) is defined as

TOP =(P/Vo)- 1

and is a positive function of 01. This property is illustrated further by the family
of curves in Figure 4 constructed for the same parameter values stipulated above
and for u varying from .05 to S O . Consistent with intuition, the higher the per-
centage increase is in the value of the target achievable by the bidder, the higher
will be the TOP.The bidder should be more aggressive about seeking control of
the target as a increases. This translates into setting a higher TOP. Observe also
278 Deckion Sciences [Vol. 20

Table 1: Tender offer premium versus a.

(2 Tender Offer Premium
,050 .12679oooE- 01
.060 .17795300E-O1
,070 .22007350E-01
,080 .25576075E- 01
.ow .28623175E- 01
,100 .31326512E - 01
.I10 .3366625OE - 01
.120 .35765625E- 01
.130 .37681112E-01
.I40 .39402187E-01
,150 .4O970112E - 01
.160 .42406475E-01
.170 .43753712E-01
,180 .44986412E-01
.I90 .46133200E-01
,200 .47204125E - 01
.210 .48207675E-01
,220 .49151075E-01
,230 .50040487E-01
,240 S0881237E- 01
,250 S1677937E-01
.260 S2434575E-01
,270 S3154675E -01
,280 .53841312E- 01
.290 .54497212E-01
.300 ,55124800E- 01
.310 S5726212E-01
.320 .56303400E-01
.330 .568580758-01
.340 S7391775E -01
.350 .57905937E-01
.360 S8401800E - 01
.370 S8898075E-01
.380 S9362187E -01
.390 S9811275E-01
.400 .60246187E- 01
.410 .60667750E-01
,420 .61076700E- 01
,430 .61473737E-01
.440 .61859462E- 01
.450 .62234487E- 01
,460 .62599346E- 01
.470 .62954530E-01
.480 .63300520E-01
,490 .63637756E-01
.500 .63%6624E-01
Note: r=10 percent, V=flO million, T=.O4.a=.15, E=O.

that TOP is more sensitive to Q! when u is higher. For a very low value of a such
as .05, TOP is almost invariant to u. Indeed, when u equals zero TOP also equals
zero regardless of a.
With the stipulation that a= .2, Table 2 illustrates that the tender offer pre-
mium is not monotonically related to the volatility of the target’s market value
as measured by u. TOP initially increases in u, attains a maximum, and then decrea-
ses in u. Figures 5 and 6 illustrate this property for the same parameter values stip-
ulated above and for a varying from .15 to .50. The rationale which underlies this
behavior is derived from the fact that both the bidder’s contingent claim on supe-
rior management of the target and the target’s put option are positive functions
of u. The tender offer price is set to maximize the difference between these two
19891 Schnabel and Roumi 279

Figure 4 Tender offer premium versus a.

n = 50
*= 45
o= 40
a = 35
o= 30

o= 25

o= 20

o = 15

o= 10

a = 05

125 250 375

A L P H A

Figure 5 Tender offer premium versus u (part I).

,129,

T
097.
D
E
R

0
F
F 064.
E
R
P
R
E
032.
U
M

I
,125 ,250 ,375
S I GMA
280 Deckion Sciences [Vol. 20

Table 2 Tender offer premium versus u.

ci Tender Offer Premium
.050 .24640350E-01
.060 .27686037E-01
.070 .30517312E-01
.080 .33156187E-01
.090 .35616225E-01
.loo .37910612E-O1
.110 .4004%75E -01
.120 .42041900E-01
.130 .438W12E-01
.140 .45613350E-01
.150 .47204125E- 01
.160 .48671612E-01
.170 S0000225E-01
.180 S1250362E- 01
.190 S238645OE-01
.2oo 53411662E -01
.210 S4329287E-01
.220 S5142662E-01
.230 .55855112E-01
.240 .564698878-01
.250 .56990125E-01
.260 S7418837E-01
.270 S7758925E-01
.280 .58066762E-01
.290 58262062E-01
.3oo .5837%25E -01
.310 S8421762E-01
.320 58390675E-01
.330 S8288450E-01
.340 S8117062E -01
.350 .57878387E-01
.360 S7574212E-01
.370 S7206150E-01
.380 .56775787E-01
.390 S6284525E-01
.400 S5733712E-01
.410 S5124562E-01
.420 34458212E-01
.430 .53735687E-01
.440 S295795OE- 01
.450 S2125837E-01
.460 .51240187E-01
.470 .50301700E-01
.480 .49311050E-01
.490 .48268837E-01
.500 .47175650E-01
Note: r=10 percent, V=SlO million, T=.O4,a=.2,E=O.

