L?"""‘l ‘4, 9.",-

unyunrmnn

H ;;~.==:_-',¢n «Est-r4 1-H;

Lu .. -24..
Lu-EVui'au.‘

 

 

 

This is to certify that the

dissertation entitled

Alternative Auction Methods for Leasing Oil and Gas

Rights on Michigan's Public Lands
presented by

Douglas J. Krieger

has been accepted towards fulfillment
of the requirements for

M. S . degree in Agricultural Economics

 

 

fimmwu ”7

Major professor

Dwell/{GALA Ll: {QS’Q

MS U is an Affirmative Action/Equal Opportunity Institution 0-12771

 

 

 

MSU

LIBRARIES
1——

5

 

 

RETURNING MATERIALS:
Place in book drop to
remove this checkout from
your record. FINES will
be charged if book is
returned after the date
stamped below.

 

 

 

 

Alternative Auction Methods for Leasing Oil and Gas
Rights on Michigan's Public Lands

33'

Douglas J. Krieger

A Thesis

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

Master of Science

Department of Agricultural Economics

1986

Abstract

Alternative Auction Methods for Leasing Oil and Gas

Rights on Michigan's Public Lands

BY

Douglas J. Krieger

Michigan's public lands yield significant oil and gas resources.

It is characteristic of these resources that their value can not be

determined prior to actual development. This uncertainty has led to the

use of auctions in many leasing markets. The question addressed in this

study is how the performance of Michigan's leasing program may be af-
fected by the choice of a specific auction method.

This research proceeds with a review of the literature on economic
decision making under conditions of uncertainty and it's application to
the modeling of auction markets. Hypotheses are developed concerning
the connection between the leasing situation in Michigan, the choice of
auction method and the desired performance of the leasing program.

The purpose of this study is not to provide specific prescriptions
concerning the auction structure appropriate in the state. The
hypotheses presented here should, however, provide information of use to
decision makers in the Department of Natural Resources in designing

leasing policy that is consistent with desired objectives.

Acknowledgements

A number of people deserve recognition for their contributions to
this research effort. My major professor, Dr. Larry Libby, exhibited
great patience and provided invaluable guidance throughout the research.
For this he certainly has my sincere appreciation.

The other members of my committee, Dr. Allan Schmid and Dr. Peter
Kakela, also provided a great deal of helpful advice and guidance. Their
help is also greatly appreciated.

My thanks go also to David Freed and Rollie Harms of the Lands
Division of the Michigan Department of Natural Resources for their
assistance in this project.

Finally, all work and no play makes Doug a dull boy. In this
respect I would like to thank my family and friends who have provided
helpful comments, encouragement, support and the necessary diversions
from the studious life. Without them my stay here would have been much

less satisfying and rewarding.

ii

Table of Contents

Chapter One - Introduction .......................................... 1
Oil and Gas Leasing in Michigan ................................ 3
Summary ........................................................ 6

Chapter Two - A Conceptual Framework for Research ................... 7
The SSCP Paradigm .............................................. 7

The Paradigm Applied to Oil and Gas Leasing ............... 11
Market Situation ..................................... 12
Institutional Strucure ............................... l6
Participant Conduct .................................. 22
Leasing Performance .................................. 23

Specifics of the Leasing Problem in Michigan .............. 26
The Auction System in Michigan ....................... 28
Michigan Leasing Objectives .......................... 34

Summary ........................................................ 36

Chapter Three - Theories of Behavior Under Uncertainty .............. 38
Risk and Uncertainty ........................................... 39
The Expected Utility Hypothesis ................................ 41
The Behavioralist Approach ..................................... 44
The Modelling of Auction Markets ............................... 48
Applications to Decision Making in Oil and Gas Leasing ......... 50
Summary ........................................................ 52

iii

Chapter Four - Bidding and Auction Theory ........................... 54

A Review of the Literature ..................................... 56

The Independent Private Values Model ...................... 59

The Common Values Model ................................... 61
Alternative Auction Objects ............................... 65

Multiple Object Auctions .................................. 7O
Asymmetric Information .................................... 73
Reservation Price ......................................... 75

Risk Aversion ............................................. 77

Summary ........................................................ 80
Chapter Five - Oil and Gas Leasing in Michigan ...................... 84
Auction Method ................................................. 85
Bidding Objects and Payment Method ............................. 92
Reservation Price .............................................. 97
Summary ........................................................ 99

A Note on Empirical Work ....................................... 102
Bibliography ........................................................ 106

iv

List of Tables

Table 1-1 - Revenue From Oil and Gas Production on State Owned lands . 3

Table 2-1 - Oil and Gas Production on State Lands ................... 27

List of Figures

Figure 2-1 - Oil and Gas Producing Areas in Michigan ................ 29

Figure 2-2 - State Lands and Oil and Gas Fields --
Michigan's Lower Peninsula ............................. 30

vi

Chapter One

Introduction

In the United States government entities do not often engage
directly in the exploration for or development of oil and gas resources
but establish exclusive rights to these resources for private develop-
ers. The manner in which the transfer of rights from public to private
hands takes place would be expected to have some effect on the way in
which these rights are allocated between interested parties and on the
way in which development takes place. Information as to how different
institutions governing these transactions affect government objectives
and the satisfaction of public preferences will be important to govern-
ment decision makers. The purpose of this study is to investigate some
possible outcomes of different transaction mechanisms in terms of
various performance criteria. This information should be helpful in
designing policy that is consistent with the management objectives of
the government.

This study deals specifically with the transfer of exploration and
development rights to oil and gas deposits on public lands under the
jurisdiction of the State of Michigan. The study is primarily descrip-
tive in nature although some prescriptions will be made based on
specific performance objectives of the state. The focus will be on
possible consequences of different mechanisms used to allocate rights to
public oil and gas deposits and will provide prescriptions consistent

with desired objectives.

These questions are particularly relevant in Michigan for two
reasons. In the first place the State of Michigan ranks twelfth in the
nation in the production of both oil and gas with over 31 million bar-
rels of oil and over 142 billion cubic feet of gas produced in 1984.
Secondly, Michigan has more land under state jurisdiction than most
states east of the Rocky Mountains. Revenues from oil and gas develop-
ment in the state could contribute substantially to the state treasury.
This diversification of the economic base is important in a state that
is often seen as being too dependent on a single industry. The fluctua-
tions in the automobile industry brought about by periods of recession
have severely taxed state resources in the past. In 1984, revenues to
the state from bonus and royalty payments on oil and gas developments
amounted to over 58 million dollars. Table 1-1 provides a summary of
revenues for the years 1927-1984 split out by form of payment.

Aside from revenues from the development of publicly owned oil and
gas deposits, the state also gains from the development of privately
held deposits through the payment of taxes. In 1982, production from
private lands accounted for almost 23 million barrels of oil and almost
110 billion cubic feet of gas. These figures represent approximately
73% of the oil and 77% of the gas produced in the state (Department of
Natural Resources, 1982b). Severance tax payments on oil and gas
development totaled over $75 million in 1984.1 It might be expected
that the choice of institutions to transfer publicly held resources to
private hands will affect the manner in which private landholders bar-

gain with developers.

 

1 "Michigan's Oil Exploration-Production Industry at a Glance",

Table 1-1

Revenues From Oil and Gas Production on State Owned Lands

Years

1927-1931
1932-1936
1937-1941
1942-1946
1947-1951
1952-1956
1957-1961
1962-1966
1967-1971
1972-1976

1977

1978

1979

1980-10/1
10/1/80-10/1/81
10/1/81-10/1/82
10/1/82-10/1/83
10/1/83-10/1/84

Totals ($)

Royalty ($)

85,262
209,125
1,302,355
1,645,462
1,813,632
2,727,410
1,879,927
1,259,162
2,160,038
31,651,761
13,327,908
18,244,756
24,269,564
24,488,630
44,912,755
45,802,471
49,142,835
49,978,858

314,901,920

Rental ($

43,821
205,349
724,330
2,021,512
2,256,913
1,989,342

769,593
1,476,949
3,258,088
5,998,726
1,190,619
1,131,238
1,100,306
1,015,879
1,216,258

908,025

790,225
1,215,116

27,312,297

27,707
87,211
515,705
601,065
1,307,470
256,186
475,840
409,809
2,168,524
18,339,720
357,005
14,483
1,414,667
-2,982
32,277,235
10,304,780
630,029
7,513,180

76,697,639

) Bonus ($) Other* ($)

1,204
4,506
7,367
3,759
4,398
3,028
3,702
4,982
6,009
9,979
3,069
1,733
1,047
1,679
1,860
17,715
21,759
33,099

Total ($)

157,995
506,192
2,549,758
4,271,799
5,382,414
4,975,967
3,129,063
3,150,902
7,592,661
56,000,186
14,878,602
19,392,210
26,785,584
25,503,206
78,408,109
57,032,992
50,584,849
58,740,225

130,895 419,042,752

* This column is the sum of application and assignment fees

Source: Department of Natural Resources (1982b)

Oil and Gas Leasing in Michigan

The state of Michigan owns 3.8 million acres of surface and mineral

rights

and an additional 2.1 million acres of mineral rights alone.

Although leasing of great lakes bottomland for oil and gas production is

currently prohibited, the state also owns 25 million acres of such land

that could become available for lease in the future.

The

recovery of hydrocarbon resources from state lands in Michigan

takes place under the provisions of Act 17, P.A. 1921, as amended, and
Act 280, P.A. 1909 as amended. These acts provide for "... management
of land and mineral resources to ensure protection and enhancement of
the public trust". In addition, regulations have established the direc-
tor of the DNR as the supervisor of wells whose duty it is to provide
day to day enforcement of "...all matters relating to the prevention of
waste...and to the conservation of oil and gas in the state" (Patric and
Kakela, 1982a). The policy of the state in leasing oil and gas rights
is specifically stated as follows.

It shall be the policy of the Natural Resources Commission to

provide for orderly development of state-owned hydrocarbon

resources, to encourage the private sector rather than the
state to risk capital in exploration and development, to
optimize revenue from state-owned hydrocarbons consistent with
other natural resource management objectives, and to provide

for regular and systematic review of the oil and gas leasing

program.

Oil and gas leasing and development on public lands shall be

established in a manner to assure (l) optimum economic return

to the state, (2) competition for the acquisition for leases,

(3) protection against drainage of hydrocarbons, and (4)

protection of the environmental, recreational and other uses

of the land, including lakebottoms and connecting bays harbors

and waterways. (Department of Natural Resources, 1982a)

The leasing of oil and gas rights on state lands in Michigan falls
under the jurisdiction of the Lands Division of the Department of
Natural Resources (DNR) with lease sales held two or three times a year
depending on demand. Tracts to be leased are nominated by interested
parties and these nominations are reviewed by the DNR to determine the
desirability of leasing them as well as any restrictions that may apply
to the lease. Leases are sold primarily through an auction method

although provisions exist for direct sales under certain conditions. The

auction used in Michigan specifies a reservation price, the minimum

number of bidders required on a tract, the way in which bids are ac-
cepted and the object of the auction. The lease, once purchased by
auction, further specifies the term of the lease, the conditions of
reimbursement for resources actually recovered under the lease, regula-
tions governing development specifics and lease assignment regulations.
Some concern has been expressed within the DNR, however, that the auc-
tion method and lease terms currently in use are not well suited to the
attainment of leasing objectives. The purpose of this study is to shed
some light on the consequences of alternative auction methods and lease
terms .

To measure the performance of alternative auction procedures it
must be decided how performance is to be measured. The consequences of
different auction methods are often stated in terms of revenues to the
seller of the lease but other variables are affected and could be con-
sidered. Auction methods can affect the distribution of leases between
various types of prospective buyers, the manner of development or the
distribution of risk and benefits between the parties to the sale. This
study will consider a number of performance criteria in addition to
returns to the seller. The choice of these criteria is, in part, a
function of the concerns expressed by representatives of the DNR in

their leasing policy.

Summary

The objective of this study is to explore the possible consequences
of alternative auction methods for the leasing objectives of the State

of Michigan. Research requires a conceptual framework that will help to

organize ideas about the problem and suggest questions that are likely
to lead to resolution of the research objectives. Chapter two intro-
duces the situation-structure-conduct-performance paradigm as concep-
tualized by the institutional economists and develops it in terms of the
subject of the study. Chapter three further develops the conduct por-
tion of the paradigm. In this section current theories of behavior
under uncertainty are explored and related to oil and gas markets.
Chapter four applies the theories introduced in chapter three to bidding
and auction behavior. A number of distinct models of bidding behavior
are introduced and related to the sale of mineral deposits. Finally,
chapter five applies the theoretical developments of the preceding
chapters to the leasing situation in the State of Michigan. The focus
of this chapter is on the possible performance outcomes of alternative

auction structures under consideration for use in Michigan.

Chapter Two

A Conceptual Framework for Research

Research requires a conceptual framework that will help to organize
ideas about the problem under consideration and suggest avenues of
inquiry that will lead to resolution of research objectives. This
research will make use of the situation-structure-conduct-performance
(SSCP) paradigm that was introduced in the field of industrial organiza-
tion and adopted by the institutional economists. This framework sug-
gests that there is a causal connection between the situation in which a
market operates, the structure of the market, the conduct of market
participants and subsequent performance.

The purpose of this chapter is to explain in more detail the struc-
ture and implications of the SSCP paradigm and to fill out the com-
ponents of the paradigm in terms of the leasing of oil and gas
resources. In the course of fitting the paradigm to this particular
problem each of the four components will be defined in terms of the
relevant variables of the oil and gas lease market and the nature of the

interdependence between them will be explained.

The SSCP Paradigm

The SSCP paradigm has its roots in the field of industrial or-

ganization which is concerned with how the organization of production

activities affects the satisfaction of society's wants. The paradigm

has been adopted by the institutional economists, with some alterations,
for use in research on questions of public policy. While the paradigm
used by these two fields in economics is essentially the same, the
variables on which the framework focuses are very different because of
the difference in the questions that each addresses. Studies in in-
dustrial organization have frequently focused on traditional definitions
of market structure, behavior and performance (Scherer, 1980) while the
institutionalists have defined these variables much more broadly. It is
the emphasis of the institutional field that will be used in this re-
search. '

Schmid (1978) presents an excellent discussion of the components of
the paradigm and the interdependence among them as used by the institu-
tional school. The emphasis of the paradigm as presented here is the
development of testable hypotheses about the performance resulting from
structural changes in institutions. Situational variables are treated
as defining the characteristics of the world within which the institu-
tions work and, as such, must be treated as fixed if the effects of
structural change on performance are to be isolated from the effects of
changing situational variables.

Externalities are defined as the impacts of the actions of one
individual upon another. Externalities are ubiquitous and their
presence implies an interdependence among individuals in their actions
and uses of goods and services. In the SSCP paradigm the situational
variables describe the characteristics of goods and services that create
the opportunity for interdependence. The nature of this interdependence
can take many forms depending on the characteristics of the good in

question and the technology available to affect the pattern of use of

the good. A great deal of energy in institutional economics has been
expended in defining a taxonomy of the characteristics of goods that
lead to certain types of interdependence. Once situations are class-
ified in terms of their common characteristics that create interdepen-
dence similar structural alternatives can be matched to similar situa-
tions to create a more predictable performance outcome.

