UNIVERSITY OF

ILLINOIS LIBRARY

AT URBANA-CHAMPAIGN

BOOKSTACKS

&2A £ 2-3

CENTRAL CIRCULATION BOOKSTACKS

The person charging this material is re- sponsible for its renewal or its return to the library from which it was borrowed on or before the Latest Date stamped below. You may be charged a minimum fee of $75.00 for each lost book.

Theft, mutilation, and underlining of books are reasons for disciplinary action and may result In dismissal from the University.

TO RENEW CALL TELEPHONE CENTER, 333-8400 UNIVERSITY OF ILLINOIS LIBRARY AT URBANA-CHAMPAIGN

BUILDING USE ONLY

NOV 2 o me

NOV 2 0 19S6

When renewing by phone, write new due date below

previous due date.

L162

Digitized by the Internet Archive

in 2011 with funding from

University of Illinois Urbana-Champaign

http://www.archive.org/details/riskaversionbidd728smit

■iw. m

BEBR

FACULTY WORKING PAPER NO. 728

Risk Aversion and Bidding Behavior for Offshore Petroleum Leases

James L. Smith

College of Commerce and Business Administration Bureau of Economic and Business Research University of Illinois, Urbana-Champaign

FACULTY WORKING PAPER College of Commerce and Business Administration University of Illinois at Urbana-Champaign November 21, 1980

RISK AVERSION AND BIDDING BEHAVIOR FOR OFFSHORE PETROLEUM LEASES

James L. Smith, Assistant Professor, Department of Economics

#728

Summary

This paper examines economic determinants of bidding behavior in offshore petroleum lease sales. Multiple regression analysis is used to measure the explanatory influence of various firm-specific and tract-specific characteristics that enter the bidding decision. The principal finding is that most bidders display significant aversion to risk. An additional finding is that the magni- tude of tendered bids is directly related to the number of competitors in the market. The hypothesis that firms which are relatively short of petroleum reserves bid more aggressively is not supported by the data.

Acknowledgment

This research was supported by grant no. 14-08-0001-G-419 of the U.S. Geological Survey. The assistance of Holly Tomlinson and Carole Green is gratefully acknowledged. However, the author alone is responsible for the contents. The views and contents expressed here do not necessarily correspond to those of the U.S. Government.

(Uf, 9-

RISK AVERSION AND BIDDING BEHAVIOR FOR OFFSHORE PETROLEUM LEASES

1. Introduction

The present paper examines economic determinants of bidding behav- ior in petroleum lease sales on the United States Outer Continental Shelf (OCS) . Factors which presumably condition the magnitude of a firm's bid include: the expected economic value of the potential re- source; the degree of uncertainty which surrounds this value; the antic- ipated degree of competition to acquire the tract in question; and the firm's capacity to bear significant financial risks. We show below that systematic differences in these factors do significantly account for observed fluctuations in the magnitude of tendered bids.

Economic factors that impact bidding behavior are of considerable interest to auction participants and to the federal government, which administers the auction procedure. The interest of participants stems from their desire to formulate successful bidding strategies appropriate to their individual circumstances. The government's interest derives from the legal responsibility to design an auction procedure and adopt policies that elicit "fair" bids from the participants. Many of the leasing policy alternatives that we presently face (e.g., royalty bid- ding, profit-share bidding, and bonus bidding with sliding-scale royal- ties) can be expected to affect significantly the underlying factors (i.e., prospective risks and resource values, and the degree of competi- tion) that influence bidding behavior. An informed choice between cur- rent policy options requires some empirical knowledge of the impact which these factors exert on the strategic bidding behavior of firms.

-2-

2. Theoretical Ambiguities

Theoretical research has helped to clarify the influence which various economic factors exert on the bidding behavior of profit- maximizing firms. A summary of results is provided in the recent survey article by Englebrecht-Wiggans . However, theoretical formulations leave several ambiguities that cannot be resolved without empirical analysis of the particular market in question.

For example, it is well known that an increase in the number of bidders for an item of certain value will cause all bidders to raise their bids. Unfortunately, very few auctions offer the chance to bid for an item whose value is known, so this simple and intuitive result rarely applies. In the presence of uncertainty, the choice of a compet- itive bidding strategy is complicated by the phenomenon of "winner's curse"; i.e., the possibility that any particular firm will outbid its competitors as a result of having over-valued the item in question. To avoid this outcome each firm will adopt a less aggressive bidding strat- egy than is indicated by the certainty model. As Smith has shown, pro- nounced uncertainty may even cause the firm to reduce its bid as the number of competitors increases. Capen, Clapp, and Campbell have pre- viously advocated the use of such "non-aggressive" strategies in the context of OCS lease sales, but the prevalence of this behavior has not been determined.

We are also relatively ignorant of the effect of risk aversion in OCS lease sales. The nature of petroleum exploration imposes signifi- cant risks on auction participants. The traditional hypothesis of de- creasing absolute risk aversion states that the premium required to

-3-

induce a firm to accept such risks is inversely related to the size of firm in question. This implies, ceterus paribus, that larger firms would tender larger bids than their smaller counterparts. The applica- bility of the hypothesis of decreasing risk aversion is supported by Ramsey's and Millsaps and Ott's observation that bankruptcy risks are systematically related to the size of OCS participants. However, to the extent that nearly all participants are large, widely-held concerns, the risk of bankruptcy may be small in absolute terms. Consequently, it is not clear whether we should expect to observe systematic bidding differentials among firms of varied size.

The theory of risk aversion also suggests that bids are discounted in proportion to the perceived financial risk attending respective tracts. The most risky tracts would then be expected, ceterus paribus, to receive relatively low bids. However, Tourihno's recent extension of option pricing theory, in conjunction with the hypothesis of risk neutrality, establishes that if an effective futures market exists for petroleum reserves (including the option to sell short), then a higher variance in prospective value will increase the expected price at which the option (i.e., resource development) is exercised; thus leading to a higher bid for the riskier tract.

In the remainder of this paper, we explore the empirical validity of these and related hypotheses regarding OCS bidding behavior.

3. The Empirical Method Multiple regression analysis is used to measure the explanatory influence and statistical significance of various factors that account

-4-

for observed fluctuations in OCS bids, both among firms and among tracts. The estimated equations consist of linear and log-linear rela- tionships between the amount of each firm's bid and a set of firm- specific and tract-specific variables. The basic model includes three factors which are thought to be of fundamental importance: (1) a pro- spective estimate of tract value; (2) the size of the bidding firm; and (3) the number of bidders competing for the tract. The simplest equa- tion is of the form:

(1)

ln(Bid ) = a + b]L«ln(NWi) + b2«ln(N.) + b3«ln(V.) + e±.;

where Bid . = bid of the i firm on the j tract, NW. = net worth of the i firm,

t*Vt

N ■ number of bids on the j tract, V = expected value of the j tract, e = a normally distributed random disturbance. Subsequently, additional variables are added whose relevance to the bidding decision is more speculative. These variables include:

(1) the degree of the firm's self-sufficiency in crude oil production;

(2) the degree of the firm's self-sufficiency in petroleum refining; and (3) two measures of tract-specific financial risks.

The implications of Equation 1 are straightforward where solo bids are concerned. However, where a single bid is tendered collectively by members of a bidding consortium (a joint bid), there is some question regarding how the equation should be interpreted. In fact, it is pos- sible to treat joint bids in the analysis, and the method for doing so opens several interesting lines of inquiry.

-5-

We apply a method of decomposition to split each joint bid into "imputed" solo bids which describe the financial stake and potential re- ward of each consortium member. Let the equity share of the k member be represented by the symbol 6, (e.g., 35%). For that member it is as if a solo bid were tendered in the amount of Bid... = 6 «Bid. . on a tract whose prospective value is V , = 6 »V.. We assume the consortium formulates a joint bid that imputes individual bids consistent with the bidding preferences of each member (cf. Equation 1):

(2) ln(Bid*jk) = a + bylnOny + b2-ln(Nj) + b3'ln(V*k) + e±jk.