contingent claims. Conceivably, then, u initially has a greater effect on the bidder’s
contingent claim than the target’s put option. Only after a high value is reached
does u have a greater effect on the target’s put option than the bidder’s contingent
claim.
%ken together, Figures 4, 5 , and 6 illustrate the need to conduct sensitivity
analysis on both u and a. When either u or 01 are high, TOP will be highly sen-
sitive; when u or 01 are low, TOP will exhibit low sensitivity. From the bidder’s
perspective, setting TOP too low would mean reducing the value of its contingent
claim. On the other hand, setting TOP too high would mean unduly enhancing
the value of that put option granted to the target by the bidder. These are the risks
involved in deviations of TOP from its optimal value-risks which a sensitivity
analysis can assist in addressing.
19891 Schnabel and Roumi 281

Figure 6 Tender offer premium versus u (part 11).

=a\‘- 50

a= 45

a = 40

a = 35

= 30

J
.-

237 .4?5 .?I2 950

S I GMA

SUMMARY
This paper has outlined a simple decision model for determining the optimal
price in a cash tender offer. It has shown that the bidder’s wealth may be viewed
as the difference in the values of two contingent claims: the bidder’s contingent
claim on superior management of the target and the target’s put option. The objec-
tive assumed was that of maximizing the bidder’s wealth. The optimality condi-
tion yielded an implicit equation for the optimal tender offer price. The behavior
of TOP was then investigated. It was demonstrated that TOP depends on the fol-
lowing two variables: the percentage increment in the target’s market value achiev-
able by the bidder and the volatility of the target’s market value. The pattern of
dependence of the tender offer premium on these variables was also demonstrated.
Although numerous simplifying assumptions were made in the development
of the model, the basic and novel approach applied in this paper-option pricing
analysis-holds promise for future refinements. For example, the informational
effects of a tender offer insofar as it causes the stock market to revise its expec-
tations regarding the target’s market value were ignored. It was also assumed that
no portion of the proposed acquisition would be consummated by an exchange
of shares. Only one tender offer price was assumed for the acquisition of all the
target’s shares. In practice, many tender offers specify a maximum number of shares
to be acquired and a two-tier tender offer price structure. These complications await
a more sophisticated option pricing analysis. [Received: September 16, 1987.
Accepted: March 4, 1988.1
282 Deckion Sciences [Vol. 20

REFERENCES
Abramowitz. M.,& Stegun, I. (Eds.) Handbook of mathematical functions. New York: Dover
Publications, 1968.
Black, F., & Scholes, M. The pricing of options and corporate liabilities. Journal of Political
Economy, 1973. 81, 637-659.
Copeland, T. E.. & Weston, J. F. Financial rheoty and corpomtepolicy. Reading, MA: Addison-
Wesley, 1983.
Galai. D., & Masulis, R. The option pricing model and the risk factor of stock. Journal of Finan-
cial Economics, 1976, 3. 53-81.
Jensen. M. lhkeovers: Folklore and science Harvard Buriness Review, 1984, 62(6), 109-121.
Jensen, M., & Ruback, R. The market for corporate control: The scientific evidence. Journal
of Financial Economics, 1983, 11, 5-50.
W o n , K. D., & Goneda, N. J. Businas combinations: An exchange ratio determination model.
The Accounting Review, 1%9, 44. 720-728.
Schall, L., & Haley, C. Introduction to financial management (2nd ed.). New York: McGraw-
Hill, 1986.
Smith, C. Applications of option pricing analysis. In J.L. Bicksler (Ed.), Handbook offinancial
economics. Amsterdam: North-Holland, 1979.
"toubh, R. Purchased affection: A primer on cash tender offers. Harvard Bwiness Review, 1976,
54(4), 79-91.
Van Home, J. Financial management andpolicy (7th ed.). Englewood Cliffs, NJ: Prentice-Hall,
1986.