Institutional structure defines the rules of the game or the
property rights of the affected individuals within a certain situation
or interdependent relationship. Institutions define, explicitly or
implicitly, the rights of individuals when conflict arises. The choice
of institutional structure is the choice of how rights are allocated and
ultimately how conflict is resolved. In the sense that institutions
define the rights of individuals in conflict resolution they also define
opportunity sets, that is, the opportunities available to the involved
parties. The choice of institutional structure must be made in light of
the situation creating the interdependence and this is the importance of
the taxonomy developed by the institutionalists.

Institutional structure defines the structure of rights; it sets
the rules by which the game must be played but does not explicitly state
the outcome or specify a decision rule. The structure of institutions
sets up rules by which individuals interact and transactions take place;
they can structure the pattern of incentives and penalties faced by
individuals in their actions., In order to design institutional struc-
ture to meet desired outcomes or performance ends the behavior of in-
dividuals in response to alternative structural changes must be under-
stood. This is the conduct part of the paradigm and refers to the

conduct of actors within the system in response to the rules and

10

incentives provided by the structure of institutions. A realistic
theory of behavior is essential to the determination of performance from
a given situation and structure.

The final component of the paradigm is performance, that is the
outcome of the behavior of actors in response to the institutions that
define their opportunity sets for a given situation. The focus of
Schmid's framework is on substantive performance, that is on who gets
what in a positive sense, rather than on normative propositions of
efficiency that are commonly the performance criteria used in industrial
organization. Since performance is the ultimate goal of public policy,
the performance objectives of policy, as well as the conduct of economic
actors and the situation that creates interdependence, must be kept in
mind when designing institutions.

Performance, as the outcome of institutional structure, presents
some problems of measurement and documentation (Schmid and Schaffer,
1983). First, categories of performance must be established and related
to structural alternatives. Performance may have many aspects and, in
order to avoid excessive detail in research, the significant conse-
quences must be addressed while the unimportant ones are ignored.

Secondly, performance categories must be aggregated into indices that
can accurately account for performance changes. Performance must be
measured in meaningful terms, impact indicators that address the real
final impact of performance rather than intermediate products. There
may also be problems in defining performance. As Schmid and Schaffer
point out, some aspects of structure and conduct are valued directly for
their own sake and can rightly be considered aspects of performance.

It is important to realize the dynamic nature of the relationship

11

between the components of the paradigm. While the paradigm implies a
causal flow from situation to performance, there can be other interac-
tions as well. Over time, conduct can influence situation and struc-
ture. Technological change can affect the situation, or the nature of
the opportunity for interdependence. Institutional structure can also
ultimately affect the situation. The dynamic character of the entire
process becomes evident when it is recognized that performance in one
time period can change aspects of situation or structure in subsequent
periods.

The SSCP paradigm provides a framework within which to explore how
rules, or institutional structure, affect individual choice and how this
in turn affects substantive performance. The emphasis is on the for-
mulation of testable hypotheses about how institutional structure af-
fects performance so that institutions can be designed to yield certain
desired performance goals. In order to effectively use the framework,
the nature of the interdependence between individuals (the situation) as
well as human behavior must be well understood. Performance goals must
be clearly stated and indices formulated in terms of impact indicators
to determine whether institutional structure is effective in reaching

these goals.

The Paradigm Applied to Oil and Gas Leasing

This study uses the SSCP paradigm to analyze the leasing of oil and
gas resources in Michigan. Stated in terms of the paradigm this study
will investigate the effects of alternative institutional structures for

leasing, that is alternative auction procedures, on the performance of

12

the oil and gas leasing market in the State of Michigan. It is
hypothesized that variations in the structure of the rules of leasing
will affect the substantive performance of the leasing program in terms
of the distribution of leases, the allocation of the benefits of
hydrocarbon development and other performance variables. In order to
proceed with the analysis the oil and gas lease market must be defined
in terms of the components of the SSCP paradigm. In the remainder of
this chapter the four components of the paradigm will be expanded in the

context of Michigan's leasing market.

Market Situation: The situation refers to the characteristics of the
good that create interdependence between users. In the oil and gas
market this encompasses the current demand and supply conditions for oil
and gas resources, certain characteristics of the resource itself and
the technology available to develop and use the resource.

Supply and demand conditions for oil and gas resources, or at least
the way they are perceived, have changed dramatically since 1970. Energy
resources in general and oil and gas in particular have attracted a
great deal of attention since the oil embargo of 1973 and the creation
of OPEC. These events not only made oil importing countries more aware
of their dependence on foreign sources of energy but spurred these
countries to seek alternative supplies in both foreign and domestic
markets (Landsberg, 1979). In the U.S., this resulted in policy
designed to increase the exploration for and the development of domestic
resources of oil. Deregulation of oil prices in the U.S. was supposed
to stimulate exploration and, in time, increase domestic supply and

reduce dependence on imports. The increased prices brought about by

13

deregulation have focused more attention on the way in which a valuable
publicly owned resource is developed.

Government policy has also been aimed at the development of alter-
natives to oil as an energy source. The U.S. has an abundant supply of
coal and production and use of this resource have been on the increase.
Special tax treatment has provided incentives for investment in less
traditional sources of energy such as solar, wind and oil shale. While
these alternatives constitute substitutes for oil the change from one
source to another is hampered by technological progress in making these
sources economically feasible and the fixed investment in oil dependent
technology.

Increased oil prices brought about by OPEC price increases and U.S.
deregulation certainly had an effect on both consumption and sources of
oil. In the short run high prices have resulted in conservation
measures which have reduced the use of some oil products, most notably
gasoline, and have increased the interest in alternative energy sources.
Deregulation has resulted in greater exploration efforts in the U.S. and
a larger known domestic reserve. In the short run substitutes for oil
have not played a major part in the market because of investment in oil
dependent technology but in the long run oil could become much less
important as a source of energy as alternatives become economically and
technically feasible.

Probably one of the most important aspects of market situation,
from the standpoint of institutional design, results from the charac-

teristics of oil and gas resources. Deposits of oil and gas lie

14

underground, sometimes at considerable depth, and their fluid nature
allows them to migrate through the porous rock in which they occur. The
technology of exploration is not developed to the point that the exist-
ence of deposits can be detected with any great degree of accuracy
before actual drilling takes place. Since drilling is very expensive,
and in most cases prohibited before the sale of the lease, the informa-
tion costs as to the value of a lease are very high. High information
costs contribute to a great deal of uncertainty as to lease value for
both the buyer and the seller of the lease.

Ramsey (1980) discusses two sources of uncertainty, those of in-
herent risk and ignorance. When uncertainty results from ignorance the
acquisition of information can reduce or eliminate the uncertainty, the
cause of uncertainty is insufficient information. While sufficient
information can eliminate ignorance as a source of uncertainty it is
important to note that some degree of ignorance is almost always
economically rational, that is it is not rational to reduce ignorance
completely. Information is expensive, even when ultimately knowable,
and perfect information is infinitely so. In this situation it would be
irrational to purchase information sufficient to eliminate uncertainty
resulting from ignorance. Inherent risk refers to uncertainty resulting
from the unknowable, no amount of information will reduce this uncer-
tainty. In the oil and gas market the size and value of deposits is a
random variable. The gathering of information through seismic tests may
be helpful in estimating the distribution of the random variable but
does nothing to actually reduce it's randomness. While this uncertainty
is caused by ignorance, implying that it is ultimately knowable, current

development rules make it essentially inherent in nature.

15

While some uncertainty in oil and gas markets is caused by the
characteristics of the resource, the technology available to discover
deposits, and the rules of development, there is also uncertainty caused
by the time frame of development. Once a sale is made it may take years
(the state allows seven) to develop the lease and begin the recovery of
oil or gas. Once production has begun it may take many more years to
deplete the deposit. Purchasing decisions are made using some estimate
of the value of the resource and this includes the future value of any
oil or gas recovered. During the time frame required for development
prices may change considerably and this adds to the uncertainty inherent
in the purchasing decision.

Another important situational characteristic of oil and gas
resources is the the fluid nature of the deposits. The uncertainty
involved in the location of deposits from the surface creates problems
in the mapping of rights to underground deposits to the surface rights
in which leases are sold. If a deposit is found to lie under surface
areas for which the development rights are held by two or more parties,
the fluid nature of the resource can lead to problems in development. A
party who pumps at a faster rate than others can draw oil from under
adjacent claims. From an engineering standpoint, there is an optimal
rate at which oil or gas can be extracted from a well in order to maxi-
mize total recovery. Too great an extraction rate will significantly
reduce the total resource ~recoverable. The fluid nature of the
resource, in effect, creates non-exclusive rights to the deposits and,
coupled with the engineering realities, can lead to a form of the
tragedy of the commons discussed by Hardin (1977). The maximizing

actions of each developer on the field will result in a substantially

 

l6

reduced total recovery from the deposit.

The situational variables define the environment within which
structural alternatives must perform. The most important consideration
from the standpoint of this study will be the degree of uncertainty
faced by both parties to the transaction. Most states, as sellers of
rights to publicly held oil and gas resources, do not engage in any
activities to estimate the value of leases prior to sale, Louisiana is
the only exception. The seller, therefore, faces more uncertainty than
the buyer as to the value of the lease and relies on the method of sale,

that is institutional structure, to ensure a fair price.

Institutional Structure: Given the interdependence described by the

 

situation, the institutional structure defines the property rights of
individuals. It imposes some order on the interdependence. In the oil
and gas lease market, the structure of leasing institutions, the rules
that govern the leasing of rights to oil and gas resources, define the
property rights, either explicitly or implicitly, accruing to the inter-
ested parties.

The most common institutional arrangement for the allocation of
property rights to oil and gas resources is the auction. An auction is
essentially a price discovery mechanism used by the seller of an object
of unknown value and mineral lease sales, having the characteristic of
uncertain value, are often structured as such for this reason. Auc-
tions, however, can take many forms and these structural alternatives
may affect the outcome of the auction process. Auction forms are
distinguished by the way in which bids are solicited, the manner in

which bids are accepted, the object of the bidding, and the relationship

17

of the winning bid to the selling price. Besides the actual structure
of the auction there are terms specified in the lease, once purchased,
that further define the rights and responsibilities of the parties to
the agreement.

Ramsey (1980) makes a distinction between a competitive and a
discriminatory auction. A competitive auction is one in which the high
bid wins but pays the amount of the second highest bid. In a dis-
criminatory auction, on the other hand, the high bidder wins and pays
the amount of his bid. In some more recent studies (Milgrom and Weber,
1982) these auction types are described as first and second price auc-
tions which is the terminology that will be adopted in this study.
Besides the first and second price concepts auctions can be distin-
guished by the method of bidding, either oral or sealed. In a sealed
bid auction all prospective buyers submit sealed bids which are then
opened simultaneously and the winner announced. In this case either the
first or second price auction can be employed to determine the amount
paid by the winner. In an oral auction, on the other hand, bids are
accepted orally and bidders have the opportunity to alter bidding be-
havior in response to the actions of others. If second price auctions
are used in this case a bidder could win simply by bidding a very large
amount, consequently only the first price auction is appropriate in this
context. It should be noted, however, that the first price concept used
in an oral situation is similar to the second price sealed bid auction.
The high bidder in an oral auction knows the second highest bid and his
bid will only be slightly higher, there will be no reason to bid as high
as his real valuation if this is much higher than the second highest bid

and he will essentially win the lease at close to the second price. In a

18

sealed bid auction, however, the second highest bid is not known and a
bidder will bid his valuation of the object.

Another common distinction is between the English and the Dutch
auction methods. Bidding with the English method is progressive, start-
ing at a low level and gradually increasing the asking price until only
one bidder remains. English auctions are also referred to as oral,
ascending, progressive or open auctions. In a Dutch auction, also
called a descending auction, the selling price is first announced at a
high level and is progressively reduced until a bidder accepts the
stated price.1 The English auction is essentially an oral auction in
that bidders must be able to respond to the bids of their competitors,
there are, however, variations that protect the anonymity of the bidder
(Milgrom and Weber, 1982). Dutch auctions are oral in nature but do not
result in any interaction between the bidders as in the English auction.

It would be expected that the choice of auction method (structure)
would be motivated by a belief that the desired performance outcome will
result. A number of articles have addressed the question of the
relationship between auction method and some performance measures, most
notable perhaps being the work of Milgrom and Weber (1982) and Mead
(1967). This study will focus on three of the auction methods discussed
in these studies that are also under consideration for use in Michigan.
A distinction will be made between first and second price sealed bid
auctions and the English auction, hereafter referred to as the oral
auction method.

Auctions of oil and gas rights are also differentiated by the

 

A more complete discussion of the Dutch auction method can be found
in Milgrom and Weber (1982)

 

19

object of bidding, the most common being the bonus, royalty, profit
share and full share. These variables specify how the seller is to be
reimbursed for the object that is auctioned and, while only one is
usually the object of bidding, a combination can be used for actual
payment.

The bonus is an up front payment, made at the time of sale, for the
right to develop a specified surface area and is usually defined in
terms of a price per surface acre. The bonus payment bears no relation
to the actual amount of oil or gas that may be discovered under the
surface leased. The royalty is specified as a percentage of the
wellhead value of resources recovered under the lease and as such the
total payment will depend on the amount of the resource recovered as
well as its price. The royalty may be used as the object of bidding in
which case the bid is in terms of the percentage of wellhead value of
the resource that the buyer will pay to the seller upon recovery. Al-
though the bonus and royalty are often used separately, it is also
common in oil and gas leasing to charge both a bonus and a royalty with
the bonus being the object of the auction and the royalty set at a
specified level. Since the bonus and the royalty are used as partial
payment for the same object it would be expected that their ultimate
magnitudes would be inversely related. The higher the royalty rate
specified in the lease the lower the expected bonus bid, other factors
held equal.

The primary difference between the bonus and the royalty as objects
of the bidding are in the timing of payments and the sharing of risk.
Payments under a bonus scheme are made in full at the time of sale.

Royalty payments, on the other hand, are made only when and if actual

20

recovery takes place.

When the bonus is the exclusive method of payment, the buyer as-
sumes the risk of finding no recoverable deposits. Under this scheme
the seller is paid the same amount regardless of whether oil or gas 'are
discovered. The buyer, on the other hand, may pay the bonus and find no
recoverable deposits, The buyer, under a royalty payment, bears less
"dry hole" risk. If deposits are not discovered the buyer has still
lost the cost of exploration but has paid nothing for the lease. A
royalty is often seen as a way to shift the dry hole risk to the seller
but seller risk is defined differently than buyer risk. The risk to the
seller with a royalty is that no payment will be received on an un-
productive lease where a bonus payment would yield revenues. The most
common case is a bonus bid with a specified royalty which, in effect,
shares the risk between the two parties, the actual share borne by each
determined by the relative size of the two payments.

The concepts of risk and the timing of payments would be expected
to be important determinants of auction behavior because of the risk
aversion and time preference of the parties to a lease agreement. This
will be discussed in greater detail in chapter four.

Two other methods of payment or objects of bidding are based on
profits and known as profit-share and full-share. In a profit share
agreement a bidder decides on a share of any positive profits made as a
result of development of the resource that will be relinquished to the
seller. The profit share can be the object of the bidding or a method
of payment in addition to a bonus when the bonus is the object of bid-
ding. A full-share payment scheme, on the other hand, does not restrict

the sellers liability for losses in development. In a full-share

II'I'Z ‘1' :-

 

21

agreement the seller is liable for a share of the losses if a tract
proves not to contain mineral deposits. In the profit-share case, the
seller bears none of the risk of dry holes.