If this were not true, the consortium would have saddled some member with an effective bid that is inconsistent with its conception of appro- priate behavior, and we would not expect that firm to remain a member of the consortium. Indeed it is common to see individual consortium members selectively withdraw from bidding on particular tracts.

Our method of joint bid decomposition does not imply that the mag- nitude of an OCS bid is independent of whether the bid is tendered jointly. If b_ < 1, the resulting concavity of Equation 2 implies that the joint bid of the consortium exceeds the amount that individual mem- bers would tender if bidding alone. Thus, the "diversification effect" of joint bidding is represented implicitly in Equation 2.

To assess the validity of the bid decomposition procedure, we sepa- rate all bids into two groups — solo versus imputed joint bids — from which Equations 1 and 2 are estimated separately.

-6-

4. The Data The primary source of tract-specific data is the U.S. Geological Survey (USGS) , which prepares pre-sale evaluations of all offered tracts. Since 1973 the method of pre-sale evaluation has been linked to a Monte Carlo simulation model of offshore exploration and reservoir development. The simulation model is designed to calculate the expected net present value of each tract, while incorporating explicit probabil- ity judgments regarding the volume and value of recoverable petroleum reserves. Three elements of the simulation analysis are of interest here :

(1) Dry-hole Risk; a tract-specific parameter that reflects the probability that no petroleum is present.

(2) Expected Value; the average simulated net present value of the tract, incorporating the probability of dry holes.

(3) Development Risk; the standard deviation of simulated net present value, conditional on the presence of petroleum.

The USGS "expected value" is the estimate of prospective tract value that enters our regression analysis. "Dry-hole risk" and "development risk" together summarize the degree of uncertainty regarding prospective value .

The USGS also provides miscellaneous accounting information regard- ing OCS lease sales. The number of bids received on each tract, the amount of each bid, and the identity of each bidding party are recorded in the Lease Production and Revenue Database: System LPR-5. If a bid is tendered by a consortium, the membership composition is indicated on the basis of equity shares.

-7-

There is no administrative counterpart to the USGS that maintains a comprehensive record of firm-specific information. Reports available in the public domain provide only partial coverage of firms recently ac- tive in OCS sales. The sample is both small and unrepresentative, with only the operations of major companies being well documented.

Accordingly, we have extracted from corporate annual reports a data set that represents financial and operating characteristics of approximately 100 firms. The sample is not exhaustive, but it is be- lieved to be representative of the group of OCS bidders. It includes all major companies plus many small and intermediate-sized firms. The information collected for each consists of annual observations over six years (1971-76) on eight operating variables:

(1) total assets (5) crude oil proved reserves

(2) net worth (6) natural gas proved reserves

(3) crude oil production (7) rated refinery capacity

(4) natural gas production (8) actual refinery runs

The annual National Petroleum News Factbook Issue compiles data taken from the annual reports of approximately thirty petroleum compa- nies; only major companies are included. The Oil and Gas Journal peri- odically reports similar operating data, but coverage is again limited to major oil companies. The International Petroleum Encyclopedia ex- tends the sample to roughly fifty major companies, but reports on a restricted number of data items. The Financial Post Survey of Oils compiles statistical abstracts of most Canadian oil companies based on corporate annual reports and other information. The Federal Energy Administration (FEA) has published (1976) the most comprehensive account of the operations of integrated U.S. petroleum refiners, forty in number, but no financial data are included in the FEA report. Separately, the FEA (1975) reports the refining capacity of 154 U.S. refiners, but no other operating information is included. John S. Herold, Inc., a pri- vate investment analysis company, provides broader operating information on 130 U.S. and foreign firms that are engaged in various phases of the petroleum industry. The major limitation of this data source is that objective and subjective appraisals of firms' operating positions are intermingled.

-8-

The variable names are mostly self-explanatory. Production and reserve volumes are taken in physical units; i.e., barrels of oil and cubic feet of gas. The complete data set is available on request from the author .

We have selected seven recent OCS sales for analysis. The partic- ular sales are described in Table 1. Among the many lease sales that have been held since 1954, analysis was confined to these seven for three reasons:

(1) Each sale occurred after 1973, when tract-specific data from the USGS simulation model became available.

(2) Each sale is classified as a "wildcat" offering of previously unexplored territory. Consequently, we avoid informational asymmetries that might otherwise arise among firms due to prior experience in the area of the sale.

(3) Each sale drew bids on seventy or more tracts of widely varied value. This provides a large and fairly rich bidding record to support statistical analysis.

5. Empirical Results The Basic Model

We initially present results pertaining to the basic explanatory variables: net worth, tract value, and number of bidders. These vari- ables appear to exert a strong influence on the magnitude of tendered bids, and establish a benchmark for judging the contribution of other variables to be considered subsequently.

Estimated coefficients of Equation 1 based on the set of solo bids are presented in Table 2. Corresponding estimates of Equation 2 based

-9-

m

60 4-1

s c <u -a o.

^3 T3 -h

S "H O

3 B -H

S3 4J

M-l l-l

o cd

CM	vO	CN	O	00	a\
00	1^	ST	00	sr	CO

3

	T1	4-1
	co	c
14-1	u	•H
o	0)	O
	T3	•n
M	c
01	<D
J2	H
§	05	o
S5	TJ	fH
	•H	O
	m	CO

CO	o	<r	C\	\o	00	r^
rs.	rH	r-^	O	ST	m	\o
CM	CN		CN	r-1	»H	CM

CM

CM	00	»o	00	CO
CTv	r~	CO	00	sr

PQ <

H

Q

i-l CJ

CO

W

CO

w

CO

to u

c

CO •H

co pa

4-1 U O M

u d B

e h >

3 -H

Z M-l CU

O O

0)

CO CM

o r-.

CO

sr

00

iH							V
cd		CO			CO		•H
u		cd	cd		cd		4->
•H c	cd	X	•H	CO	X		3
.c o	c	01	C	c	0)		cd
P. vH	cd	H	H	cd	H		t-i
Cd 6C	•H		o	•r(		cd	u
U 01	0)	.C	U-l	CO	x:	a:	<
oo os	•H	4-1	•H	•H	4J	CO	1
o	3	3	rH	3	3	cd	13
0)	O	O	CO	o	O	rH	•H
c	►J	w	O	1-4	CO	<	s

	u
0)	0)
rH	.a	CO	sr	m	vD	r^	cr.	o
CO	e	CO	CO	CO	CO	CO	CO	sr
to

01
4J	<r	sr	CO	sr	m	*X3	v£>
cc	r>-	r^	r-»	r-^	r->	f^	r^
Q	>^	^^	~*^	^^.	■*-^	-*^	■^^
	00	o\	t-i	v£>	sr	CO	r~
0)	CM	CM	iH	t-l	o	rH	iH
ca	CO	m	CM	O	CM	sr	00
CO	O	o	rH	iH	o	o	O

-10-

so

m <r • r^

II II

CM 14-4 P2 T3

co SO

CO CM

• 00

II II

CM K-l

erf t3

vO

o

sr \o

• 00

ii II

a xi

00

o •** CM <-i

ii ii

CM <4H

o

CM

CO O rH

II II

OS TJ

CO 00

II II

erf -o

OS

00 SO -3- CO

II II

Csl <4H

►J

m <:

H

w c o

u

M CO

<:

PQ

CO

B

o

H H

<

Q- W

m

o

►J o

CO

CO CD CO

0)

•c

4J

c

0)

l-J c«

a

w o

4J CO

ce

4-1

CO

I

4J

			*			*
	*		*	*	*	*
^^	^"\	•~s	^%	/-s	^-s	•-s
rH O	m m	o oo	so m	r^ O	00 SO	rH SO
O CM	CM rH	i-i i-»	CM i-l	rH -tf	iH m	CO CO
o o	O CM	C rH	o <•	C CM	C CM	c m

*

SO CM

r- r->.