APPENDIX A
Valuing So
To value So, split C, into two components, S,' and S,", where

s,=S,'+s,:

aV, if PzV,,
S,'=
if P<Vfi

and

if P r V,,
S,"=
if P < V,

Define So' as the value of S,' and So" as the value of S," at time 0. Clearly S,=So' +So".
Consider fiist the valuation of So'. If V, follows the diffusion process given by equa-
tion (1) and if a variable Q , is defined as follows:

if V , EnX,

if nXr V,>qX,

if V , < 3 X ,

where q,I& X, and T are arbitrary parameters and p is the average expected rate of growth
in Vl (i.e., E(Vr/Vo]=t@), then the expected value of QT, E(QT),is given by the follow-
ing expression [9, p. 831.
19891 Schnabel and Roumi 283

In( V,/*X)+ (p+(2/2))T In( V@X) + (p+(2/2)) T

E(QT)=epTxvo 1-N(
UJT uJT

When the risk-neutrality argument usually invoked in the option pricing literature is
applied, p = r in equilibrium. Setting n=1,X=a, and T=*=O in (Al) makes Q T iden-
tical to Sr'. Apply (A2) with these values to derive E ( S r ' ) = ~ a V , [ l - N ( d i ) ]Since
.
So' =e-flE(ST'), this implies that

So' = avo[
1-N(dl)]. (A31

Consider next the valuation of So". N(d9 is the probability that Vj->ls As E(S,? is
merely arB times the probability that V T s E this implies that E (ST")= d [ 1 -N(d2)].Since
So" =QflE(Sr"), this implies that

S, " = d e - f l [ l-~(d31. (-44)

Combine (A3) and (A4) to achieve the following valuation model:

so=~ V , [ I - N ( ~ ~d e) -If+l [ 1 - N ( d 3 ~ .

APPENDIX B
Maximizing W,,
The following property allows us to conclude that a stationary point of W,, P, which
exceeds Vo yields the maximum of Wo.
Property: For a given set of parameters (a,r, T a, and V$ there exists a value P, such
that Wohas a unique maximum P in the interval (Ps, a),where P, can be approximated
by the formula

P p v, exp[(r-(2/2)Tl (B1)

Proofi Suppose P is a root of equation (11). Thus aW,(F)/aP=O, that is, P is a sta-
tionary point of Wo.
To simplify the notation, define

f ( P )= aW,(P)/aP and f'(P)= a2W0(P)/&.

P yields a maximum of W, if f'(p)< 0 where

f'(P)= 1/(PaT~2~~(ad2/Pul(V~(exp(-.5~)+Bscp(-i;T-.5d~))

- J(1 +(aB/P))exp(-rT-.5d31. (B2)

Qpically, T has a low value and aB is less than P, so the second term in equation (B2)
is much smaller than the first term. Thus, as an approximation, the sign of
f'(P)is the same as that of the first term in equation (B2). Except for d,, all elements in
2 84 Decision Sciences [Vol. 20

the first term are obviously positive. Therefore, the sign of f ’ ( P ) is the same as the sign
of d,. To solve for P,, set d2= 0. The equation

In( Y d P )+ (r+(d/2))T
d, = -aJT=o
adT

results in

P,’ vo exp[(r-(2/2))T]. 034)

The preceding discussion implies that f’(P) c 0 for P > P,. Thus, if P > P,, it yields a unique
maximum of W,,.
The fact that the exponential function in equation (B4) is evaluated at a value close
to zero suggests that P, can be approximated with V,. Thus, for P > Vo, tFe second deriv-
ative of Wo with respect to P is negative. Therefore, a stationary point P which exceeds
Voyields a unique m;urimum of Wo. This is the value of the cash tender offer price of interest.

APPENDIX C
Approximation to N ( * )
The cumulative unit normal probability distribution, N(*),was computed using a poly-
nomial approximation formula described in Abramowitz and Stegun [l]. The approxima-
tion, ad),
is as follows:

where
X = 111 +pd,
p =.231649,
b, = .319381530,
6, = -.356563782,
b, = 1.781477937,
b4 =-1.821255978,
b, = 1.330274429,
(dd)<7.5x104,
Q(-4 = 1 - Hd),
Md) =1-Qo,
and e(d) is the error term.

Jacques A. Schnabel is Professor of Business and Head of the Finance Area of Wilfrid Laurier
University. He obtained his Ph.D. in finance at the University of New South Wales. Dr. Schnabel has
held teaching appointments at the New South Wales Institute of Technology and the University of
Calgary. He has published in the Journal of Buriness Research, Journal of Financial Research, Finan-
cial Review, and Managerial and Decision Economics.
Ebrahim Roumi is Associate Professor of Business Administration, University of New Bruns-
wick, St. John. He received his B.Sc in engineering from Arya-Mehr University, ‘khran-Iran, and
his M.A.Sc in engineering and PLD. in management sciences from the University of Waterloo. His
research interests include design and implementation of management information systems as well as
applications of operations research and operations management techniques to industrial and mana-
gerial problems.