As with the choice of auction method, the choice of payment may
also be expected to influence some aspects of performance. Robinson
(1984), Reece (1978b) and Leland (1978) explore the effects of alterna-
tive payment structures on such measures of performance as revenues to
the seller and expected price.

Besides the method and object of bidding there are several other
factors that can distinguish one auction process from another. In some
auction markets a minimum number of bidders may be specified; that is,
an object will not be sold unless a certain number of bids is entered.
The seller may also specify a minimum acceptable price for the object to
be auctioned. This is referred to as a reservation price and represents
the value of the object to the seller. Robinson (1984) investigates the
effect of the choice of a reservation price on some aspects of auction
performance.

In addition to the actual auction process, the institutional struc-
ture of the sale is further defined by specific terms set forth in the
lease. These terms may specify regulations applying to development of a
lease or additional payments that must be made to the seller. Many
leases require an annual rental payment in addition to the bonus or
royalty, this is usually a charge per surface acre per year over the
specified term. The term, or the amount of time for which the lease is
in effect, is also specified as part of the conditions of sale. In
addition provisions are usually made for the assignment of a lease to

another party and for extensions of the initial term under certain

 

22

conditions. Finally a lease will specify the activities that can take
place during the exploration and development process on the site. Dif-
ferences in specific terms may also affect the performance of the auc-
tion system in terms of how the resource is developed, who obtains the
rights and how the proceeds are distributed between the buyers and the

seller.

Participant Conduct: Individual and group behavior provide the link

 

between institutional structure and market performance. Knowledge of
how economic actors will behave when confronted with the incentives and
penalties dictated by institutional structure is essential to the ac-
curate prediction of performance. The behavioral models of economics
generally assume that individual behavior is both purposive and ra-
tional, that is that individuals act with some purpose and that they do
so in a rational manner. These assumptions imply that the economically
rational man chooses so as to maximize an objective function such as
profit or utility. Furthermore these models assume perfect information
and the cognitive abilities to determine the maximizing, or rational,
course of action.2

Markets in oil and gas leases are characterized by uncertainty
which makes rational behavior, in the sense described above, impossible.
If accurate predictions of performance in these markets are to be made,
an understanding of behavior with less than perfect information and

limited cognitive resources must be considered. The expected utility

 

For a discussion of the neoclassical theory of economic behavior see

a good text in microeconomic theory such as Deaton and Muellbauer
(1980).

23

hypothesis (EUH) has been used to create a model of rational decision
making under uncertainty that has gained wide acceptance in economics
and has been applied frequently to oil and gas markets. Simon (1978)
and others have proposed models of the decision making process when
faced with uncertainty and limited cognitive abilities. Both of these
types of models will be considered in this study.

Besides the uncertainty faced by participants in the oil and gas
market there is another violation of the assumptions of the perfectly
competitive model that bears consideration in this study. It is
generally assumed that markets are competitive in nature and that the
actions of one player do not affect those of another. Because of the
limited number of participants in the market for oil and gas leases and
the heterogeneity of the leases these markets are often seen as oligop-
sonys, that is a market with a small number of buyers. In this case the
behavior of other bidders must be accounted for when deciding on a
course of action, decisions are not independent. Game theoretic models
have been developed for such situations and have been widely applied to
oil and gas markets.

Theories that describe both bidding and development behavior will
be important to the determination of performance in oil and gas leasing
markets. Because of the importance and the complexity of the behavioral
component of the SSCP paradigm for this study it has only been intro-

duced here and will be developed in more detail in chapter three.

Leasing Performance: The ultimate goal in the design of leasing institu-
tions is the attainment of program objectives, the performance of the

program. Performance has many aspects, however, and it must be decided

24

how it is to be measured so that real changes can be observed and
hypotheses tested. Performance can be measured in terms of normative
concepts such as economic efficiency or in the more substantive terms of
who gets what. The focus of this study will be on substantive perfor-
mance. Efficiency in the context of the economics of production and
consumption implies equality between marginal conditions and price
ratios for input allocation, product mix and consumption.3 Efficiency so
defined depends on the status quo and says little about substantive
performance.

The substantial literature on auction theory deals with performance
primarily from the standpoint of revenue to the seller and the capture
of economic rent. The distribution of leases 'among buyers, however,
has generally received very little attention in this literature. Ques-
tions of distribution do not address final impacts but are related to
larger issues of performance that may arise from different distribu-
tional patterns. It is conceivable, for instance, that the way in which
leases are allocated among different types of buyers will affect such
variables as the development of the lease or employment in the region of
development. Furthermore the fact that these are public rights being
transferred may lend added importance to the distributional issues.

Public ownership may be used to justify certain distributional goals,
for instance ensuring that Michigan based developers receive a specified
share of the offered leases.

From the seller's perspective a relevant measure of the performance

of auction alternatives is the capture of the economic rent associated

 

3 For a good discussion of the conditions of efficiency applied to
resource economics see Randall (1981), Chapter 6.

 

 

 

25

with the oil and gas resource. Economic rent is defined as the surplus
paid to a factor of production exceeding the minimum amount necessary to
call forth its services. This surplus results from supply inelas-
ticity. If supply were elastic surpluses would bring about increases in
supply until demand and supply were equated. Land, or exhaustable
natural resources, are prime examples of perfectly inelastic goods. In
the literature, economic rent is often discussed in terms of land rents.

The concept of economic rent can also be applied to oil and gas
resources and would consist of any revenues from the lease that exceeded
the amount necessary to produce and market the oil. Costs of explora-
tion, development, risk bearing and a return to the developer on in-
vested capital would be subtracted from revenues to determine rent.

Howe (1979) explains economic rent in oil and gas markets as aris-
ing from differences in recovery costs of deposits. Demand may be
sufficient to support the development of deposits with varying quality
or recovery costs. With certainty, the least expensive deposits will be
the first to be developed with more expensive ones following. In this
model the price of oil is the cost of recovery from the marginal
deposit, that is the one with the highest recovery cost. Developers of
intra-marginal deposits earn economic rent because development costs are
below price.

Perhaps a more significant source of rents in oil and gas markets
arises from uncertainty. When deposit existance, size, quality and
recovery costs are unknown a developer may pay less for a lease than it
is ultimately worth and hence earn positive economic rent. If lease
value were known, competition among buyers would eliminate rent on each

lease. The problem in institutional design for leasing under conditions

 

 

 

26

of uncertainty is to extract the economic rent in spite of the uncer-
tainty as to its magnitude.

The emphasis of this study is on performance in terms of the cap-
ture of economic rent by the seller and the possible effects of the
distribution of leases among different types of buyers on other objec-
tives of the state leasing program. Some attention will also be

directed to the impacts on development of various leasing institutions.

Specifics of the Leasing Procedure in Michigan

The SSCP paradigm has been described in terms of the oil and gas
lease market in general but the issues discussed are relevant to the
more specific problem of the leasing of Michigan state lands for
hydrocarbon development. The market in the State of Michigan is unique
in a number of ways and these specifics need to be understood in the
context of the paradigm used in this study.

Oil and gas in Michigan are found in relatively small deposits,
effectively increasing the difficulty of locating and developing the
resource. The size of a deposit will, to a large degree, determine the
revenues generated from a producing well. Smaller deposits generally
increase the risk faced by developers and imply that successful wells
will yield a smaller profit than would be the case when deposit size is
larger.

Oil and gas have been produced in Michigan for almost 100 years and
have been part of a significant industry in the state since 1925,
(Patric and Kakela, 1982b). The information in Table 2-1 illustrates

the historic development of the oil and gas industry in the state for

27

Table 2-1

Oil and Gas Production on State Lands

Oil Production Gas Production Acreage Under

Years (barrels) (1,000) cubic ft. Lease
1927-1931 555,101 ---------------------
1932-1936 1,665,578 220,620 27,120
1937-1941 9,810,671 732,423 160,473
1942-1946 8,270,928 2,315,755 291,644
1947-1951 5,277,096 5,085,920 446,021
1952-1956 7,272,536 5,161,244 933,531
1957-1961 4,543,205 8,354,698 264,725
1962-1966 3,146,716 6,974,406 438,539
1967-1971 5,337,231 6,497,155 284,651
1972-1976 21,324,223 91,132,019 1,563,465
1977 8,980,542 33,329,183 1,643,488
1978 7,936,277 41,310,258 1,535,088
1979 9,228,560 37,898,681 1,666,320
1980-10/1 1,185,380 29,786,521 1,584,850
10/1/80-10/1/81 8,229,847 39,188,686 1,707,986
10/1/81-10/1/82 8,450,258 32,296,848 1,888,286
10/1/82-10/1/83 8,901,078 18,298,733 1,841,761
10/1/83-10/1/84 8,488,415 39,343,217 1,693,189
Totals 128,674,097 397,926,372
Source: Unpublished figures collected by
Michigan Department of Natural Resources
state lands. Referring to the map of Figure 2-2, the first phase of

development was primarily in the Michigan Basin which covers the center

of the lower peninsula but in recent years the primary producing area
has become the Salina Niagaran trend which in 1980 yielded 73% of the
state’s oil and 85% of the natural gas. The Salina Niagaran trend,
which runs northeast across the northern lower peninsula, is thought to
have reached its production peak and some interest has shifted to the

application of new extraction technologies in previously abandoned

28

fields. In 1980, however, a large natural gas find was located in what
is known as the Mid-Michigan Gravity High which is a formation at almost
the 11,000 foot depth that runs from south of Detroit to the Traverse
City area. This find generated a great deal of interest and prompted
speculation about a major resurgence in the industry as the "deep play"
was explored. Drilling wells to this depth, the current average is
around 5,000 feet, is very costly an d if this signals a new stage in
exploration and development the cost of discovery and extraction in the
state could rise significantly. Figure 2-3 illustrates the location of
oil and gas fields relative to land under state jurisdiction.

There are currently approximately 491 companies or individuals that
are active producers of oil or gas in Michigan although an average lease
sale draws thirty to fifty participants. Developers range from private
individuals and small local firms to representatives of the major oil
companies. In a recent sale held in December of 1984, for instance,
there were 75 registered bidders 55 of which were from Michigan. Also
among the bidders were representatives of major out of state oil com-
panies such as Shell, Amoco and Atlantic Richfield. It might be ex-
pected that the ability to obtain a lease would differ among these
various types of bidders and leasing institutions will need to account
for this if allocation is important to the performance of the leasing

program.

The Auction System in Michigan: The State of Michigan currently conducts
two or three sales of oil and gas leases annually with the tracts of-
fered chosen from lands nominated by the industry. At the present there
exists a large backlog of nominations that were not offered at the time

they were nominated. Some of these backlogged tracts may be so old that

Figure 2-1

Oil and Gas Producing Areas in Michigan

 

 

 

 

 

Salina

Niagaran
Reef

    

29

30

Figure 2-2

State Lands and Oil and Gas Fields --
Michigan's Lower Peninsula

IIlOil Fields
c3 Gas Fields
[:3 State Lands

 

31

there is no longer development interest in them and the previously
interested parties may have nominated additional land. The offering of
leases for which there is no longer interest increases the number of
leases offered at a sale and results in many tracts receiving no bids.
When the backlog is worked through, it would be expected that the number
of acres offered would be reduced and the percentage of tracts receiving
bids in a sale will increase.

In order to avoid, or at least decrease, the problem of several
developers drilling the same structure and the attendent common property
related problems the state requires that drilling take place only on
drilling units of specified size. Drilling units are usually 40 or 80
acres but may be as large as 640 acres (Patric and Kakela, 1982c). The
size of drilling unit required depends on the geological conditions in
the area leased.

Lands nominated by the oil and gas industry are reviewed by the DNR
in order to determine whether they will be offered for sale and if so
under what conditions. The state distinguishes between development and
non-development leases, the former are those that allow exploration and
development activities on the surface of the land area covered by the
lease. Non-development leases, on the other hand, restrict these ac-
tivities because of environmental or other concerns. Leases are class-
ified as non-development if they fall within parks, state forest
campgrounds, natural or wilderness areas or areas with unusual natural
or recreational values where these would be disturbed by oil and gas
development. Non-development leases are often purchased to fill out a
drilling unit or because they can be developed by drilling at an angle
from adjacent lands. The Commission can be petitioned to change the

classification of a non-development lease and, based on a field review

32

and information provided by the applicant, may reclassify the lease if
it is determined that the non-development classification is no longer
applicable.

Once the DNR has determined which lands will be offered for lease,
the proposals are published and interested parties will be allowed to
contest leases and classifications before the Natural Resources Commis-
sion. The Commission reviews any contentions of the proposed class-
ifications and may require that they be changed. A list of the tracts
finally determined to be leasable is then published prior to the sale so
that parties interested in acquiring a lease are aware of the offering
at the sale.

Lease sales in Michigan are held by oral English auction. That is,
bids are elicited orally in an ascending fashion and the winning bidder
pays the amount of his bid. The object of the bidding is a bonus with
the royalty specified in the terms of the lease. Professional auc-
tioneers are hired by the state and, starting at the reservation price
of ten dollars per acre, they elicit progressively higher bids until
only one bidder remains. If there are no bidders willing to offer the
reservation price the tract is not sold.

The ten dollar reservation price represents a recent change in
leasing policy. Until 1983 the state had a reservation price of two
dollars per acre with a minimum of three bidders required in order for a
lease to be sold. It was found that bidders would immediately offer two
and then three dollars leaving a third bidder to decide, at a price of
four dollars, whether to continue the bidding. This amounted, in prac-
tice, to a reservation price of four dollars and no restriction on the
minimum number of bidders. The minimum was instituted because it was

thought that it would ensure competition for the lease but in practice

33

it did nothing to affect competitive behavior. Because of the ineffec-
tiveness of this policy, it was changed to the current reservation price
with no minimum bidder requirement.

The Commission reserves the right to refuse any bid and states a
number of reasons for doing so. Bids may be refused because a bidder
has failed to maintain good credit by failing to pay the total amount of
the bid at the time of sale, is in violation of rules or terms on a
previously purchased lease or for any other stated reason. The Commis-
sion also has the power to authorize alternative auction methods.

In addition to auctions, the Commission is authorized to enter into
directly negotiated sales under certain conditions and for certain
lands. Direct sales are the result of negotiation between the state and
a buyer and do not require the presence of other prospective buyers.
Direct sales are authorized in order to complete a drilling unit, on
lands previously offered but not leased at two previous lease sales
within a one year period, or on lands on which it is suspected that
drainage is occurring. Direct leases may not provide for less than the
standard royalty or rental rates and may not have a term longer than
those offered at public sales.

The lease specifies further conditions of the sale and reimburse-
ment to the state for oil and gas recovered. The lease currently in use
in Michigan specifies a term of seven years during which oil and gas
must be found if the lease is not to revert to the state. If oil and
gas are found within the seven year period the lease extends for as long
as oil or gas are recovered in paying quantities. The lease specifies a
royalty rate of 16.66% (1/6) of the wellhead value of oil or gas
produced. Beginning in the third year of the lease term a rental of two

dollars per acre will be required. Once a producing well of oil or gas

34

is established the rental is no longer required.

In order to ensure compliance with these terms the state requires a
performance bond of each leasee, the amount of which is determined by a
schedule. Subject to the approval of the Commission, any lease may be
assigned to another person.

The term of the lease may be extended beyond the initial seven year
period provided that the extension is determined by the Commission to be
in the best interests of the state. Extensions may be for a period of
one year only and if additional extension is requested it must be on an
annual basis. Rental rates on extensions will be three dollars for the

first year and increase by one dollar for every additional year.