*

<f SO sC <•

rH CO

00 CO

rH 00

m oo

*

CM OS

so c

CO SO O CO

•K

O so

m o

c m

o -<r

c m

o cm

c sr

o co

	«				*	*
*	*				*	*
/— s	•— •.	^~*	^^	^~s	y— s	•"V
so l-»	r-t 00	o> m	r» m	r-» <t	<* CO	<* o
rH rH	CM CO	O CO	rH m	O CM	S3- rH	CO CO
O CM	O CO	O rH	O rH	O rH	O CO	O so
x_>	Sw'	1 '"-'	^-/	1 N~'	s-^	SW

•	c
c	•H
o
•H	X)
rH	4)
<-t	n
•H	D
6	CO
	CO
</>	01
	s
c
iH	01
	u
*o	CO
9)
1-4	0)
2	3
en	rH
CO	cfl
01	>
e
	4-1
CO	u
•H	CO
	u
X	4-»
4-1
P	■o
o	c
s	cfl
4-1	T3
aj	•H
S=	m

leve leve

*

vr so

l-» CM

82 79)**

58 32)**

* *

m sr

rH rH

86 58)**

.69 .15)**

.79 .96)**

at 5% at 1%

m cm

rH

s»/

co sr

m m

rH

<T O i-l

CM <t

cm r-

4-1 4-1

e c

CO co

o o

CO

en

m

CO

SO

CO

CTl

CO

o si-

*sign **sign

CO

o

CO

u o

CO

o o

CO

o

CO

u o

CO

u

o

CO

u

o

-li- on the set of imputed joint bids are presented in Table 3. The most striking feature of the results is the significance of the three vari- ables in accounting for observed fluctuations in the magnitude of bids. Each of the three factors is apparently of fundamental importance to the firm's behavior.

The number of competitors enters thirteen of fourteen equations at the 1% significance level, always with a positive sign. The magnitude of the coefficient ranges widely between 0.23 and 1.09, averaging roughly 0.65. This indicates that a 100% increase in the degree of com- petition has been associated historically with a 65% increase in the magnitude of tendered bids, ceterus paribus. Neither set of results (solo versus joint bids) produces a response elasticity that is signif- icantly higher than the other, although the dispersion of results admit- tedly leaves much uncertainty regarding the precise magnitude of the stimulative effect of competition. There can be no doubt, however, that the effect is positive. None of the fourteen equations supports the hypothesis that the phenomenon of winner's curse has caused widespread use of non-aggressive bidding strategies. This may indicate that bid- ders have been misinformed regarding the optimal formulation of bidding strategies. However, the industry's long experience with this particu- lar auction market creates a strong presumption to the contrary; i.e., that the phenomenon of winner's curse does not play a role so pronounced to require the adoption of non-aggressive behavior.

All coefficient estimates presented in this paper were obtained by the method of ordinary least squares.

-12-

.566 647

.390 505

CXi

C-g C

-a- m

• CM

.390 552

.149 502

.368 564

.462 1,09

ii ii

II II

n ii

ii ii

CM U-l

pa -o

CM M-l 05 T3

CM U-) OS 73

CM 14-1 PS T3

CM <4-4 OS T3

CM 14-4 OS T3

CM M-l

os -a

	*	*	*	-K	*	*	*
	*	*	*	*	*	*	*
	x~\	/— \	/^	*"->.	X-N	/"^	/— \
	O CM	CO CO	cr> r-~	on tH	O 00	V40 OS	o sr
>	rH O	co oo	CM VD	CM iH	<H st	<-t \D	en *jo
a	o m	O 00	O m	C 00	C co	O >X>	o CO

■J

W

^s

o

03

o

01

s

CO

a)

*

cj

J2

*

■K

*

M

4-1

/— N

^-%

•— N

*^N

*— N

s~\

^-N

CO

c

CT»

m

o>

00

o

CM

st

CO

00

r-^

VO

CO vO

<:

01

!S

C

Oi

st

m

CO

m

Ol

CM

CO

c\

CM

m o

M

n

I

a)

c

i-t

co

O

oc

O

CM

c

m

O

CO

o

ON

c \C

1

ex

rH

r-l

s^

* —

v— '

s^

s '

*w*

CO

c

S^

W

o

•H

►J

M

oa

£-i

(a

<

o

H

S3

c

i-l

4J

to

•H

Q

4-1

■K

*

■K

*

M

ct

*

■K

*

M

4-1

/— ^

/— »

•^

•*~s

/*~\

•~N

rf- s

CO

^

CM

O

st

<J\

o

i-l

st

OO

rH

CM

CT*

CM

■H m

1 4-1

s

st

v£>

CM

CO

\C

CO

m

CM

St

st

iH

CO 00

c

O

J-~

O

v£>

o

rH

O

r~.

C

CT\

O o>

O

fH

cH

S*'

* — '

iH

s*-'

*w*

s-^

>-}

\_^

s- *

•	C
C	•H
o
•H	•a
i-l	gj
.-)	i-<
•H	O
B	CO
	co
«■	0)
	S
C
1-4	0)
	>-4
X)	CC
0>
(-4	a>
D	D
CO	i-l
cfl	U)
Oi	>
e
	4-1
CO	O
-H	CO
	^
J=	4-1
•u
i-4	•w
O	c
IS	«
4-1	T3
01	•H
53	ea

i-H

1-1

*

0)

01

4->

*

>

c

X"S

/~N

/— s

^^N

^^

•^

s~\

0)

01

m

co r~

•H CO

I~- 00

vo sr

O &\

M3 CO

cr> st

i-H

r-i

4J

r» r~

CM CTi

O CO

m sr

CO CM

is- O

o o

CO

i>S

i-i

c

CO CO

co m

CO CM

co m

sj- \C>

cm sr

cm r~»

m

r-i

o

CO

i-i

*-i

CM

iH

CJ

4-1 0)

4-1

c cd u

•H

<4-l

iH

4-> Cfl

4J C

cd a

•H CH •H

CO

St

m

vO

r^

o

c

co

CO

st

oo

•H

ao

T-l

CO

co

GO

CO

CJ

u

o

u

o

*

o

O

o

■K

-13-

The influence of tract value is very much as expected. Tract value enters all fourteen equations with a positive sign, significant in twelve cases at the 5% level. The finding that bids are monotonically increasing in tract value confirms the theoretical results of Matthews, Vickrey, and Wilson. Of greater interest is the finding that none of the fourteen coefficients exceeds unity, which is the pivotal value for determining the effect of diversification. As discussed in Section 3, the concavity implied by b < 1 is sufficient to demonstrate that joint bids are of greater magnitude than would be tendered by Individual con- sortium members. The joint bid is evidently an effective institutional device for diversifying risks that are by nature indivisible. The di- rect effect of diversification is reflected in higher bids; the indirect effect is reflected in the rapid historical growth of consortia forma- tion relative to the practice of solo bidding.

The evidence regarding diversification is one indication that OCS bidders are risk averse. This conclusion is supported further by the performance of net worth in Tables 2 and 3. The coefficient of net worth enters positively in twelve of the fourteen equations, significant eleven times at the 5% level. Large firms have bid significantly more than small firms, after controlling for tract value and the degree of competition. This finding is consistent with the hypothesis of decreas- ing absolute risk aversion. An alternative explanation is that large firms are more efficient in the development of OCS reserves and therefore

The percentage of all bids tendered jointly, and of winning bids tendered jointly, increased from 32% in the late 1950 's to nearly 70% by the early 1970' s.

-14-

place higher values on the potential resource. However, all firms typi- cally contract with drilling specialists who provide development ser- vices, so this explanation is not compelling. It would appear more likely, a_ priori, that the bids of small firms reflect higher premia re- quired as compensation for bearing a relatively large risk. This inter- pretation is reinforced by additional corroborating evidence (presented later) that firms systematically discount bids on relatively risky tracts.