Michigan Leasing Objectives: The specific objectives of the leasing
program are stated as being, (1) optimum economic return to the state,
(2) competition for the acquisition for leases, (3) protection against
drainage of hydrocarbons, and (4) protection of the environmental,
recreational and other uses of the land. These stated objectives are
too vague to be meaningful measures of performance and, it seems, con-
fuse intermediate objectives with ultimate goals in some cases. Discus-
sions with managers of the state leasing program have refined these
objectives and turned up other concerns that will be reflected in the
measures of performance used in this study.

The concern with an optimum economic return to the state is
primarily aimed at the concept of economic rent or a "fair" price for
the lease. Auction models generally imply that competition is a neces-
sary condition for the capture of the full economic rent and the con-
cerns of the DNR in this area again seem motivated by the concept of a

fair price. One component of performance that will be used in this

35

study will be the capture of economic rent. Due to difficulty in the
measurement of economic rent, the analysis of this component of perfor-
mance will rely on theoretical models in describing the link between
institutional structure and rent capture given the situation in
Michigan.

The orderly development of hydrocarbon resources is also listed as
a concern in the design of leasing policy in Michigan and, coupled with
a stated mission of stewardship of public resources, this seems to imply
a concern for non-wasteful development practices. Another area of focus
for this study is on the effects of alternative institutional structures
on the pattern of development of publicly owned oil resources.

Michigan relies very heavily on the automobile industry and during
the recent recession high unemployment in the state caused financial
hardship for the state and fostered concern for a broader industrial
base that would be less vulnerable to fluctuations in a single industry.
With the total value of oil and gas recovered in the state in 1980
listed at 1.5 billion and the level of drilling activity rising, oil and
gas development is big business in Michigan and is an important aspect
of the state economy (Patric and Kakela, 1982b). Oil and gas activities
brought around $494 million to the state in 1984 in the form of bonus
payments, royalties and severance taxes. Besides the revenue to the
state generated by oil and gas activity there are also employment oppor-
tunities to be considered. It is estimated that approximately 21,000
individuals in Michigan depend on the oil and gas industry for their

livelihood.4 It is conceivable that the allocation of leases between

 

"Michigan's Oil Exploration-Production Industry at a Glance",
Michigan's Oil and Gas News, May 24, 1985, pp. 24

36

different types of firms, local as opposed to national or international,
may affect the employment of Michigan residents in the industry and the
tax revenue and other benefits to the state. While this study will not
directly estimate the impacts of various ownership patterns on employ-
ment, state income or other factors it will describe some of the pos-
sible allocative consequences of alternative leasing institutions in the
event that these factors are found to contribute to state performance

objectives.

Summary

The SSCP paradigm provides a framework within which to explore how
rules, or institutional structure, affect individual choice and how
this, in turn, affects substantive performance. The choice of institu-
tional structure is the choice of how property rights are to be assigned
given the interdependence between individuals dictated by the charac-
teristics of the good and existing technology. Individual behavior in
response to the property rights, incentives and penalties imposed by the
structure of institutions dictates the performance of the system in
terms of who gets what.

Oil and gas lease markets are characterized by uncertainty as to
the value of resources. Because of this uncertainty rights to these
resources are allocated using an auction method. This study will inves-
tigate how various auction alternatives allocate property rights and
what sort of substantive performance results. Measures of performance
will be linked to state objectives in leasing oil and gas resources and
will focus on rent capture and distributional issues. Although the

allocation of leases among bidders is not a direct measure of

37

performance, it is possible that this distribution may be important to

some objectives of the state in leasing or may be politically desirable.

Chapter Three
Theories of Behavior Under Uncertainty

In economics assumptions are made in order to simplify analysis and
yield tractable mathematical models that economists often rely upon for
prediction.1 Assumptions, by necessity, abstract from reality and it
must be determined in each case whether the usual assumptions can be
applied without seriously affecting predictions. One of the common
assumptions in economic models is that economic actors possess perfect
knowledge. In some cases, however, the very fact that economic agents
do not know all suggests very different behavior from that implied by
the general model. If meaningful predictions are to be made in such
situations we must be able to describe how the presence of uncertainty
affects the behavior of the economic decision maker and how, in turn,
this affects the performance of the system.

The development of oil and gas resources is inherently a very risky
venture. The uncertainty involved is greater than in many other
economic activities and has been instrumental in shaping the institu-
tions that deal with oil and gas development. The objective of this
study is to describe the consequences of alternative auctioning institu-

tions and lease terms in a situation characterized by uncertainty.

Because of this uncertainty we can not rely on theories based on the

 

A more complete coverage of assumptions in economics can be found
in a microeconomic text such as Leftwich (1976).

38

39

assumption of perfect knowledge and must explore alternatives that can
explain behavior under conditions of uncertainty. The purpose of this
chapter is to present two theories of behavior under uncertainty that
will be useful in evaluating the performance of alternative structural
forms for the allocation of oil and gas resources.

There are two primary theories of behavior in the face of uncer-
tainty that merit discussion in the context of this study, the Expected
Utility Hypothesis (EUH) and what will be referred to as the Be-
havioralist Approach. These two theories make very different assump-
tions about the rationality or the cognitive abilities of economic
actors but they can both contribute to an understanding of how decisions

are made when uncertainty exists.

Risk and Uncertainty

Decision making under uncertainty is often seen as a problem of
deciding on a course of action when the consequences of the action are
unknown. The payoff, or consequence, of a particular course of action
will depend on the value of a random variable which is unknown at the
time of the decision. This random variable is known as the state of the
world and is defined to include all variables that will affect the value
of the payoff. Since this value depends on the value of a random vari-
able the payoff is itself a random variable. Formally stated, the
problem for the decision maker is to choose a course of action when the
payoff is an unknown random variable depending on the state of the

world. Of course the decision maker may have knowledge of the

40

probabilities of occurrence of the random variable and this fact can be
used to define two distinct cases.

Knight (1921) makes a distinction between the concepts of risk and
uncertainty, defining the latter as the case where the decision maker
has no knowledge of the probabilities of occurrence of the random vari-
able. In the case of risk, however, this knowledge is assumed. Knowledge
of the probability distribution for the state of the world implies
knowledge of the probability distribution of payoffs as long as the
state of the world is defined as above to include all elements of uncer-
tainty. The amount or type of knowledge possessed by the decision maker
would be expected to influence his behavior when faced with a situation
involving uncertain outcomes.

Ramsey (1980) dismisses the distinction made by Knight as one that
is irrelevant in most practical problems with the justification that
neither case is likely to occur. The most common case, according to
Ramsey, is one in which the decision maker has some knowledge of the
form of the probability distribution and can use this, along with infor-
mation from sampling, to estimate its unknown parameters. This view
emphasizes the importance of gathering information for the purpose of
estimation. A decision maker would also be expected to revise his
estimates over time as he learned more about the importance of various
sources of information and the suspected form of the probability dis-
tribution. In light of this discussion Ramsey proposes to use the terms
risk and uncertainty interchangeably which is how they will be used in

the remainder of this paper.

41

The Expected Utility Hypothesis

More is needed to determine behavior under uncertain conditions
than what is given above. The general model of uncertainty puts the
economic decision maker in the position of choosing a course of action
when the payoff is a random variable. Predicting behavior requires that
the decision rule be known. In standard economic theory, assuming cer-
tainty, the rational decision maker is a maximizer. That is, he makes
decisions so as to maximize an objective function such as profit or
utility. The decision rule is to make decisions consistent with this
objective. In the case of uncertainty, maximization of utility or
profit is no longer a feasible decision rule as it is not known what
action will produce the maximizing outcome. The problem facing
economists is to determine a decision rule for the case of uncertainty
that yields results consistent with observed behavior.

Historically economic actors were thought to maximize expected
income when faced with uncertainty as to the outcomes of their actions,
this expectation being the mathematical expectation over all possible
values of the random variable, the state of the world. This sort of
decision rule, however, led to predictions of human behavior that were
at odds with empirical observation. Perhaps the most famous contradic-
tion was the St. Petersburg paradox addressed by Bernoulli (1738). In
this example an individual is unwilling to pay a large sum to play a
game with an infinite expectation. In an attempt to explain this
paradox Bernoulli introduces the concept of utility and hypothesizes
that individuals make decisions based not on money but on the utility

that they can obtain from it.

42

"...in their theory mathematicians evaluate money in propor-

tion to its quantity while, in practice, people with common

sense evaluate money in proportion to the utility they can

obtain from it." (Bernoulli, 1738, pp. 33)
Bernoulli goes on to propose a decision rule stating that individuals
act so as to maximize the expectation of utility rather than income or
profit. Although Bernoulli recognizes that individuals will exhibit
different reactions towards risky prospects he falls back on a belief in
the diminishing marginal utility of money to support his claim. Accord-
ing to Bernoulli the reaction to a risky prospect will depend on the
initial wealth of the individual although no rational individual will
ever accept a fair gamble, this latter point being a consequence of a
belief in the concept of the diminishing marginal utility of money.

While Bernoulli's theory seemed to be a start towards an under-
standing of human behavior in risky situations it was not accepted by
many economists because it was not able to explain riskless choices. If
the concept of diminishing marginal utility of money is accepted, then a
decision rule of expected utility maximization restricts rational in-
dividuals to never accepting a fair game, an obvious contradiction to
observed behavior such as gambling. Von Neumann and Morgenstern (1947)
and Friedman and Savage (1948) among others have pointed out that the
concept of a cardinal utility measure such as that of diminishing mar-
ginal utility is not necessary to explain riskless choices. With an
ordinal concept of utility, a decision rule of expected utility maxi-
mization can be consistent with behavior in riskless, as well as risky,
situations.

With diminishing marginal utility no longer necessary it is not

required to restrict study to the class of utility curves that Bernoulli

discussed, namely concave utility functions. Without this restriction

43

there are three cases that can be explored -- risk aversion, risk
neutrality and risk loving behavior. Individual reactions to risk can
be explained in terms of the shape of the individual utility curve
rather than initial wealth. Concave utility functions correspond to the
case discussed by Bernoulli and would indicate a risk averse individual,
one who would not accept a fair gamble. A risk averse individual would
have to be paid in order to accept a fair gamble, the amount determined
by the degree of risk aversion. To this person there is a cost as-
sociated with bearing risk. A risk loving individual, on the other
hand, gains some utility from risk and will accept an unfair gamble
without payment, again the gambles that he will accept are restricted by
the degree of risk loving. The third type of individual will accept a
fair gamble, and any that are better than fair. In other words, gambles
are evaluated on their mathematical expectation. The last two cases
describe convex and linear utility functions respectively. Arrow (1971)
and Pratt (1964) develOped a measure of risk aversion, that is essen-
tially a measure of the concavity of the utility function, that can be
used to measure the cost of risk to an individual with a given utility
function.

Very simply the Expected Utility Hypothesis can be stated as a
decision rule that has individuals making decisions so as to maximize
expected utility. Each person's behavior, given this theory, will
depend on the shape of their personal utility function. Bearing risk
imposes a cost on some individuals and increases utility for others

depending again on their personal utility function.

44

The Behavioralist Approach

The Expected Utility Hypothesis described above is one way that
economists deal with decision making under uncertainty but there are
some who would disagree with the basic behavioral assumptions implied by
this theory. Specifically the EUH assumes that individuals exhibit
"rational" behavior, that is that people will always choose correctly,
from all possible courses of action, the action that is best in accord-
ance with their objectives. One of the advantages of a decision rule of
rationality is that it serves to narrow the range of behavioral alterna-
tives that must be considered and yields tractable solutions to problems
of predicting behavior. If rationality is assumed, then behavior, given
the individuals' objectives, is determined entirely by the environment
in which an individual finds himself. There may, however, be other
factors that influence behavior that are ignored under a theory that
assumes rationality and the predictions of such a theory under such
conditions could be incorrect.

The behavioralist viewpoint is that individuals do not always act
rationally, in the sense implied by economic theory, because of cogni-
tive and informational limitations. Rational behavior requires that the
individual be able to use all the relevant information at his disposal
and to accurately evaluate, and be aware of, all possible courses of
action. The proponents of the behavioralist theory would state that
this is next to impossible in all but the simplest of situations. The
complexity of the environment in which we make decisions is such that we
can not evaluate our actions in a manner consistent with rational be-

havior. Simon (1957) proposes a theory of satisficing rather than

45

optimizing behavior as a decision rule that is more consistent with the
actual decision making process. Individuals act not in a rational
manner but in accordance with "bounded rationality", or what Simon
refers to as subjective, as opposed to objective, rationality. Subjec-
tive rationality is rational behavior given the cognitive and perceptual
limitations of the individual whereas objective rationality is rational
behavior as described in the standard economic theory.

The basic tenet of Simon's satisficing theory is that the com-
plexity of the environment necessitates methods of simplification for
the purpose of decision making. To better predict human behavior we
must have a better understanding of the manner in which individuals make
these simplifications. It might be expected that the simplification
mechanisms employed by the individual will depend on the persons' objec-
tives, their cognitive abilities and on their perceptions of the en-
vironment. The view of the satisficing theory is that behavior is
related to the concepts of perception and cognition rather than to a
universal decision rule of rationality.

Simon (1979) develops an outline of individual satisficing behavior
as follows. An individual faced with making a decision engages first in
a process of search for a course of action consistent with his objec-
tives. The first step in the simplification process is the use of
Standard Operating Procedures (SOP's) by which the individual limits the
portion of the problem space in which alternative courses of action are
sought. SOP's are developed over time through a learning process as
different areas of the problem space are found to yield more profitable
actions in certain situations. In this sense, SOP's are a product of

past experience and are subject to alteration as the individual more

46

clearly defines the situations in which different SOP's seem to result
in satisfactory outcomes. SOP's can also become obsolete as the en-
vironment in which they are used changes.

As stated above, SOP's serve to narrow the area of search for
alternative courses of action but rarely will this result in but one
choice. After reducing the set of alternatives the individual must still
search for a course of action. Search is not a costless process,
however, and the individual must decide when to terminate the search
process and proceed with a chosen course of action. The theory of
satisficing suggests that search proceeds until a course of action is
found that yields a satisfactory outcome, an optimal outcome being
unknown in the face of uncertainty and cognitive limitations. The
decision to terminate search is made not on the basis of a comparison of
marginal costs and benefits of continuation but on the discovery of a
choice that is satisfactory in the view of the decision maker.

While this presents a general model of choice it leaves many ques-
tions unanswered as to how SOP's are arrived at and what sort of
simplifications are employed by individuals in an attempt to reduce a
decision problem to manageable decision proportions. Tversky and Kah-
neman (1974) provide some interesting insights on how people may arrive
at systematically biased perceptions of probabilities. This paper
hypothesizes that individuals employ heuristics to estimate the prob-
abilities of occurrence of an event. Heuristics serve to reduce the
complexity of calculating estimates but the type of heuristic used can
lead to systematic bias in the estimates. The authors tested the
hypothesis in a laboratory setting and found that people did seem to

ignore some important information and rely on heuristics when

47

formulating probability estimates. This finding would seem to support a
belief that people do not act rationally because of the complexities
involved in such behavior.

The decision of whether to buy insurance is a well used example of
decision making under uncertainty. A person must decide whether to
insure against a loss before it has occurred with no assurance that it
ever will. The assumptions of the expected utility hypothesis have been
used to derive many predictions of how the rational person would act
when considering the purchase of insurance. Kunreuther and Slovic (1978)
argue that the empirical evidence of behavior is not consistent with the
predictions of utility analysis and we must look instead at the
psychological, economic and environmental factors that influence the
purchase of insurance. The authors make a study of the market for flood
insurance and reach some interesting conclusions concerning the percep-
tion of probabilities. They conclude from the study that people ignore
low probability events because of a "finite reservoir of concern". If
people concern themselves with all possibilities they would have too
much to worry about. It was also found that the way in which a prob-
ability was stated would make a difference in the way in which people
reacted to it. The bottom line of this study is that people do not act
rationally because of limited abilities to correctly perceive probabil-
ities or to be aware of all possible threats or even of the conditions
of insurance. Generally people do not make decisions from a position of
perfect knowledge nor do they act "rationally" in the objective sense

because they perceive probabilities incorrectly.

48

The Modeling of Auction Markets

Auctions are often used when the value of the object for sale is
unknown and hence the calculation of a bid involves making a decision
with incomplete information. Attempts to model the behavior of auction
participants have drawn from both utility analysis and from game theory
models to explain decision making under these conditions. Whether an
auction market is modeled using a game theoretic or a decision theoretic
utility analysis approach will depend on the nature of the interaction
between market participants and the structure of the market. Gilley and
Karels (1981) claim that the two methods, if correctly specified, yield
generally the same qualitative predictions of bidding behavior.

The decision theoretic approach assumes that bidders act so as to
maximize the expected value of utility defined over net worth. Pre-sale
utility for a bidder will be U(w) where U(.) is the bidder's utility
index and w is net worth. If a bidder wins a lease with value v and
pays a price of p then the level of utility after winning the lease will
be U(w+v-p). If a non-winning bid is entered the level of utility will
be the same as before the sale if it is assumed that there are no bid
preparation costs. The bidder chooses p, the bid price, so as to maxi-
mize the objective function, the expected value of the utility function
defined over net worth. The expectation of utility is taken with
respect to lease values, given the distribution of these values condi-
tional on the bidders information about the lease and the distribution
of maximum losing bids by competitors, which is a function of the number
of bidders. Information about lease values is assumed to be obtained

through independent random sampling. Each bidder is usually assumed to

49

have access to the same information and each bidder's valuation is
assumed drawn from the same distribution of lease values.

The game theoretic approach searches for a Nash equilibrium bidding
strategy in which a firm's bidding strategy maximizes expected profit
given the bidding strategies of other firms. It is characteristic of a
Nash equilibrium that no firm can increase profits by changing its
strategy unless at least one other firm does the same. Strategies in
this formulation of the problem are bid amounts.

As Ramsey (1980) points out, the game theoretic approach yields not
only the optimal decisions for each player but also the outcome of the
game. The decision theoretic approach, on the other hand, while it
derives optimal decisions for each player, says nothing about market
outcomes. The decision theoretic approach yields the form of the market
supply and demand curves to which a traditional equilibrium analysis can
then be applied.

The choice of a theoretical approach to modeling bidding behavior
will depend on the structure of the market and the interactions among
participants as well as the ease of application of the model. In oil
lease markets, and particularly offshore markets, the supply side is
dominated by the government as the sole supplier of leases while the
demand side is seen as a competitive market. For onshore leases in
Michigan a substantial number of leases are offered by the private
sector and there is probably more competition for leases than in the
offshore markets. If it can be reasonably assumed that the demand side
of the market is competitive, the actions of one actor do not affect
others and a decision theoretic approach is appropriate. If, on the

other hand, the interaction between actors is thought to be important a

50

game theoretic approach may be easier to apply.

Applications to Decision Making in Oil and Gas Leasing

Two very different theories of decision making under uncertain
conditions have been presented in this chapter, as well as an applica-
tion to a specific decision setting, auction markets. The following
chapter will contain a much more in depth review of current models of
auction behavior and will use these models to compare alternative in-
stitutional structures for the leasing of oil and gas rights.

While the two theories of behavior presented in this chapter are
very different in their assumptions about cognitive abilities and actual
decision processes, they both contain ideas that can be helpful in
understanding the effects of uncertainty on oil and gas lease sales.
When participating in a sale a prospective buyer must decide how much to
offer for a lease of unknown value. Decisions are made subject to the
information available at the time and the ability to process that infor-
mation. The type of information that a prospective buyer possesses may
be used to divide the decision making process into two distinct parts in
which different decision processes may be used.

Before the sale, interested parties nominate tracts that they are
interested in purchasing. This interest implies some level of informa-
tion concerning the relative value of the tract. At this stage tract
value information can be obtained from seismic tests that detect certain
geological structures beneath the surface or through observation of

adjacent tracts that may already have been leased and developed.

51

Exploration by seismic testing is far from an exact science but these
tests do provide some information about the prospects on a tract. With
experience firms have developed probability distributions for the exist-
ence of oil or gas given certain types of geological formations found by
seismic means. If a firm can determine, before the sale, the probabil-
ity of finding oil or gas then some sort of expectation of oil or gas
deposits can be formed. In preparation for the sale, a prospective
buyer would use this information along with other expectations, prices,
the firms attitude towards risk, etc., to arrive at a bid.2 It seems
reasonable to assume that decision making before the sale is "rational"
given the knowledge and cognitive limitations of the individuals.

Before the sale there is time to collect and interpret information
relevant to the purchasing decision and make the decision in as "ra-
tional" a manner as possible. There will always be limitations to the
knowledge available, however, and to the ability to assimilate all the
relevant information. Once the sale is underway the actual method used
to make the sale may serve to transmit information. In an oral sale
situation, for instance, a bidder will gain information about the number
and identity of competitors for a lease and their valuations of the
tract. This new information may lead to different assessments of cer-
tain leases and may alter the predetermined bidding behavior of some
bidders. Since this information is gained in the process of bidding on
a tract there is not time for careful calculation as to how this infor-
mation alters bidding strategies and it would be expected that decisions

made at this time would rely on SOP's and would not be as ”rational" as

 

2 Capen, Clapp and Campbell (1971) claim that the distribution of
tract value is approximately log-normal.

52

those made before the sale. While a decision made before the sale may
be more "rational", a split second decision made in the course of bid-
ding may be based on more information and may be a better decision
despite its spontaneity. A decision maker in this situation must decide
quickly on the value of information gained during bidding and how it
will affect his behavior in the sale. Such spontaneous decisions may
well be based on SOP's rather than being the result of the usual

economic rationality.

Summary

Two quite different theories of behavior in the face of uncertainty
are discussed in this chapter. The main difference between the two
theories is in the degree of rationality assumed in individual behavior.
While the expected utility hypothesis assumes rational behavior, the
satisficing approach regards objectively rational behavior as impossible
because of cognitive and informational limitations. Because of this
difference in assumptions the two theories focus on different factors as
the determinants of behavior. A utility analysis will be concerned with
how characteristics of the environment affect behavior, which is assumed
to be rational, while a satisficing analysis will focus more on the
actual process by which information is perceived and used to arrive at a
decision given limited cognitive abilities.

While Simon's theory that no decision making takes place under the
rationality assumptions of the EUH seems very reasonable there are
certain concepts from the traditional theory that will be useful in

determining bidding behavior and in evaluating performance. In

53

particular the concepts of the cost associated with hearing risk and the

dependence of this upon the level of risk aversion will be important.

Chapter Four

Bidding and Auction Theory

Starting with a seminal paper by Vickery (1961) a number of models
of auctions and bidding behavior have been presented in the literature.
For the most part these models have focused on describing bidder be-
havior in order to define an optimal auction structure from the seller's
point of view, usually in terms of maximizing expected revenues (Meyer-
son, 1983). These models vary considerably in their assumptions con-
cerning bidder behavior, knowledge and the interaction between par-
ticipants. Some of the models are quite simple while others incorporate
a variety of the complex variables that can influence auction outcomes.
The purpose of this chapter will is to review the current development of
the theory of auction markets so that it can later be applied to the oil
and gas leasing situation in the State of Michigan.

The writers on auction theory have yet to adopt a standard ter-
minology to describe alternative auction structures so in order to avoid
confusion the definitions employed in this paper will be presented here.
Ramsey (1980) makes a distinction between the competitive and dis-
criminatory auctions based on the relationship of the winning bid to the
selling price. In a competitive auction the high bid wins but pays the
amount of the second highest bid while in a discriminatory auction the
high bid determines the price paid by the high or winning bidder.

The definitions used by Ramsey cut across the traditional distinc-

tion between oral and sealed bid auctions, a division which he dismisses

54

55

as not theoretically useful. An oral auction is essentially a competi-
tive auction as the winner need not bid more than a small increment
above the second highest bid to win the object. A sealed bid auction,
on the other hand, may be either competitive or discriminatory depending
on the auction rules. The Dutch auction, used in The Netherlands for
the wholesale auction of flowers, uses a descending price counter until
the registered price is accepted by one bidder. While this is a dis-
criminatory auction, it is also referred to as an oral auction as is the
standard English auction which is competitive in nature. The distinc-
tion between oral and sealed bid auctions in this context is the ”real
time" element of the oral variety. During the course of an oral auction
a bidder has the opportunity to observe the bidding behavior of other
bidders and react to their bids while this is not possible in a sealed
bid auction.

While Ramsey dismisses the distinction between oral and sealed bid
auctions, it can be a useful distinction when some of the more complex
models are considered. Some more recent work, most notably Robinson
(1984) and Milgrom and Weber (1982), distinguishes between an oral, or
English, auction, which is necessarily a competitive auction; a Dutch
auction; and two forms of sealed bid auctions, the first and second
price types corresponding to discriminatory and competitive auctions
respectively. This study will follow the lead of the more recent work
and distinguish between English auctions, Dutch auctions, first-price
sealed-bid and second-price sealed-bid auctions with the following
definitions.

An English auction is a progressive auction in which bids are

openly solicited at increasing amounts until only one bidder remains who

56

is willing to pay the announced price. The English auction is competi-
tive in that the winning bidder pays the amount of the second highest
bid, or a small increment more. The Dutch auction is a descending
auction in which the seller announces a high price that is lowered until
a bidder claims the object at the stated price. The Dutch auction is
discriminatory in that the winner pays the amount of his bid, that being
the high, and only, bid. The first-price sealed-bid auction is one in
which sealed bids are collected and opened simultaneously with the
highest bidder winning and paying the amount of his bid. In the second-
price sealed-bid auction the highest bidder wins the object but pays the
amount of the second highest bid. The Dutch and first-price auctions
have been shown to be strategically equivalent while the English and
second-price auctions are also equivalent but not as strongly so

(Milgrom and Weber, 1982).

A Review of the Literature

There is a considerable literature on auctions and bidding theory,
the Stark and Rothkopf (1979) bibliography includes almost 500 articles
and many more have been published since. The current theory has been
developed for a wide variety of applications including the sale of
treasury bills, the letting of procurement contracts and the leasing of
mineral rights. The range of applications of the theory has led to a
number of distinct models that employ different assumptions and seek to
answer a variety of questions concerning auction markets. While these
models differ in a number of ways perhaps the most important is in their

assumptions concerning the knowledge each bidder possesses and the way

57

in which it is related to that of other bidders.

Much of the literature on auction theory has focused on the
revenues to the seller under alternative auction methods. The usual
procedure is to establish the model in terms of the characteristics 'of
the object being sold and the assumptions concerning the bidder's
knowledge and interaction. Bidders are then assumed to act rationally
within the confines of the model towards the end of maximizing expected
profits. In the game theoretic approach, the auction game is usually
modeled as a non-cooperative game in which a Nash equilibrium is sought.
Under this formulation, bidders choose strategies so that it is not
possible for any bidder to gain by changing strategies given the
strategies of the other players. A strategy so chosen is termed a
dominant strategy. Given the equilibrium strategies resulting from
rational bidder behavior, the revenues to the seller can then be calcu-
lated and compared across auction types.

Under the decision theoretic approach the profit maximizing bid of
the one strategic bidder can be calculated and the expected revenue to
the seller derived. The difference between this and the game theoretic
approach is that a market equilibrium is not found using decision
theoretic methods.

Engelbrecht-Wiggans (1980) describes a very general auction model
that is essentially a game with incomplete information. In this model
there is an underlying true state of nature that is defined by the
characteristics of the object for sale, the number of objects being
auctioned, the utility functions of the players, the number of strategic
players and the behavior of the non-strategic players. A strategic

player is one whose bidding strategy is not specified by the model while

58

the behavior of the non-strategic players is endogenous. The decision
theoretic approach to auction modeling corresponds to the case of one
strategic player where the bidding behavior of the non-strategic bidders
is incorporated into the true state of nature through the reservation
price. In the game theoretic approach, on the other hand, the number of
strategic players can be any number larger than one and is occasionally
modeled as random.

It is assumed that all players know the true states of nature that
are possible and the probability distribution over the states but do not
know which state will apply in any given situation. Each player may
receive information regarding the state of nature through the observa-
tion of some random variable. In the case of oil and gas leasing this
information may be the result of seismic tests or other sources of
knowledge about probable tract value.

Each player chooses a bidding strategy that defines how the ob-
served information will be used to formulate an actual bid. Strategies
are assumed to be chosen before any information is actually observed,
therefore the choice of strategy is independent of information. The
choice of a bid, using the strategy, will depend on the strategy assumed
to be employed by the other bidders. A common strategy of many models
is to determine the fraction of an objects value to enter as a bid.

Finally the model contains a payoff function, known to each player,
that determines to whom the object is sold and the price paid. The
payoff function will depend on the player's bids, the true value of the
object, the auction mechanism employed, the payment scheme and the
reservation price. There may be other factors involved in the payoff

function as well, including costs associated with bid preparation or

59

participation in the auction. Although the payoff function is assumed
to be known the actual payoff may remain uncertain if the true value of
the object is unknown.

The four components of the general theory presented here, players,
objects, payoff functions and strategies, are further broken down into
their component parts by Englebrecht-Wiggans (1980). The classification
of these components serves to define a taxonomy of auction models that
will be useful in the discussion of the various theories to follow. Some

of the variables that will be important to this study are as follows.

1) Players:
a.Number of strategic participants
b.Utility functions, Linear (known), Non-linear (known),
Random
2) Objects:
a.Number of objects, known or random
b.Information on object's value
c.Physical characteristics of objects
d.Type, divisible or indivisible
3) Payoff Function
a.Award Mechanism, first-price, second-price, other
b.Payment scheme, bonus, royalty, profit share, etc.
c.Reservation price
d.Other Transfers, participation costs, bid preparation
costs, etc.

The Independent Private Values Model

Much of the literature on the theory of auctions analyzes the sale
of a single object using the independent private values model. In this
model, it is assumed that a single indivisible object is to be sold to
one of a number of bidders. Each bidder is assumed to be risk neutral

and knows the value of the object to himself but not to the other

60

bidders, hence the term private values. The values placed on the object
by the bidders are assumed to be independently drawn from some con-
tinuous distribution. Furthermore bidders are assumed to behave com-
petitively, without collusion.

In terms of the independent private values model, the Dutch and
first-price auctions are strategically equivalent (Milgrom and Weber,
1982). This is intuitive in that in each auction the bidder must choose
at what level to make a bid without knowledge of the actions of other
bidders. In each case the bidder does not know at what level other
bidders will enter bids and must make his decision in a static environ-
ment. This strategic equivalence means that the strategies chosen in
the two models will be the same as will the expected revenue to the
seller.