It is not coincidental that the influence of net worth is more erratic among solo bids (Table 2) than among joint bids (Table 3). In the solo bid equations, net worth enters significantly in only four of the seven cases; and actually enters with a negative sign on two occa- sions (but without statistical significance). This result should not be considered to be a statistical aberration, but to accurately reflect an important difference between the sets of solo and joint bids. The solo bids in our sample were tendered primarily by the largest firms in the industry, whereas the joint bids have been tendered by a wide range of small and large firms. Consequently, there is considerably less sample variation in the size of firms represented in the solo bid equa- tions (see Table 4). This difference, in conjunction with smaller sam- ple sizes, would account for the weaker statistical significance of solo bid results.

Moreover, if we adopt the working hypothesis of decreasing risk aversion, the disparate performance of net worth between solo and joint bids follows as a logical consequence of the behavioral model. The

-15-

TABLE 4 SAMPLE VARIATION IN FIRMS' NET WORTH

Net Worth ($ billion)

PCS Sale

Sale 33: Solo Bids Joint Bids

Sale 34: Solo Bids Joint Bids

Sale 35: Solo Bids Joint Bids

Sale 36: Solo Bids Joint Bids

Sale 37: Solo Bids Joint Bids

Sale 39: Solo Bids Joint Bids

Sale 40: Solo Bids Joint Bids

Mean	Minimum	Maximum
5.274	0.039	15.634
1.571	0.005	8.394
4.001	0.008	15.634
1.650	0.005	15.634
7.859	0.051	15.634
2.033	0.001	8.394
5.762	0.164	15.634
1.630	0.005	15.634
7.870	0.021	15.634
1.213	0.005	6.089
13.272	0.817	18.470
2.621	0.078	9.002
9.810	0.090	18.470
2.310	0.035	9.002

-16-

average size of firms tendering solo bids ranges across sales from $4 billion to $13 billion, as compared to roughly $2 billion for firms tendering joint bids. It is not surprising, then, that fluctuations in net worth have been less influential in the determination of solo bids, since the firms involved there are, by hypothesis, less averse to risk. The preceeding argument suggests that the log-linear specification (which implies a constant response elasticity) may be inappropriate as a model of bidding behavior. To further explore this question we re- specify the basic estimating equation as a quadratic relationship that permits the relative influence of net worth to diminish and eventually disappear where larger firms are concerned:

(3) Bid.. = a + b »NW. + b_«NW? + b.»N. + b , -V. + e...

ij Ii2i3j4jij

Consistent with the hypothesis of decreasing risk aversion, the expected

signs of b. and b„ are positive and negative, respectively. Moreover,

the "critical" firm size to attain risk neutrality (NW ) is determined

c

by: NW = - -^b./b,,. Increments to wealth beyond NW fail to increase c a X i. c

the tendered bid.

Estimates of Equation 3 based on solo and imputed joint bids are presented in Tables 5 and 6, respectively. Both tables lend considerable support to the hypothesis of decreasing risk aversion. Among joint bid equations, the anticipated sign pattern occurs in six of seven cases. Although the quadratic term is statistically significant in only four instances, a consistent pattern is established across the seven sales. Evaluation of the critical size to attain risk neutrality yields the estimates reported in Table 7. None of the seven sales generates an estimate that falls below $5 billion.

-17-

VO CO

• CO O 1^

II II

vO

n

o CO

II ii

CM "4-1 CM

OC X)

>l

	O	*— V
	fe	<u
	u	(0
	M	<u
	H	j=
	2	4J C
	5	0)
	t>	tO
	C	a
m	1 1	c
u	CO	•H
►J	2;
«	o	CO
•<	M	U
H	H	•H
	<	4J
	3	W
	o-	•W
	w	4-1 ce
	a	4-1
	M	co
	oq	1 4-1
	O	N— '
	J
	o
	co

cm in

ec -v

VO

in CO

• m o co

II II

CM <4-l 0= XI

CO

r^

cm en • rH

O rH

II II

CM <-M

(C T3

r-~

O O

• co

O rH

II II

CM 14-1

CM

CO

• VO

o r-^

II II

CM >4-l

PS TJ

sr o

CTV CM

rH ST

rH VO

en ST

r- i-i

VO St

•H		CO
ai	/— s	0) *-\
o	O-i	CO CT>
CTN	c	O m

CO c

CD ^-s CO CO CO rH

• •

•tf o

I v~'

CM

CU <~s rH CM O VO

• • «* O

I ^

CM

0) /-«.

iH vO

CM vj-

• •

i-l C

*

CO *

01 ^-\

CTi O

i-l CT\

• •

rH CM

CM O-i

CM m

• CO

O iH

II II

CM 14-1

05 T3

r-l		CM	CM ■*	*	iH	i-l *	■H *
1		1	1 -K	o *	1 *	1 *	1 -x
0)	•~N	m ^-s	01 /—.	oi *-s	4> x-s	01 ^-N	cu ^-s m i— i
o	m	CO CO	«* CO	rH O	CO co	iH CM
m	o	co <r	O VO	CO vo	m cm	<r m	rH CO

*	■K	*			•JC
VC *	VD -K	m *	in	m	m *	m *
CU -	CU ^^	cu /-s	CU /^N	cu *■%	CU ^-v	cu <— in >d
O CO	«* o	c-i r-»	CM Ov	O rH	o m
rH rH	sj- <r	vo m	O VO	m on	m vo	co in

0\ CM

CM \|	CM \|	CO	CM	CO	CM *	CM
CU /— >	CU x-\	0) *->	CU *-n	1 CU »*••	1 * CU ^~v	1 CU /"N m o
•<* o	00 VO	O CO	O rH	CO CM	rH m
vD vO	VO vO	m m	vo r-»	r^ vo	c vo	cm r~»
• # CO C	• • vO rH 1 V	• • CO o	• • CM o v_x	• • CO O 1 ^	• • m cm 1 s-'	• • sr O 1 ^

CO

CU <~\

00 CO

CM O

• •

rH rH

•	C
c	•H
o
•H	-a
H	CD
tH	i-i
■H	3
C	to
	cd
</>	01
	0
C
•H	0)
	Vj
-a	CO
01
u	CD
o	3
to	•H
CO	CO
<u	>
E
	4J
(0	o
•H	CO
	l-J
rC	4-1
4-J
M	-o
O	c
3	CO
4-1	TD
41	•H
E	PQ

4J

c

CO

VO

CU /-n

VO

CU <~«

CU ^-\

VO

CU *-*

m

cu *-»

*

VO *

vC * 0) x-s

in oo

CM <•

4J

CO

rH CT>

• •

rH CM

CO o>

• •

O r^ Ov O

<3\ vo

in vo

ST CM

m sj-

00 o r^ m

o o

CM O

1 *-*

rH O

1 "^

rH O

v/

CO o

• •

•* CO

I "^

• •

VO CM

1 s«-x

rH	rH
0)	41
>	>
at	4)
rH	H
.X'	K
m	^H

CO to

							c c
							tO tfl
							o o
							•H -H
							>4-i <4-l
CO CO	CO	in CO	VO CO	CO	CT> CO	o	•H -H C C 00 60
CO u o	o o	CO u o	CO u o	CO u o	CO u o	w u o	-H -H (0 (0 * * *

-18-

VO

vO vO

CO -3"

• VO

.306 504

.211 249

.228 551

.069 501

.435 563

.252 1,093

II II

ii it

ii ii

g M-4 05 TJ

CM 14-1 OS T3

CM <4H

os -a

CM C4H

os -a

CM 4-1

OS xs

CM 4H

oi -a

CM 4-1 OS TJ

	o	•— \
	PK	(0
	u	ca
	M	0)
	H	.c
	2	C
	§	0) u
	:=>	cd
VO	c i i	a, c
w	to	•H
rJ	z
«	o	CO
<	M	CJ
H	H	•H
	<;	4-J
	£3	(0
	c	•rl
	W	4-1 CS)
	Q	4J
	M	co
	«	1 4-1
	H	•w
	Z
	l-l
	C
	>-]

CM

r-l *	rH	rH *	t-l *	CM	rH *	rH *
1 *	1	1 *	1 *	1	1 *	1 *
CD >->>	0) /-N	CD <— ,	a) *— »	0) *-N	<u •-%	0) x-n
vO m	CO o	00 ».	i-t iH	m vO	CM CM	on m
vo co	O r-	o o	vO rH	n r-«	Ov CM	en <r