Also in the context of the independent private values model the
English and second-price sealed-bid auctions are equivalent. In the
setting of the English auction, the bidder must choose a price at which
to stop bidding if the auction is still in progress. If the bidder
knows the value of the object to himself, he will bid until the price
reaches his valuation, a dominant strategy (Milgrom and Weber, 1982).

This strategy is independent of the valuations of the other bidders.

In the second-price sealed-bid auction, when the value of the object is
known, the dominant strategy is to submit a sealed bid equal to the
bidder's valuation of the object (Meyerson, 1983). In both auctions,
the winning bidder will be the one who values the object most highly and
he will pay the amount of the second highest bid. It is important to
note that this result requires that the bidders know the valuation of

the object to themselves, an assumption which is not necessary for the

61

equivalence of the Dutch and first-price sealed-bid auctions.

The dominant strategy in the independent private values model is
for the bidder to bid his valuation of the object. In this case the
bidder who places the highest value on the object will win it. The
result is that the allocation of prizes in the English and second-price
auctions is Pareto optimal. This result carries over to the Dutch and
first-price auctions as well if the model is symmetric. Another impor-
tant result of the independent private values model is that the expected
revenue to the seller of the object of the auction is the same regard-
less of which of the four auction methods is used (Milgrom and Weber,

1982).

The Common Values Model

The assumption of the independent private values model, that bid-
ders know their own valuation of the object of the auction, is not very
realistic in the context of the auction of such objects as mineral
rights. In such cases the value of the rights will depend on the actual
quantity of minerals found, quality, ease of recovery and prices
prevailing at the time of extraction and sale. Since these variables
are unknown at the time of the sale, the value of the rights is not
known with certainty. Given these uncertainties it is impossible for a
bidder to know the true value of the rights at the time of the sale. The
common values model assumes that the object of the auction has a common,
although unknown, value to all bidders where each bidder may have a
different estimate of this value. It is, of course, unlikely that an

object such as mineral rights will have the same value to all bidders.

62

The actual value of the rights to a bidder will depend on costs that
may vary among bidders as well as the possibly different utility func-
tions of the participants. The common value assumption is, however,
useful as a first approximation.

This model usually assumes that bidders make independent and un-
biased estimates of the value of the object of the auction. This means
that the valuation of one bidder is not related to the valuation of
others and the expected value of a bid is equal to the true value of the
object. Rothkopf (1969) shows that in a symmetric common values model,
that is one in which all bidders have equal costs and can be considered
identical, the winning bidder will be the one with the highest value
estimate. The actual amount bid by this bidder will depend on the
bidding strategy he chooses which is a function of the number of bid-
ders. In this case the bidding strategy is to bid a fraction of the
estimated value of the object. Wilson (1977) obtains the same
strategies with the additional result that the expected selling price
converges to the true value of the object as the number of bidders
becomes large. The results of these two papers are valid for the first-
price or Dutch auctions only.

When bidders are uncertain of their valuation of the object of the
auction information becomes an important consideration. Early models
made the assumption that each bidder's value estimate was based on a
sample independently drawn from a common distribution that was known to
all bidders. The resulting valuations were unbiased estimates of the
true value of the object, implying that some estimates were higher than
the actual value and some lower. In the symmetric common value model

the bidder with the highest valuation will make the highest bid and win

63

the object (Milgrom and Weber, 1982). This formulation led to bidding
strategies in which the winner was the bidder who had most overestimated
the value of the object and consequently paid too much, a problem
referred to as the "winner's curse" by Capen, Clapp and Campbell (1971)
who find that winning bidders often pay too much for mineral rights.

Wilson (1977) points out, however, that winning an auction is, in
itself, an informative event and failure to account for this source of
information yields the winner's curse result with bidders making nega-
tive expected profits. Reece (1978), in the context of a first-price
bonus bidding model, demonstrates that when this source of information
is accounted for bidders will revise their bidding strategies and
eliminate the winner's curse. Both Reece and Smith (1981) found that
the winner's curse phenomenon led to non-aggressive bidding behavior
where optimal bids fell as the number of bidders increased. Intui-
tively, if a bid wins against a large number of competitors it is more
likely that the object has been overvalued. These results were further
corroborated by an empirical test of offshore lease sales by Gilley and
Karels (1981). In contrast to the winner's curse predictions Reece
found bidders to make positive, and often large, profits over a wide
range of conditions. '

While the Dutch and first-price auctions are strategically equiv-
alent in the common values model, as they were in the private values
case, the English and second-price auctions are no longer so. In the
common values case the English auction generally leads to higher prices
(Milgrom and Weber, 1982). This result stems from the information about
other bidder's valuations that a bidder can gain in the course of an

oral auction. The extra information weakens the winner's curse and

64

leads to more aggressive bidding in the English auction. The structure
of the auction can enhance or impede the transmission of information
among bidders and this is the major reason why the distinction between
oral and sealed bidding is a useful one.

The common values model, unlike the private values model, allows
for statistical dependence among value estimates but does not provide
for different valuations due to different tastes or costs. When value
estimates are statistically dependent the second-price auction generates
a higher average price than does the first-price auction. Milgrom and
Weber (1982) rank the four auction methods in terms of expected prices
with the English auction first followed by the second-price and then the
Dutch and first-price auctions.

The above ordering of auction methods in terms of expected prices
depends on the assumption of competitive behavior among bidders. Mead
(1966) discovered evidence of collusive behavior in federal timber
auctions and analyzes the effects of collusion on auction choice (Mead,
1967). Game theory would indicate that collusive behavior needs to be
enforced, if a party to the collusive agreement can cheat and not be
punished it will be in his best interests to do so. Collusion in auc-
tion markets will tend to lower selling prices as bidders agree not to
bid against each other. The choice of auction method can facilitate the
policing of collusive behavior by the bidders and, if collusion is
present, can affect the level of prices received by the seller. In an
oral auction bidders can observe who is bidding and can identify, and
later punish, cheaters on a collusive agreement. In a sealed bid auc-

tion, on the other hand, the identity of bidders is not known and

65

collusive agreements will be more difficult to police.1 Given these
considerations a ranking of auction methods in terms of expected prices
can be very different than that found by Milgrom and Weber (1982). Mead
(1977) states that in the presence of collusion the auctions can be
ranked in the reverse order, that is first-price sealed-bid followed by

second-price and English auctions.
Alternative Auction Objects

The theory of auctions presented so far has been in terms of the
choice of auction method but there are other factors, as well, that can
influence auction outcomes. There are many methods by which payment for
the object of the auction can be made and these are usually associated
with the object of the auction. Some common auction objects and payment
schemes used in the auction of mineral leases are the bonus, royalty,
profit share and full share. There is a small, but growing, body of
theory dealing with the consequences of alternative auction objects and
payment schemes, most notable among the recent articles are Robinson
(1984), Reece (1979) and Leland (1978).

Reece (1979) analyzes the bonus, royalty and profit-share as bid-
ding objects, and as payment methods, in terms of a symmetric common
values model. The comparison is made on the basis of expected prices.
Using a decision theoretic approach he shows that the amount of rent

captured by the seller is, in part, a function of the uncertainty

 

1 This result assumes a single sale, with repeated sales including the

same bidders a cheater on a collusive agreement can be identified
once sale results are announced and can be "punished" at a later
date.

66

associated with the value information of the bidders. Regardless of the
level of uncertainty, however, he finds that the profit-share dominates
the royalty which, in turn is better than the bonus in terms of rent
capture. The comparative advantage of the profit—share and royalty is
found to be greater when the number of bidders is low and the uncer-
tainty high.

Leland (1978) addresses the choice of payment form in terms of the
"optimal" sharing of risk between the buyer and the seller of mineral
rights and the effects on production decisions. Excessive risk aver-
sion by development firms, when higher than the applicable social level,
imposes social costs as it affects firm decision making with respect to
a publicly owned resource. The optimal sharing of risk will depend on
the risk aversion of both the buyer and the seller of the rights, the
seller in this case being the government. The optimal sharing of risk
between the buyer and the seller depends on the relationship between the
measures of risk aversion of both parties. The choice of payment method
can be used to adjust the share of risk borne by either party towards an
optimal level. Oil and gas lease sales, however, typically involve a
number of potential buyers who most likely exhibit varying degrees of
risk aversion. Because of differences among buyers in their attitudes
towards risk an auction that is optimal from the point of risk sharing
for one buyer may provide an advantage or disadvantage in the sale for
other buyers. The choice of payment method therefore is likely to
benefit some buyers relative to others.

The payment forms dealt with in Leland's paper are the profit-share
and the royalty. The analysis concludes that uncertain costs can not be

adequately insured under a royalty payment. The firm must bear this

67

cost uncertainty. This non-optimal sharing of risk leads to prices that
are lower than optimal. Furthermore the royalty payment leads to a host
of inefficient production decisions by the firm.

The physical characteristics of oil and gas wells dictate that
production costs rise as the age of the well increases. In the context
of the discussion of economic rent as presented by Howe (1979), as
production costs rise the amount of rent falls. At the point that rent
becomes negative, that is when marginal extraction costs exceed marginal
revenues, the rational developer will cease production. A roylaty
payment acts as a tax, a deduction from marginal revenue. The larger
the amount of the royalty the greater will be the deduction from mar-
ginal revenue and the earlier abandonment will occur. A lower royalty,
or a form of payment that does not affect marginal revenues, will result
in greater recovery from each well.

The profit-share payment, on the other hand, tends to share risk
more effectively, leading to higher expected prices and more efficient
outcomes if the bidders are more risk averse than the sellers. Further-
more, the profit-share scheme does not distort production decisions as a
royalty does. The primary disadvantage of profit-share payments is the
relative difficulty of effectively and accurately monitoring them.
Profits are much more difficult to monitor and define than is the physi-
cal production of oil or gas that is necessary for a royalty payment
determination.

Robinson (1984) notes that the most common current practice in the
leasing of oil and gas rights is a bonus bid with a fixed royalty, a
practice which clearly contradicts the ordering of payment methods found

by Reece and Leland. To explain the use of the bonus and royalty in the

68

face of the current theory he introduces the concepts of monitoring
costs, offer comparability and landowner risk aversion.

As Leland mentions, but dismisses as relatively unimportant given
the magnitude of possible benefits, the cost of monitoring profit-share
or full-share arrangements may present problems. These plans require
that the seller keep track of output, prices, capital costs, operating
costs and overhead. As Robinson points out under a profit-share system
the developer has an incentive to shift as much cost as possible from
dry holes to producing wells since he is not reimbursed for losses on
dry holes. Furthermore the developer has less incentive to control
costs than would be the case with either a bonus or a royalty. These
problems are further exacerbated by a full-share arrangement. The
problem of monitoring the developer may be greater for a private
landholder than for a government body as the government could audit a
firm's financial records.

Another problem is that of comparing offers if bids differ in more
than one dimension. In the profit-share and full-share arrangements the
costs faced by the firm will affect the ultimate rent recovered by the
landowner. If firm's costs differ, perhaps because one is more effi-
cient than another, then it will be difficult to compare bids in terms
of profit shares. A high bid by an inefficient firm could result in
less revenue than a lower bid from a more efficient one. Operating
costs can also affect production decisions which, in turn, affect the
amount recovered under a royalty system of payment. If these costs
differ greatly among firms a bonus payment may be more desirable than a
royalty.

Finally the risk aversion of the landowner and the bidders may

69

influence the choice of payment. Leland's results depend, in part, on
the assumption that bidders are more risk averse than sellers. Robinson
disagrees with this assumption and finds it more likely that the bidder
is less risk averse. Bidders are able, through diversification and
stock offerings, to spread risk among many prospects and may be less
risk averse than small landowners who do not have these opportunities.
This argument loses some of its validity when the seller is a government
offering a large number of leases.

The analysis of Robinson, while it is not developed in a model,
suggests a ranking that is the reverse of that obtained in the earlier
results of Reece and Leland, that is bonus, royalty, profit-share and
full-share. While Robinson admits that these results are not conclusive
they do present some important variables for consideration in auction
theory and help to explain the current practice in mineral rights auc-
tions.

While the literature addresses the bonus versus royalty question in
terms of the optimal sharing of risk between the two parties to a lease
the concept of time preference is not introduced. Time preference
refers to the preference of an individual for consumption (money) in the
present as Opposed to the future. The time preference of an individual
may depend on a number of factors such as current wealth or expecta-
tions. The higher a persons time preference the higher current consump-
tion is valued relative to the future. The choice between a bonus or a
royalty affects the time of payment, either in the present or in the
future. A buyer of leases with a high time preference may prefer to
delay payment through a royalty bid relative to a buyer with a lower

time preference. Similarly, the time prefernce of the seller will affect

70

his preference for payment forms.

It seems clear from this discussion that the choice of an "optimal"
auction should consider the relative time preference of the parties
involved. As with risk aversion the choice of a payment method will
benefit some relative to others because of differences in time

preference.
Multiple Object Auctions

Most auction models consider the auction of one object while the
common practice in many auction situations, including oil and gas leases
is to auction a number of objects at a single sale. Very little work
has been done on the theory of multiple-object auctions but some results
are available.

Weber (1983) distinguishes among three types of multiple-object
auctions, the simultaneous-dependent auction, the simultaneous-
independent auction and the sequential auction. In the simultaneous-
dependent auction each bidder must take a single action after which the
objects are distributed and payments made. An example of this sort of
auction is the sale of U.S. treasury bills. A simultaneous-independent
auction, on the other hand, describes the case where the bidders must
act simultaneously in a number of auctions where the outcome of each
sale is independent of the outcomes of the others. The sealed bid
auctions of mineral rights are a good example of this sort of auction.
Finally the sequential auction is one in which the multiple objects are
sold sequentially with each sale concluded before another is begun. The

sale of mineral leases by English, oral, auction is an example of the

71

sequential type of auction.

Multiple-object auctions, like the single object type, can be
either discriminatory or competitive in nature. In the context of a
model in which each bidder wants only one of the items offered, the
discriminatory auction would award an object to each one of the highest
bidders, up to the number of items offered, and these bidders would pay
the amount of their bid. In a competitive setting each high bidder
would pay the amount of the highest rejected bid, this is also called a
uniform price auction.

As in the single object case for the independent private values
model Weber (1983) finds that the total expected revenue to the seller
is the same under either the discriminatory or the competitive auctions.
It must be remembered that this theory deals with the sale of identical
objects, a uniform price auction with non-identical objects, oil and gas
leases for instance, would not be expected to yield the same results.

The simultaneous-independent auction is the one that most closely
corresponds to the typical sealed-bid oil and gas rights auction. An
important consideration in this type of auction market is that bidders
risk winning more objects than they desire. Since the sales are simul-
taneous a bidder does not have the choice of submitting contingent bids,
that is bids with the condition that if a particular object is won then
other bids are withdrawn. This is a major complaint about sealed-bid
auctions of multiple objects and particularly in the presence of budget
constraints for the bidders. Rothkopf (1977) derives optimal bidding
strategies in this case where bidders have a constraint on exposure.

The off-shore auctions of oil and gas rights conducted by the

federal government have often exhibited substantial differences between

72

the winning bid and the second highest bid, this difference is referred
to as money left on the table. For example Capen, Clapp and Campbell
(1971) analyze the bids offered in a number of offshore sales and show
that the ratio of high to low bids is sometimes as high as 100 and is
often in the range of five to ten. An explanation for this has always
been the uncertainty involved in tract value estimation. Differences in
value estimates may result from bidders having access to different
information or from different interpretation of the same information.
Differences in valuation resulting from interpretation can be attributed
to inherent uncertainty. It is this uncertainty that accounts for the
log-normal distribution of estimated tract values cited by Capen, Clapp
and Campbell. By definition inherent uncertainty can not be reduced.
Differences in value estimates based on different data is, on the other
hand, the result of ignorance. Additional information gathering could
reduce this source of uncertainty.

While second-price auctions eliminate the problem of money left on
the table and weaken the winner's curse phenomenon there may be other
causes of the large differences in bids. Mead (1966) points out that in
the timber industry a firm may bid very aggressively in order to gain at
least one sale in an area that is necessary to the firm's survival. Oil
and gas developers must also obtain leases if they are to remain in
business. In a sequential auction it would be expected that firms would
enter a few high bids to ensure winning something and a number of smal-
ler "fishing" bids. Sequential auctions should be expected to increase
the variance of bids.

The sequential auction has the same properties as the other multi-

object auctions in the context of the independent private values model.

73

In these models a bidder did not have to consider how his actions might
affect the strategies of other bidders since their strategies were
independent of their perceptions of him. In the oil and gas lease
markets, however, it is realistic to assume that bidders have dependent
value estimates and the situation becomes much more complex. Bidders
can bid strategically by altering bids in one round in the hopes of
influencing the bids of competitors in subsequent rounds.

The complex situation that would occur in a common values model
with dependent value estimates and non-identical objects has yet to be

modeled and would provide a rich area for further research.

Asymmetric Information

A common assumption in many bidding models is that all bidders have
access to the same information concerning the value of the object of the
auction. Different valuations and bids, in such a case, are a result of
differences in interpretation of the available information. It seems
reasonable to assume that in most cases bidders do not have access to
the same information about lease values because of the cost of such
information or because of information monopolies discussed by Reece
(1978). Gathering information on oil and gas leases is expensive and it
would be expected that some firms would face fewer constraints in its
purchase than others. Furthermore, the results of seismic tests are
very sensitive to the actual procedure. Two firms could carry out tests
on the same tract and obtain very different results. Information monop-
olies may occur when a firm developes tracts already under lease. The

information gathered from these leases may yield significant information

74

regarding adjacent tracts that may be offered at a later sale. Since
this information is proprietary to the firm it may result in an informa-
tional advantage in the future sale.

Milgrom and Weber (1982) claim that a bidder's expected profits may
depend more on the privacy of his information than on its accuracy. In
the equilibrium of the first-price auction two bidders with the same
information will receive zero expected profits while a bidder with
private, although inferior, information may have positive expected
profits. For a first-price sealed-bid auction Hughart (1975)
demonstrates that a bidder with better information on tract value can
restrict the profits of a less informed bidder to zero while maintaining
positive profits for himself.

Both of these results depend on a first-price sealed-bid auction
that offers no possibility of a less informed bidder learning from the
bids of a more informed one. If the auction structure is changed to one
in which a bidder is able to revise his bids upward in response to the
bids of others then very different results are obtained. Assuming that
the less informed bidder knows the strategy of the more informed bidder
and that the less informed bidder is risk neutral Hughart (1975)
demonstrates that there is no strategy available to the informed bidder
by which he can profit from his superior information in an English or
second-price auction. The two assumptions of the model are crucial to
the result, if either are violated the more informed bidder is able to
make positive expected profits although they would probably not be as
high as in the case of the first-price auction.

The importance of information in bidding strategies and rent

75

capture by bidders raises questions about access to such information and
its effect on auction markets from the seller's point of view. Analysis
by Hughart (1975) and Reece (1978) suggests that in some cases, depend-
ing on the promise of the region in which the leases are located, it may
be in the best interests of the seller to explore the region and make
the information available to all bidders. Milgrom and Weber (1982)
further demonstrate that in an English or second-price auction the

seller benefits from complete disclosure of information.

Reservation Price

In many auction situations the seller will specify a reservation
price, that is a price below which the object will not be sold. The
seller sets a reservation price to assure that the object does not sell
for less than that amount, usually the seller's valuation of the object.
Clearly the seller suffers a loss if the price is set too low and the
object sells for less than his personal valuation. While the use of the
reservation price protects the seller from loss some recent studies
indicate that the choice of a reservation price may affect not only the
minimum price but the maximum as well.

The reservation price represents the lower bound on bids and, in an
oral auction, is the level at which the auctioneer begins the bidding.
With a reservation price set higher than the seller's valuation there is
a chance that the seller will retain an object that a bidder values more
highly. While this represents a risk of loss to the seller, studies by
Meyerson (1983), Robinson (1984) and Riley and Samuelson (1981) indicate

that it may be optimal, under certain conditions, to bear this risk.

76

Riley and Samuelson (1981) demonstrate that when bidder's valua-
tions of the object are independent the optimal reservation price should
be higher than the seller's valuation. While the seller risks some loss
of revenue by setting a high reservation price, he will be compensated
by higher revenue extracted in the case of a lone high bidder. Also in
the independent private values model the choice of optimal reservation
price does not depend on the number of potential bidders or the type of
auction.

In the common values model the expected value of the object to a
bidder will depend on the number of bidders and their value estimates
since the "winner's curse" must be considered. Robinson (1984) shows
that, in this model, the optimal reservation price depends on the ex-
pected value of the object to the winning bidder given the information
available to him. Since some auction methods convey more information
than others, the optimal reservation price depends on both the number of
bidders and the auction type. As the number of bidders increases the
expected high bid also increases and the potential gains from a high
reservation price are reduced while the potential losses are unaffected.
The optimal reservation price, therefore, depends negatively on the
number of bidders.

The results given so far depend on the assumptions of symmetry and
non-cooperative bidder behavior. Robinson (1984) demonstrates that in
both the independent and common values models the optimal reservation
price increases with an increase in the probability of a stable bidder
cartel. The logic behind this result is that collusion reduces effec-
tive competition and decreases the expected level of a lone high bid. In

the extreme case of perfect collusion, there is only one bidder and the

77

reservation price should be set quite high. Similar results would be
expected to apply in other cases of reduced effective competition such
as the case of informational asymmetry. In an auction where one bidder
is thought to have better information than others the optimal reserva-

tion price is higher than in the symmetric case (Robinson, 1984).

Risk Aversion

Most auction models consider only the case of risk neutral bidders,
a condition that is often implicitly met in a model with the assumption
of expected profit maximizing behavior. Remember that risk neutrality
implies a linear utility function. If utility is defined over wealth,
then maximization of profits is equivalent to utility maximization. If
an individual is risk averse, however, the utility function is not
linear and expected profits are discounted for the level of risk. In
this case the maximization of expected utility and expected profits do
not correspond.

Very few models consider the implications of risk aversion of
either bidders or sellers. It would be expected that a risk averse
bidder who faces uncertainty as to the value of an object, in an auction
would discount his bid as the uncertainty increased. If two bidders
placed the same value on the object the less risk averse bidder would
make the highest bid. To the extent that the choice of auction struc-
ture can affect the uncertainty associated with the value of the object
it could also influence the bidding behavior of bidders who are not risk
neutral.

Milgrom and Weber (1982) demonstrate that the first-price auction

78

leads to higher prices than the second price auction when bidders ex-
hibit risk aversion in the independent private values model. When risk
averse bidders are uncertain of the value of the object of the auction,
as in the common values model, additional information may affect their
bid. Unless some restrictions are placed on the form of the utility
function, however, it is not possible to determine whether expected
prices will rise or fall when additional information about the value of
the object is made available.

Milgrom and Weber show that if information is perfect, so that
uncertainty is completely resolved, or bidders exhibit constant absolute
risk aversion, expected prices will rise with the disclosure of informa-
tion. When the resolution of uncertainty is only partial expected
prices may either rise or fall. Reductions in expected prices would
only be possible when the range of possible wealth outcomes of the
auction is large2 or when unresolved uncertainty is great. In addition,
under the assumption of constant absolute risk aversion the expected
price under the English auction is at least as large as under the
second-price sealed-bid auction.

Holt (1980) uses an independent private values model of bidding for
contracts to show that the English and first-price sealed-bid auctions
yield the same prices. An important assumption of this model is the

risk neutrality of the the bidders. If bidders are assumed to be risk

 

2 Absolute risk aversion is defined by Arrow (1971) as -u"(.)/u'(.)
where u(.) is the utility function. A measure of absolute risk
aversion that is constant implies that an individual will have the
same risk premium regardless of assets. In other words the amount
by which the individual discounts a particular risk will not depend
on his assets. In the above case a bidder with non-constant ab-
solute risk aversion may still increase bid amounts if the wealth
outcomes of the auction are not too large.

79

averse then the first-price sealed-bid auction is shown to yield the
higher price. At first glance this result seems counterintuitive since
bidders may be able to reduce value uncertainty in the English auction
by observing the behavior of other bidders, a source of information that
is not available with the sealed-bid auction. When value uncertainty is
reduced the bids of risk averse bidders would be expected to rise. Holt,
however, does not seem to consider the value information that bidders
acquire but only information as to the number of competitors. The uncer-
tainty as to competition inherent in the sealed-bid auction induces more
aggressive bidding from risk averse bidders and increases the expected
prices of the sealed-bid auction.

Leland (1978) addresses the optimal degree of risk sharing between
the buyer and the seller of an uncertain prospect, in this case a
mineral lease, from the standpoint of social welfare. He shows that the
optimal sharing of risk will depend on the relative risk aversion of
both parties. The greater the degree of risk aversion of the lease
buyer relative to the seller, the more risk the seller should bear. As
mentioned previously the choice of a bonus, royalty, profit or full
share as a method of payment will affect the share of risk borne by the
buyer and the seller. Leland goes on to discuss payment schemes in
terms of risk sharing. Robinson (1984) uses these results to help
explain the common use of the bonus as a form of payment. If sellers of
mineral leases are more risk averse than buyers, as Robinson suspects,
then sellers will specify payment schemes that will shift the risk to

the buyer.

80

Summary

There is a great deal of diversity in models of auction markets but
the one that most closely fits the mineral rights leasing situation
seems to be the common values model. This model, in its simplest form,
assumes that the object of the auction has a common, but unknown, value
to all prospective buyers. Each bidder has an independent estimate of
the true value of the object drawn from a common continuous distribution
which implies that all bidders have access to the same information about
the value of the object. It is generally assumed that there is only one
object to be sold to a number of bidders who are risk neutral and behave
non-cooperatively.

When the value of the object of the auction is unknown the role of
information becomes an important consideration. In the private values
model information about value was unimportant since each bidder knew
with certainty his valuation and it was independent of that of other
bidders. In the common values model, however, information becomes more
important and, since valuations can be dependent, the valuations of
other bidders can yield useful information. When the value estimates of
other bidders were not accounted for in bidding strategies, winning
bidders often paid too much for leases and suffered the winner's curse.
The possibility of the winner's curse leads to non-aggressive bidding
and lower expected prices.

In the common values model, as in the private values case, the
first-price and Dutch auctions were shown to be strategically equiv-
alent. The second-price and English auctions no longer yield the same

expected prices, however, because of the possibility of information

81

exchange. An oral auction structure allows bidders to revise bidding
behavior in response to the bids of others. This provides for better
information and weakens the winner's curse leading to more aggressive
bidding and higher expected prices in the English auction. Judged in
terms of expected prices, the English auction is preferable to the other
forms in the common values model.

When bidders have access to different information, the choice of
auction method can also affect the outcome. It was shown that a bidder
with superior information could restrict a less informed bidder to zero
profits while gaining positive profits for himself. This advantage
breaks down, however, when an auction method is used that allows
response to the bids of others such as an English auction.

The advantages of the English auction method depend, in part, on
the assumption of non-cooperative behavior. In the presence of a stable
cartel among the bidders the English auction may lead to lower prices
than the first or second-price sealed-bid auctions. The availability of
information about bidders and bids in an oral auction facilitates the
policing of a collusive agreement and, in the presence of such behavior,
reduces effective competition and lowers the expected price.

While much of the literature has addressed the question of auction
type there has been some attention given to the method of payment. The
payment method, either bonus, royalty, profit share or full share,
serves to share risk between the buyer and the seller of the lease. In
terms of expected prices the profit share payment dominates the royalty
method which, in turn, is better than the bonus. The comparative ad-
vantage of the profit share and royalty is greatest when the number of

bidders is low or the uncertainty high. The choice of payment method

82

can also affect production decisions and this must be accounted for as
well.

The ranking of payment methods given above contradicts the empiri-
cal dominance of the bonus with a fixed royalty. This inconsistency is
explained by introducing the concepts of landowner risk aversion, offer
comparability and monitoring costs. The inclusion of these factors
suggests that the ranking could be reversed with the bonus being the
preferred choice.

Although few results are available when either buyers or sellers
are assumed to exhibit aversion to risk, it is expected that this will
influence the choice of auction and payment method. Alternative auction
and payment methods subject the parties to the sale to different types
and shares of risk and, given aversion to risk, this would affect be-
havior and possibly auction outcome. Optimal risk sharing is found to
depend on the level of risk aversion of the two parties but no clear
linkage to payment schemes or auction methods is made.

It is also suggested that a reservation price above the seller's
valuation may, in most circumstances, result in higher expected prices.
The optimal reservation price is shown to depend on the number of bid-
ders as well as the auction method in the common values model. As the
number of bidders increases the optimal reservation price falls while an
increase in the probability of a bidder cartel will increase the optimal
price. A high reservation price serves to protect the seller in the
case of inadequate competition.

The common case in mineral lease auctions is for a number of items
to be offered at the sale. While some work has been done with multiple-

object auctions the assumptions of such models are very restrictive.

83

These models generally assume that items are identical and value es-
timates are independent which does not correspond well to the mineral
rights leasing case. An important consideration in the choice of auc-
tion method in multiple-object sales is the effect of a budget con-
straint in simultaneous sales.

In using the results of these models to prescribe actions in actual
auction situations it is important to bear in mind the assumptions that
have been made. Actual auction markets will never conform fully to the
requirements of a particular model and the results of such a model
should not be expected to apply exactly to a real situation. While most
assumptions of these models are made explicitly the assumption of ra-
tional behavior by bidders is often overlooked. As seen in the work of
Simon, economically rational behavior may be impossible because of
cognitive limitations. Despite the restrictions of the models, however,
they can serve as a useful starting point for applied work. Where
actual conditions deviate from model assumptions it is often possible to
estimate the probable affect of these differences on theoretical predic-

tions and hence to provide some guide to actual market outcomes.

Chapter Five

Oil and Gas Leasing in Michigan

Chapter two of this paper introduced the SSCP paradigm and expanded
it in terms of the oil and gas leasing market in the State of Michigan.
The situation was presented both in terms of general characteristics of
mineral leasing markets and those specific to Michigan. Also in that
chapter structural alternatives in auction markets were discussed along
with some possible performance measures. Chapter three developed more
fully the conduct portion of the paradigm by presenting two very dif-
ferent theories of behavior under uncertainty, one based on an assump-
tion of rationality and the other on bounded rationality. Chapter four
provided a review of the literature on the theory of auctions with
special attention to oil and gas leasing markets.

Most of the literature is oriented towards an analysis of the
performance consequences of alternative auction structures. These
theoretical models, however, define performance quite narrowly, almost
exclusively in terms of expected revenues to the seller of the lease.
Likewise, bidder behavior was considered only in terms of rational
economic models, that is in terms of expected utility or profit maxi-
mization. While these simplifications are understandable in the
development of theory it is necessary to go beyond these models in
applied work. The purpose of this chapter is to adapt the results of
these models to the concerns of the DNR with the oil and gas leasing

program in the State of Michigan.