-<f r-

co m

*	*		*		■K	*
m *	m *	m	m *	-* *	m *	m *
CD ^-n	CD *-n	CD /-n	CD ^">	CD *~-	CD ^N	CD *"N
CM 00	i-- <r	rH <t	vc in	rH O	vo <r	vO -tf
vo m	rH vr	CO ON	CO ON	CO ST	CO CO	rH r-
m on	co r-.	rH rH	CM CM	CO CM	co m	rH CM

rH -K	CM	CM	CM *	CM	CM	CM
1 *	1 *	1	1 *	1	1 *	1
CD *-»	CD <-s	CD ^-s	oi /— .	0} /~-	CD ^-s	CD •— »
cm r-~	O t-^	ON CM	-tf rH	rH m	VO 00	m on
-d- <*	-* O	r-» m	VO vO	cm r--	■* <r	■H -vf
cm m	CO CM	CM o	00 CO	CM rH	VO CM	rH O
I ^	1 ^	v_/	1 ^	1 ^	I ^	1 ^

*	*		*	■K	*
CO *	CO *	CM	CO *	CM *	CM *	CM *
CD ^>	CD ^^	CD *-»	CD ^s	CD *^	CD /~.	CD <-n
ON O	CM CO	00 rH	rH O	co oo	r^ r--	sr <t
sf O	CO si-	o m	rH CM	cm on	00 O	00 rH
CM CO	rH t-s	CM O	rH VO	CM CM	oo sr	CO CM

• c
o
•H	TJ
H	<u
iH	H
•H	3
e	CO
	CO
</>	CJ
	e
C
•H	a
	n
X)	CD
CD
U	0)
3	3
i/i	t-i
cO	CO
0)	>
6
	4-J
CO	CD
•H	CO
	rl
.C	4-1
4J
u	TJ
o	c
s	cO
■u	TJ
CD	•H
SS	P5

4J	■K				*	*
c	VO *	m *	•o-	m	m *	VO *	m *
CO	CD «r^	CD <-x	CD /-n	CD /-\	CD •"%	<u •— .	Q) /-s
4-1	00 00	sf CO	vO Ov	vO CO	rH CO	O vo	CM CO
CO e o	m oo	CO sj-	o o	CO o-.	cm r^	rH rH	m rH
cm m	ON CM	m o	CO o	CM CM	CM >3-	ON CM
o	1 •-'	1 -	1 -	1 ^	v-^	1 ^	1 ^

a>	01
>	^
CD	cD
rH	H
6~^	S^
m	■H
4-1	4-1
CO	CO
4-1	4J
a	G
CO	cO

cj o

CO	>*	m	\0	t-s	ON	o	a c
CO	CO	CO	CO	CO	CO	-*	00 00 •H -H
CO	<Ji	w	CO	ca	CO	to	CO CO
'_>	u	c_>	o	o	u	CJ	* *
O	O	o	o	o	o	o	•K

-19-

Three of the four statistically significant estimates of NW range between $5 billion and $7 billion. The fourth exceeds $19 billion. The one estimate that is of marginal statistical significance (OCS 37) ex- ceeds $5 billion only slightly.

TABLE 7: ESTIMATES OF CRITICAL FIRM SIZE (in $ billion)

OCS Sale NW c-

33 $ 5,145

34 19,412

36 6,424

37 5,045 39 6,865

For perspective, the firms in our sample range in size from less than $5 million (Oxoco Oil Co.) to more than $18 billion (Exxon) in terms of net worth. The seven "major" companies are of particular in- terest. Of the seven, four hold net worth in excess of $7 billion (Exxon, Texaco, Mobil, and Standard of California) . Net worth of the remaining three (Gulf, Standard of Indiana, and Shell) ranges between $4.6 billion and $6.9 billion. Thus, the seven majors are the only firms that can reasonably be said to have attained risk neutrality. This conclusion is reinforced by the analysis of solo bids (Table 5) , where the behavior of the seven majors is dominant and where the influ- ence of fluctuation in net worth is of limited statistical significance.

Vertical Integration and Self-Suf ficiency

Many participants in the OCS leasing market are vertically inte- grated companies that engage in both production and refining of petro- leum. It is frequently alleged that firms relatively short of crude

-20-

oil supplies bid more aggressively to acquire petroleum development rights. We are not aware of any bidding models that establish this result formally; rather, the assertion is usually supported by general references to "security of supply" and "enhanced coordination" between different segments of the industry. These justifications are most seri- ously questioned because they ignore what many believe to be superior mechanisms for ensuring and coordinating supply; i.e., the many alloca- tive mechanisms which comprise the world petroleum market.

We have tested the "self-sufficiency hypothesis" by including two additional variables in the basic estimating equation. The first vari- able measures the ratio of the firm's production to proved reserves. The second variable measures the ratio of the firm's refining through- put to production. Both measures are taken from corporate annual re-

2 ports. The resulting estimating equation takes the following form:

(A) BidjLj = a + b1-NWi + b^^ + ^ + b^ + b$^ + bg^g- + e±j

If self-sufficiency is advantageous to firms, then both ratios should

enter Equation A with positive coefficients.

The estimated coefficients are presented in Tables 7 and 8, which

reflect solo and imputed joint bids, respectively. Although these

estimates are not strictly comparable to those presented earlier (the »

A recent example of this argument appears in the October 13, 1980 issue of Business Week.

2

Self-sufficiency ratios used here refer to proved reserves, pro- duction, and refining throughput in the North American Region. Alterna- tive specifications which include proved plus provable reserves and which incorporate world-wide operations were not found to materially alter the results.

-21-

CN

CO sr	m vO	00 vO	r~- CM	c sr	1^- CM	CM
CO • CM c m	m • m o sr	CO • m C vC	CO • CO O CO	1-1 r^ • iH O iH	vO O VD	co sr • iH O iH
II II	II II	II II	II II	II II	II II	II II
4 VM PS X>	CM M-l Pi TJ	CM 14M	CM IH OS "O	CM 14-1 PS XI	CM <4H PS no	CM <4M PS -D

-o o

Cm

CO CD

PS

-a o

Cm

	>h
	CJ
	M	CQ
	CJ	0)
	M	CO
	Pn	01
	fc	.C
	3	*j
	CO 1	c 01
	Ch	u
	►J	cd
	w	(X
is.	CO
w	1	c •H
hJ	CO
K	S3	CQ
<;	o	u
Eh	M	•H
	H	4-1
	<	CO
	3	■H
	O*	4J
	w	a) •M
	Q	CO
	M	1
	m	•u
	o
	►j
	o
	CO

Til

CO

01 »-N

oo r^

cm O

I •>-'

01 /-N

r^ o

i-l CM

• .

rH O

cm sr

sr cu m

o

01 /->

sr r^ rs O

• •

r-l O I N-'

CO

CU ys VO 00

sr m

CO C

CO

0) ^

co r-»

vo sr

• .

CM O

I s^

CO 0)

o sr

• CO I

I

*

sr *

0) <-N

VO 00

CO VO

. .

CO CM

I '-*

sr	CO	CO
01 ^-»	<b ^"N	01 <~\
O vo	r^ cm	rH 1^
rs oo	1-^ vo	co sr

sr *

04 ^-N C CM

CO rH

• .

CM CM

VO rH

ST *

01 /-\ CO r-H

vo o

• . T-i CM

iH VO

in co

r- iH

CM CM

oo sr

sr

04 •-> 00 O vO ov

. . CM iH

rH

i-i

T-i

•K

rH

T-i *

-1.77e- (0.96)

1.37e- (0.74)

l.Ole- (2.44)*

O *

01 /-n

cm r~.

. .

2.89e- (2.37)*

1.69e- (5.49)*

3.50e- (3.21)*

*

vO *

0) *s

T-i O

sr i-i

.81e6 .54)**

,93e5 .86)**

,55e5 ,61)

,94e5 12)*

09e5 98)**

*

vo *

at *->

VD CM CO Ov

CM |

CO |

CO

CM

T-i

CM

01 »-N

sr o

CO CM

0) ^v

oo r-« sr o

. .

,81e- ,18)

1

at *-\

00 o

m sr

1 *

0) y^ 0\ vD

rs o\

1

0) ^-s i-l 00

vo r^

40e0 09)

CM O 1 ^

m o

v— >

o-i o

CM o

T~i T-i

sr o

• •

oo o

.06e3 .49)

CO *

O vo CM O

. .

CM

<u ^-n

CM r-»

sr t-i

,19e2 ,48)

19e3 87)

CM

01 »-» cm in vo r-~

CM

01 ^ •H rH rH i-l

T-i O

CO CM

T-i o

m o

• • CM r-l

1 *•*

. •

Ov O 1 s^

CM O

•	c
G	■H
O
•H	-a
iH	0)
rH	Jm
•H	3
e	CO
	cd
<r>	ai
	e
c
•H	cu
	m
T3	cd
01
U	0)
3	3
CO	rH
cd	Cd
9J	>
S
	4-1
CO	CJ
■H	cd
	u
J=	u
4-1
u	-a
p	c
5	cd
■U	•c
0>	•H
z	PQ

•K

01 01

.27e6 .25)

.22e8 .29)*

.69e6 .76)

01 *"N

CO CO

vo r^.

0) *~\

OV rH vO 00

12e6 47)

vo

0)

sr in

vO

sr

% lev % lev

vO O 1 ^

•H sr

1 ***

vO O 1 ^^

•H rH

sr ^i

• •

r^ rH 1 s^

•

1

•

rH

ficant at 5 ficant at 1

CO

sr

CO

m co

vO CO

CO

o%

CO

o

sr

•H -H

a a

00 00

CO

CJ

o

CO

u o

CO CJ

o

CO

o

CO CJ

O

CO CJ O

CO CJ

o

*sl

**si

-22-

	S
	W
	M •-*
	U CO
	M 0)
	fe (0
	Pn 0)
	S3 .C
	CO JJ
	1 c
	[ft 0)
	►J M
	w co
CO	co c
w	1 c
h-1	c/> -w
«	Z
<	o to
H	H U
	H -H
	< W
	S3 (0
	0*-H
	W 4-1
	CO
	Q -W
	l-l (0
	m i
	4-t
	H v-1
	23
	M
	O
	►■J

0.308 214

0.315 239

0.208 80

0.163 190

0.069 174

0.415 386

0.282 704

ii ii

CM IH

ei xi

CM l*-l CXi X)

CM 4-1

a! x>

CM 4-1

CM 4-1 « X>

CM 4-1

PS XI

CM 4-1

oi -a

XI

o l-l

4-1 d) Pi

M 01 &

X)

o M

27e2 26)	CM 01 ON ITl	oo	CO 0> sr CM	* O o	rH 0)	VO rH	rH 0> st CM	m si-	rH 01 oo	CM CM	93e2 82)
iH O 1 ^	1	o	I	CM	1	O	1	t-i v->	CM	o	rH rH 1 *->
68e2 23)	CO 01 oo o	* i-l CM	CO Oi r-l		CO Ol CO CO	ON CO	CM 0> si-		CM 01 in si-	00 si-	24e2 52)
a> o l ^	r»»	CM	CO	O	r-l	o	sr	O	m 1	o	m o

rH

*

rH

*

rH

*

H

rH

«

rH

*

1

*

1

|

*

1

*

1

*

1 *

0)

^-N

01

^N

01

^^

0>

/~*

01 ^-v

0)

*^N

0) ir-S

>

St

o

CO

Ov

O

CM

O

r-~

O CO

oo

VO

rH vr

CM

si-

CM

O

VO

r^

si"

m co

Ov

CO

m oo

ST

CO

rH

CO

CM

oo

st

rH H

CO

CM

m cm

s_-

v— -

^s

w

s_>

rH

*

VO

*

m

*

m

*

sr

m

*

m

Ol

/— s

o>

^■s

0)

*~\

o>

^^

01 /-n

0)

•-N

01 <<"s

rs

CM

rH

m

o

m

CO

CM

m vo

VO

o

CO CM

H

00

m

si-

o

CM

m

CO o

o

m

vO O

l-l

VO

sr

m

CO

rH

m

CM

CO rH

»3-

>*

rH rH

rH 1

*

CM

*

CM 1

CM

1

CM 1

*

CM

1

at

^■N

0)

s~\

0)

s-*.

o>

y— \

0) <-N

01

/-^

01 <-N

i

o

m

si-

r»-

CM

r^

CT>

CO rH

f-.

vO

CO rH

g

r~

r~-

CO

m

o

CM

r-

m

O vO

r^

rH

co in

CM

r>»

CM

ST

O

VO

rH

CM O

r».

CM

rH O

1

V— '

I

^— '

v-/

1

*-^

*~s

1

v«/

1 «-'

*

•K

CO

■K

CO

*

CO

CM

*

rH

CM

*

CM

0)

•^

0)

^-s

01

/■^

01

^v

01 »""N

01

^*\

0) /-N

s

o

ST

rH

r^

Ov

O

m

Ov

in m

m

O

CJ\ rH

§

OX

S*

O

CM

i-»

CO

Ov

rH rH

r-~

o

Si- rH

CM

CO

CM

in

rH

o

co

rH

CO o

a»

CO

CO rH

4J

*

c

vO

*

VO

*

VO

vO

m

vo

*

m

a

01

^-\

01

•-N

0)

y-N

a)

^^

0) •->

0)

s-\

01 <"->

4J

m

CM

o

a>

CO

vO

Ov

CO VO

cy>

r^

co r~

10

r^

rH

CO

Ov

vO

\o

o

m

CO rH

CO

CM

vo rH

c

o

in

CM

ST

CM

VO

rH

o

>* rH

CM

rH O

o

I

N-/

1

«— /

N^»

'

s— '

K^

1

v^

1 ^

co

St

m

vO

r-~

CT»

O

CO

si-

CO

co

CO

co

o

CJ

u

o

C_)

O

o

O

c	c
o	•H
■H
rH	X)
rH	0)
•H	M
B	D
	CO
rH	CO
</>	0)
	e
c
•H	01
	rJ
XI	CO
o>
u	01
3	a
CO	rH
CO	CO
0)	>
a
	4-t
CO	O
•H	CO
	u
£	4J
•u
u	T3
o	c
&	CO
4J	X)
0)	■H
ra	«

0> 01

> >

0> 01

6^ 8^

in rH

to a)

c	C
CO	CO
o	O
■H	•H
VM	IH
■H	■H
C	C
CUJ	OO
•H	-H
09	en
*	*
	•>:

-23-

sample has been condensed to eliminate firms for which self-sufficiency data are not available), the previous pattern and significance of coef- ficients is hardly affected by the added variables. However, the self- sufficiency ratios themselves perform quite erratically. In only three cases out of twenty-eight do the measures of self-sufficiency enter sig- nificantly with the expected positive coefficient. In the majority of cases the measures enter with a negative sign (three times with statis- tical significance), in direct contradiction to the self-sufficiency hypothesis. Public assertions aside, the hypothesis that self- sufficiency is valued per se does not appear consistent with the actual bidding behavior of firms in the industry.

Tract-Specific Risks

In this section we consider the influence of tract-specific risks on bidding behavior. Analysis of this factor provides another test of the hypothesis of risk aversion, and generally reinforces previous con- clusions. To account for tract-specific risks we add two further vari- ables to the basic estimating equation: (1) USGS pre-sale estimates of dry-hole risk (DHR) , and (2) the standard deviation (SD) of development value ascribed to each tract by the USCS. The revised estimating equation takes the form:

(5) Bid., - a + b- -NW. + b-»HWT + b «N. + b.«V. + bc«DHR. + b^SIT + e,. ij Ii2i3j4j5j6jij

If greater financial risk is a factor that systematically depresses bid levels, we would then expect both coefficients (b_ and b,) to enter negatively. Alternatively, if the option pricing theory of asset

-24-

valuation applies (cf. Section 2) and firms are risk-neutral, we should then expect the coefficient of development risk (b,) to enter positively.

Estimates of Equation 5 are presented in Tables 9 and 10, which correspond to solo bids and imputed joint bids, respectively. The dry- hole risk factor enters negatively in nine of twelve cases and displays considerable statistical significance. In no case does this factor en- ter with a significant positive sign. Because expected tract value (inclusive of dry-hole risk) has been controlled elsewhere in the equa- tion, our results imply that firms value small but sure prospects more highly than those that are large but risky — even though their actuar- ial value is the same . What may be surprising is that this pattern is reversed in both solo and joint bid equations for OCS Sale #39 (Gulf of Alaska) . Members of the industry may have their own explanation for this reversal, but it is conceivable that the significantly higher costs and harsher operating environment of the Alaskan province diminish the attractiveness of all but the largest geological structures located there.

The measure of development risk (SD) also performs consistently across equations. The variable enters five of six joint bid equations with a negative sign, four times with statistical significance. This result lends additional support to the hypothesis of risk aversion (at least among the firms that have tendered joint bids), and strongly con- tradicts the prediction of the option pricing model. Both measures of risk (DHR and SD) exert less influence in the sample of solo bids. In- deed, development risk enters none of the solo bid equations with sta- tistical significance, and on only two occasions with a negative sign.

-25-

SI

co

u

CO

M ^s

M <U

C_> CO

tC 0)

(in .C

CO I

H

O

2

4-1

c

CU M CO

I -H

I CO

CO CJ

Z iH

O *->

M CO

!=> <a

K CO

Q -u

M *~>

O CO

Zl

co

CO

ce co

i- o

43 CO

CO >

TO

CO

XJ

CO

a

u

0) fX

CO

I

4-1 o CO r4

4J

o

c

co

co o O

0.416 66

0.467 60

rH CO

ST

. en o sr

0.057 99

0.541 64

0.271 119

ii ii

II II

ii ii

CN. CM

PS X)

eg ih

Pi XI

CNI M-l

PS 13

cm vh PS xJ

CM <4-l PS X!

CM 14-4

PS -o

CM 1	rH	CM 1	rH 1	CO 1	CM 1
01 •—>	CD •-s	cu <-n	cu •-*	a) /^	OJ /-N
00 ST	ON CM	r-~ vo	CM r--	rs r--	oo r-»
ST rH	cm t-»	00 l*»	ST CM	rH CM	00 00

sr

cu ^~»

oo o

-* ON

CM I

I

cu .--s o sr

CM ON

r-^ C I ^

*

NO *

CU *"N

to o CO oo

. • rH sr

CM I

CU /-N

oo to cm to

I

CO

(U *-N

00 o

O sr

rH rH

ST CO

CO

u o

rH O rH

rH O

CM O

*

sj- *

CU #"N

O CO CM rH

CO *

(U r"N rH 00 VO rH

CM

CU «""» ON 00 ST rH

CO

CM 00

NO rH

CM CO

I •*•*

CO

I

cu ^~N NO CO CM O

r»- CM

I «-*

CM I

cu ^-> oo o

CM rH

vO I

CM * I *

<U <-<•

rH ON rH VO

ON rH

rH * I *

CU /— >

t-» ON

ST vO

CM O

CO

I

NO CM

rH CM

*

to *

CU /"N

CM ON

O 00

. •

r^ sr

CM

I

CU /-N

ON CM

cm r~»

*

NO *

cu /"N

C to

CM CO

• .

rH St

CM I

<U /-s CO vO

on oo

*

to * at •->

m sr

CM CO

* .

CM CO

CM

<U ^"N

vO O rH vO

•K

to * oi ^*s

sr sr oo to

. .

NO CO

CM I

CU s~%

o c

vO vO

rH rH O

0) r-»

CO NO CO rH

rH O

a> /— »

CO U0 O rH

CM *

cu ^"N

CM St

vO ON

CO rH

CM *

(U /-N

no sr

CO On

vO O

I n-"

CO

01 ^-v 00 CO

in cm

r-» o I s-'

CU /-N

ON IT) VO 00

CO

to *

0> r-»

CO rH

00 O

ON CM

CM I

01 o-N CO NO

co oo

in

I

CO

(li /-s CO co vO CM

tO I

vO O I n-'

ON

rH rH

	*
vO	IS. *	VO *	to	VO	NO
CU r-N	01 '"N	0) /-s	cu /-v	01 r-N	01 /-»
co sr	ON CM	rH r-	on sr	cm sr	00 to
sr sr	co rs	VO ON	to 00	CO 00	CO o
Ov rH	rH CM	ST rH	00 rH	rH rH	rH O

to

CO

co u o

vO CO

CO

u o

r^

CO CO

u o

ON

CO CO

o o

c

sr

co u o

	c
	■H
	*C
	QJ
•	U
c	3
o	CO
•H	nt
rH	cu
rH	e
•H
e	CU
	S-J
<j>	CO
c	O
•H	CO
X)	XI
(1)	c
r4	CO
3
CO	*
cO	cu
cu	3
s	rH
	CO
CO	>
•H
	4J
4S	o
U	CO
n	r4
o	4-J
3
4-1	X>
01	•H
2	P3

cu	CU
>	>
CU	0)
rH	rH
M'	B<
to	rH
■u	*J
CO	CO
4-J	4-1
c	c
CO	CO
O	o
•H	•H
14-1	U-(
•H	•H
c	C
QU	00
•H	■H
CO	CO
*	*
	■K

-26-

	CO
	W
	co
	H
	PS
	**■%
	O 03
	H <U
	En CO
	M CO
	U .C
	W 4->
	P-< c
	co at
	1 u
	H cd
o	u a
rH W	§5
rJ	i
5	1 CO CO o
H	55 -H
	o 4J
	H CO
	H -H
	<: *j
	£3 OJ
	O* 4J
	W 0)
	Q «

P3

O

			vO	Sf	en	O	CN	vo
		o	vO	rH	o	oo	in r~
		si- r-»	en en	cn sr	rH rH	sr en	cm en
		• CN	• CT>	• vo	• m	• CM	• o
		o sr	O rH	O rH	O st	o m	O rH
		II II	II II	II II	ii n	n II	II II
		CM *4-l	CN M-l	cm y-i	CM vw	CM vh	CM M-l
		erf xt	« XI	pd xi	tH XI	P<S x)	»i XI
		CN	CN *	en	CN	en *	CM *
		1 *	1 *	1	1 *	1 *	1 *
			0) «""»	CO r~»	eo <"■■>	01 *s	CO «-N	0) t-\
			r~- vO	vo si-	O CTl	vo oo	r- st	rH rH
o\|			iH CN	r*» m	en ov	ST rH	en vo	!»-. vO
co\|			rH CN	vo en	CM o	en cm	m cn	en m
			1 ^	1 '"-'	I ^	N-^	I ^	I ^
				*		*		4S
		cn	sr *	en	en *	en	st *
	CO	0) *"N	at ^^	01 r^	a> ^N	co <"->	0) *"N
etjl	cn	rH CM	vO OV	CM 00	I-. rH	o\ r-»	st m
§1	01 rH	r-. en	ITl VO	r-» en	O CM	m oo	m o
i-i i-l	en vo	rH rH	CM st	m rH	cm in
	CO	^	1 ^	1 ^	1 ^	^->	1 ^
	CO
	U o MH
	rH *	rH *	rH	en	rH *	rH *
		1 *	1 *	1	1	1 He	1 *
	0)	0) »"■>	01 /-N	CO ^"N	0) r-N	CO «■-••	01 r-S
	rH	sr o	CTv i-l	00 st	en m	m cm	i~» m
>l	.£>	o en	cn m	r-» m	m O	CTi CM	CM st
	n)	• •	• •	• •	• •	• •	• •
	rH	vo sr	st en	rH rH	en O	st st	st O
	•H	V	s^	W	1 '-'	rH	rH
	tt)					«^/	Vw>
	>
	cd
		*	*	*		*
	cd	m *	m *	in He	en	m *	in
	4-1	o) ^-\	CO <-»	01 <— •	0) ^	01 <■*»	cu »~.
	cd	o o	00 vo	r^ r-»	vo en	r-» vo	H rH
21	X)	en tr>	en vo	o r-	VO VO	oo oo	CM vo
		• •	• •	• •	• •	• •	• •
	^j	en vo	si- cn	CM CM	CT\ c	en m	r-i rH
	CO	N— »	N-'	v^	^s	«w>	N->
	•H
	U
	o •H UH
	CN	CN	CM	CM	CM	cn
	•H	1	1	1	1	1 *	1
1 I	U	co •-»	01 <~>	CO <"N	01 ^"»	0) r*»	Q) «*"<
	01	O vo	i-l oo	rH r-»	00 CTi	oo m	o r~
*z\	CU	00 U">	sr oo	00 vo	st 00	i-~ cm	st CM
	CO 1	CN t-H	m o	CN r^	CM rH	m cm	VO c
	4J	^	%m*	*->	1 v->	1 s- '	i ^"*
	u
	cd
	u	*			*	*
	u	CN *	rH	rH	CM *	CM *	CM
		co *-»	01 *""»	CO -	01 *»»	0) /— v	0) »~ «
£5 l	a	o r-»	VO rH	<y\ rH	vo m	o oo	m o
§\|		a	rH st	en rH	st en	t^i o	m st	sr ov

en en

co c_>

o

oo sr

m o

sr o

cm en

en rH

	He		«		*
vO	i-~ He	vo He	vo He	vo He	r^ He
CO r«S	CO ^^	0) r"S	CO r-N	01 «-»	01 /-»
CM VO	en en	cn r>	r- vo	rH in	vO Oi
en vo	CM 00	r-» cn	vo vo	<r st	rH r^
rH rH	CN VO	rH CM	h sr	VO CM	CN st
1 ^	v^	V	N-'	1 ^

st

cn

oo U O

m en

co u o

vO

en

co U O

en

co O O

CTi

en

co O o

o

st CO

u o

	<^>-
	C
	tH
	X)
	01
•	M
c	3
o	CO
•H	Cfl
r-H	0)
rH	i
•H
e	0)
	l-J
<n-	cd
C O
•H	CO
XI	XI
01	c
u	cd
3
01	*
cd	a>
01	3
y	rH
	cd
cn	>
iH
	4J
J3	o
4-1	cd
H	m
g	4J
4J	X)
01	•H
2	K

CO 01

> > a) o>

m rH

cd <d

(3 C

cd cd

u o

•H >H

<tH «4H

•H -H

c c

•H -H

co co

He He

He

-27-

This result is again consistent with the hypothesis of decreasing risk aversion among large firms and cannot be explained by the hypothesis of differential efficiency in the development of offshore resources.

6. Summary Bidding patterns in recent OCS lease sales provide strong evidence in support of four important hypotheses :

(1) Greater competition for OCS leases inspires all participants to bid more aggressively. There is no evidence that the phenomenon of winner's curse is sufficiently important to lead to the adoption of non-aggressive bidding strategies.

(2) The majority of firms in the industry exhibit significant aversion to risk, but the degree of risk aversion decreases among larger firms and may eventually disappear.

(3) The mechanism of joint bidding, which facilitates the pooling and diversification of risk, has enabled consortia to bid more than individual members would under a regime of solo bidding. This conclusion may not apply to the seven major firms, who appear to have attained a position of risk neutrality. Not surprisingly, it is precisely this group of firms that have been the most active solo bidders.

(4) Self-sufficiency in crude oil production is not, per se, of sufficient importance to cause "crude-short" firms to bid more aggressively for OCS development rights.

These conclusions pertain closely to the debate regarding current leasing policy options. A traditional goal of leasing policy has been

-28-

to increase the degree of competition for leases. Our results indicate that this desire is well directed; the degree of competition is the most significant determinant of bidding behavior in our analysis. Of course, the analysis does not directly indicate which policy actions, if any, would encourage greater competition. However, the finding that most OCS participants exhibit significant aversion to risk suggests that any policy that is effective in mitigating OCS risks might induce more extensive and aggressive competition for leases.

The current thrust of leasing policy reform appears to be moving in this direction. Numerous auction methods currently in experimental use (e.g., royalty and profit-share bidding) are designed to substitute production-related payments for the present lump-sum bonus. By effect- ing this substitution the government would assume a significant portion of exploration risk, and presumably permit the companies to compete more aggressively. A disadvantage is that the government's new vehicle for collecting mineral rents may interfere with subsequent resource develop- ment by distorting marginal costs, but this carries us beyond the scope of the present paper.

Finally, an informed policy regarding joint bidding must reflect the benefits of risk-sharing. Because OCS risks are by nature indi- visible, and because most firms display considerable aversion to bearing risk, the benefits of joint bidding appear to be significant. Current policy, which prohibits joint bidding among the seven largest U.S. oil companies, but permits it elsewhere, is well-supported by our finding of decreasing risk aversion. The largest companies have not behaved as if the problem of indivisibilities were a serious one, but almost everyone else has.

-29-

REFERENCES

Business Week, "Mobil's Successful Exploration," pp. 112-18; October 13, 1980.

Capen, E. C, Clapp, R. V., and Campbell, W. M. , "Competitive Bidding in High Risk Situations," Journal of Petroleum Technology, pp. 641-53, June 1971.

Englebrecht-Wiggans, R. , "Auctions and Bidding Models: A Survey," Management Science, vol. 26, no. 2 (1980), pp. 119-42.

Financial Post Survey of Oils, Toronto: Maclean-Hunter Ltd., 1977.

International Petroleum Encyclopedia, Tulsa, Oklahoma: The Petroleum Publishing Co., 1976.

John S. Hero Id, Inc., Oil Industry Comparative Appraisals, Greenwich, Connecticut: John S. Herold, Inc., 1976 (monthly).

Matthews, S., "Risk Aversion and The Efficiency of First and Second Price Auctions," Working Paper No. 586, College of Commerce, University of Illinois, Urbana, July 1979.

Millsaps, S. W. and Ott, M. , "OCS Oil Leasing, Consortia Formation, and Bidding Behavior: An Application of The Theory of Risk Aversion," paper presented at the meeting of The American Economic Association New York, December 1977.

National Petroleum News Factbook Issue , New York: McGraw-Hill, Inc., 1976.

Oil and Gas Journal, "U.S. Oil's Profits Up But Still Shy of 1976 Levels," May 2, 1977, pp. 111-16.

Ramsey, J. B., "The Economics of Oil Exploration: A Probability of Ruin Approach," Energy Economics, January 1980, pp. 14-30.

Smith, J. L. , "Non-Aggressive Bidding Behavior and The 'Winner's Curse'," Economic Inquiry (forthcoming).

Tourihno, 0. A. F. , "The Option Value of Reserves of Natural Resources," (draft), Graduate School of Business Administration, University of California, Berkeley, September 1979.

U.S., Federal Energy Administration, The Petroleum Industry: A Report

on Corporate and Industry Structure and Ownership, Washington, D.C.: Government Printing Office, 1975.

U.S., Federal Energy Administration, Petroleum Market Shares: A Report

on Indices of Market Structure in the Petroleum Industry, Washington, D.C.: Government Printing Office, 1976.

-30-

U.S., Department of the Interior, Geological Survey, Conservation

Division, Lease Production and Revenue Database; System LPR-5, compiled by The General Services Administration, Fort Worth, Texas, December 1972.

Vickrey, W. , "Counterspeculation, Auctions, and Competitive Sealed Tenders," Journal of Finance, vol. 16, no. 1 (1961), pp. 8-37.

Wilson, R., "A Bidding Model of Perfect Competition," Review of Economic Studies , vol. 44, no. 3 (1977), pp. 511-18.

M/D/317

HECKMAN IXI BINDERY INC. |S|

JUN95

^.T0,^NNSHESTER