84

85

The analysis in the present chapter will expand on the theoretical
models by considering a wider range of performance measures and by
relaxing some of the assumptions to more closely coincide with the
actual leasing market in the state. The analysis will focus an specific
questions of auction structure that have been raised as concerns by the
Department of Natural Resources. Measures of performance used in this
chapter will also be defined in a somewhat broader context that includes
such variables as the distribution of leases among bidders and the
development consequences of structural alternatives. These performance
measures are also dictated to some extent by the concerns of the DNR
with the state's leasing program. This adaptation of auction theories
to the market in Michigan and the performance concerns of the DNR will
permit some conclusions to be drawn concerning the auction structure

appropriate for use in the state.

Auction Method

The State of Michigan currently uses an oral, English auction for
the sale of rights to publicly owned oil and gas resources. Some con-
cern has been raised within the DNR, however, that the oral auction
method does not capture the economic rent associated with a lease and
that a sealed bid auction should be adopted. The alternative to the
oral auction considered for use in the state has been the first-price
sealed-bid auction and several bidding experiments have been conducted
using this method.

Milgrom and Weber (1982) demonstrated, in the context of the common

values model, that the auction methods could be ranked in terms of

86

expected revenue with the English auction first followed by the second-
price and first-price or Dutch auction. The model used to derive these
results has an attractive feature when oil and gas lease auctions are
considered. One of the important assumptions of the model is that
bidder's valuations of the object are dependent, or "affiliated" in the
terminology of the model. In terms of oil and gas leasing models the
assumption of dependent value estimates is appealing from an intuitive
standpoint. Information concerning possible tract value comes from many
sources and it is reasonable to assume that different bidders will have
similar information sets. Valuations will not be identical because of
differences in information and interpretation.

The ranking of the English auction above other, sealed-bid methods,
in the Milgrom and Weber study was a direct result of the oral nature of
the sale. When bidders are uncertain of the value of the object they
can gain useful information by observing the bidding behavior of others
during the course of an oral auction. Extra information weakens the
winner's curse and leads to more aggressive bidding in the English
auction. More aggressive bidding leads, in turn, to higher expected
prices. The added information also implies a decrease in value uncer-
tainty which would lead risk averse bidders to enter higher bids.

While information about valuation can lead to more aggressive
bidding and higher prices it also introduces the possibility that some
bidders will act as "free riders". Mead (1967) discusses the pos-
sibility that some bidders will not invest in information but will enter
bids based entirely on others bidding behavior. The money saved on
geophysical tests and other information gathering techniques could be

used to outbid the competition. This would be expected to be more of a

87

problem in situations where exploration costs were very high such as in
the offshore markets.

The primary argument against oral auction methods also stems from
the resulting accessibility of information. Bidders gain not only
information about other's valuation of the lease but also discover who
is bidding and the amount of their bid. When this sort of information
is easily available collusive practices become more easily policed.
Collusion would tend to decrease effective competition and lower ex-
pected prices. In addition to outright collusion, where agreements are
made before the sale, the oral auction method allows the possibility of
implicit signaling during the sale. A bidder may signal, by his bidding
behavior, his determination to gain a particular lease and other bidders
may not compete on the implicit condition that they will receive similar
treatment on "their" lease. Discussions with both bidders and auc-
tioneers at Michigan's lease sales would indicate that there is no
perceived problem of collusive behavior in lease sales in this state.

Head (1967) presents an excellent discussion of how industry
characteristics can influence auction preferences. Resource and in-
dustry characteristics may dictate that a firm have access to a reliable
source of specific raw materials if it is to remain in business. Tim-
ber, for instance, can not be economically transported over great dis-
tances, furthermore mills represent large fixed investments that require
a steady supply of logs. The timber industry, states Mead, prefers the
oral auction method because they can more easily ensure frequent access
to sales within a specific geographical area. The opportunity to react
to the bidding behavior of competitors gives the bidder greater control

over auction outcomes for specific sales and provides a means to protect

 

88

investments in production facilities. In the sealed bid setting it is
difficult to guarantee that an object will be won and this causes plan-
ning problems for bidders who must obtain the resource.

Oil and gas, unlike timber, is easily transportable over long
distances and, argues Mead, the major investment in drilling rigs and
other equipment usually takes place after the sale. It is not as impor-
tant in the oil and gas industry to acquire a specific lease and Head
concludes that the arguments for an oral auction will not be as strong.
It is important to note that oil and gas development firms have capital
invested in drilling equipment and, in order to remain in business, must
discover new resources. While leases in specific geographical locations
are not often required it is still important to obtain some leases in
order to utilize capital investments and ensure a future supply of the
resource. This will be particularly true for small firms that do not
have a large number of tracts under development. There are also cases
where, in order to fill out a drilling unit, a firm must purchase a
specific lease. It seems that many of the arguments for an oral auction
preference apply to the oil and gas industry, and particularly a market
with a number of small firms.

The preference for oral auctions may be weakened somewhat by the
presence of alternative sources for the resource. Acquisition of the
resource in an auction will not be as critical if the opportunity exists
to purchase from private landowners or from other sources. In this case
a buyer will not be as dependent on the outcome of the auction for
resource supplies. In Michigan there is a large market in oil and gas
leases on private lands and this provides an alternative source of

resources for the oil and gas industry in the state.

89

In certain types of markets the sealed bid auction could reduce the
number of bidders competing for a given lease and reduce the selling
price arrived at in the auction. In an oral auction bidders have the
ability to respond immediately to bids offered by competitors and to
know at any given time during the auction how much money has been spent
and which leases acquired. If a desired lease is not acquired the
bidder can reallocate money to the purchase of another lease and be more
certain of winning a lease even if not the most desired. Since most
bidders do not win every lease that they bid on, bidders will par-
ticipate in the auction on more tracts than they actually purchase. In
a sealed bid auction, on the other hand, once bids are submitted there
is no chance to change bidding strategy in response to the number of
leases won or the behavior of others.

In a sealed bid situation a prospective buyer will be able to
submit bids on a limited number of tracts, limited by his budget. He
will not know until the sale is concluded which leases were won and how
much of the budget was actually spent. In an oral bid situation,
however, if a lease is not won the bidder can immediately use the un-
spent money to submit a bid on another tract. The effect of this
characteristic of sealed bid auctions will depend on the capital limita-
tions faced by the bidder. A small firm would be more constrained in
bidding than a larger one with a larger budget and would submit fewer
bids. If the number of leases won depends in part on the number of bids
submitted on different tracts, which is a reasonable assumption, then
limiting the number of different leases on which a buyer can submit a
bid reduces his chances of winning a lease. A buyer could, of course,

bid on any number of leases in a sealed bid auction, regardless of a

90

budget constraint and simply default an any unwanted leases so obtained.
The DNR would, however, take such behavior into account in evaluating
the buyers bids in future sales so such a strategy may not be in the
buyers best interest in the long run.

This analysis suggests that if a market exhibits a disparity in
size and capital levels among bidders then a sealed bid auction would
tend to yield an allocation of leases that was biased in favor of the
wealthier bidders who could submit a larger number of bids. The market
in Michigan consists of large national firms, smaller local firms,
speculators and some individuals or private investment clubs. There is
wide disparity among bidders in capital available for the purchase of
leases. In terms of an "equitable" allocation of available leases among
the prospective buyers in this market it would appear that an oral
auction would be preferred. Furthermore an oral auction will allow more
bids to be submitted and would be expected to raise the overall level of
competition and the level of the winning bids.

While the allocation of leases among bidders is not listed as a
specific performance measure by the state, it may be that this alloca-
tion will affect other measures of performance. It may be found that
some types of firms are more likely to develop a lease in accordance
with the state's leasing objectives than others. Some firms may provide
more employment for state residents than others or it may be politically
desirable to favor certain firms. If the allocation of leases among
different types of bidders is important to the performance of the leas-
ing program then the choice of auction method given the market in the
State of Michigan could have significant consequences.

It should be noted that the problems faced by bidders who are

91

subject to a budget constraint are caused less by the sealed bid method
than by the simultaneous nature of the multiple-object auction. It is
possible, although not common in practice, to conduct a sequential
sealed-bid auction in which sealed bids are accepted for each object and
opened before bids are accepted on the next item.

A final consideration in the choice of auction method is the ease
of application. An oral auction requires the physical assembly of all
prospective buyers and can take a substantial period of time. The oil
and gas lease sales in Michigan typically take between two and three
days to conduct. Simultaneous sealed bid auctions, on the other hand,
do not require the presence of the bidders and the opening of bids and
awarding of leases can be accomplished in much less time and at much
less expense.

In summary the English, or oral, auction method is shown theoreti-
cally to yield higher expected prices than the sealed bid auction when
bidder's valuations are dependent on each other. This result depends
on the reduction in value uncertainty in the oral auction as information
is gained from the bidding behavior of others. If bidders are risk
averse, which seems reasonable, the reduction in uncertainty will also
lead to higher bids because of lower risk discounting. The ease with
which information is transferred in the oral auction also leads to
concerns about free riders and collusive behavior. The free rider
problem will be most severe when the cost of exploration is high, as in
offshore leasing, and would not be expected to be as much of a concern
in Michigan. Collusive behavior is not thought to be much of a problem
in Michigan's lease sales but it is often difficult to detect and should

be watched for. Finally the sealed bid method, as most commonly used

92

for multiple-object auctions, can lead to a biased distribution of
leases among bidders subject to different budget constraints. To the
extent that this affects other leasing objectives it is an important
consideration in the design of auction structure. The large disparity
between the size and budgets of bidders in the state's lease sales would
indicate that this could be a significant factor in Michigan's leasing
program. When budget constraints are binding in sealed bid auctions an,
oral auction may result in increased competition and higher expected

prices as bidders are able to enter bids on a greater number of leases.

Bidding Objects and Payment Method

Michigan currently utilizes a bonus as the object of bidding with
an additional royalty payment of 16 2/3 percent specified in the lease.
There has been some discussion of adopting the royalty as a bidding
object in cases where a drainage lease is to be sold. Remember that a
drainage lease is one in which it is thought that hydrocarbon resources
are being drained from under state lands. In the case of a drainage
sale the uncertainty to the buyer is considerably reduced because
resources are being recovered on lands adjacent to the lease.

The choice of payment method determines, in part, the shares of
risk borne by the buyer and by the seller of a lease. When buyers and
sellers are risk averse their attitudes towards bearing risk will affect
the relative outcomes of the use of alternative payment forms. Payment
methods can also affect development of a lease and this may have conse-
quences for the performance of a particular auction market.

The theoretical results of Reece (1978) rank payment methods in

93

terms of expected revenue for the common values model with profit-share
first followed by the royalty and bonus methods. The ranking of Reece
is in inverse order of the value estimate to the bidder. This result is
intuitive for a risk averse bidder since he will pay more when risk is
reduced. The ranking given by Reece is in decreasing order of the
amount of risk borne by the buyer. Under a profit share arrangement the
landowner shares dry hole risks and cost uncertainty with the developer.
With a royalty payment the landowner shares only the dry hole risk and
with a bonus he bears neither.

Leland (1978), approaching the problem from the standpoint of the
optimal sharing of risk, obtains the same ranking as Reece under the
assumption that bidders are more risk averse than landowners. In
Leland's model the optimal share of risk to be borne by the landowner
increases as his level of risk aversion falls relative to the bidder.
The choice of an optimal payment method will depend on the relative risk
aversion of both parties to the lease.

Robinson (1984) introduces three explanations to account for the
empirical preference for the bonus payment over either a royalty or
profit share. These are landowner risk aversion, offer comparability
and monitoring costs. The profit share payment requires that sellers
keep track of costs, output, prices and overhead which may be very
difficult. Furthermore the buyer has an incentive to shift costs from
dry holes to producing wells and this may be difficult to detect. The
costs of accurately determining a profit share payment may .make it
unattractive for most landowners despite the possibility of greater
revenue. While monitoring may not be as difficult for a state govern-

ment as for a private landowner it may still entail considerable cost. A

94

royalty determination requires only a knowledge of output and prices
which will be more easily obtained than the information to determine a
profit share and this may account for the greater popularity of the
royalty, often in conjunction with a bonus.

When firm costs differ because of differences in efficiency or
production processes it may be difficult to compare offers from compet-
ing firms when a profit share or royalty is the object of the bidding.
The problem is greater with a profit share than a royalty since develop-
ment costs as well as production costs must be accounted for in payment
determination. With the royalty payment only costs of production, which
determine production rates and influence the abandonment decision, need
be considered.

Finally, if the landowner, in this case the State of Michigan, is
more risk averse than the bidder it may be Optimal from the standpoint
of the state to shift as much of the risk as possible to the bidder
through the use of a bonus payment. Robinson argues that it is likely
that landowners are more risk averse than bidders since development
firms can diversify their risk over a large number of leases. This
argument loses some of its validity when the seller is a government body
that sells thousands of leases and diversifies more in this respect than
most of the bidders. Government bodies, however, may reflect the risk
attitude of the public in decision making and may have other, internal
reasons for being risk averse.

Michigan specifically states as a criterion for the choice of
auction method that the development risk be borne by the buyer. While
this supports the current use of a bonus as the object of the bid it

contradicts the inclusion of the royalty as an additional payment. The

95

use of both a bonus and a royalty achieves a level of risk sharing that
is intermediate between either one used alone and the adjustment of the
royalty rate can further fine tune the risk sharing. The expected
amounts of the two forms of payment, when used together, will be in-
versely related.

In addition to risk aversion the relative time preference of the
buyer and the seller must be considered when deciding on a payment
method. If the seller has a higher time preference than the buyer, that
is has a greater preference for current consumption, a bonus may be the
prefered payment method from the sellers point of view. In a manner
analogous to the sharing of risk an "optimal" auction should consider
the relative time preference of the two parties.

In addition to affecting the sharing of risk between the two
parties to a lease sale the choice of payment method may affect some
aspects of lease development. The royalty payment acts as a tax on
production and will influence the abandonment decision, the higher the
royalty rate the earlier the well will be abandoned. The affect of the
royalty payment on the abandonment decision can be ameliorated through
the use of a sliding scale royalty. With this scheme the royalty rate
is tied to the amount produced. As a well ages and production drops the
royalty rate also falls and premature abandonment is avoided. The time
of abandonment will also affect the total revenue ultimately recovered
from a lease through royalty payments as it affects the physical output
from the well.

The expected price to the seller of a lease would be expected to
increase as the share of risk borne by the buyer falls if buyers are

risk averse. The seller, in choosing a payment method, must decide on

96

his own level of risk aversion, that is what amount of risk he is will-
ing to bear. Other considerations include the cost of monitoring any
payment method to insure that the proper payments are being made and the
ability to meaningfully compare bids when the payment method is the
object of bidding.

Discounting for risk involves applying a discount rate to the
expected revenue from a lease. When only a few leases are being
developed a firm may not be able to afford to play the averages and may
be risk averse. A government body, however, that sells thousands of
leases can diversify very effectively and would probably increase ex-
pected, and actual, revenues by increasing the amount of risk borne.

Given the difficulty in monitoring profit share payment methods, in-
creasing the royalty seems an effective way to accomplish this. When
using a royalty, however, a sliding scale should be considered in order
to avoid the negative effects on well abandonment and subsequent royalty
revenue.

In the case of drainage leases the uncertainty concerning tract
value is considerably reduced because of the presence of an adjacent
producing well. The same argument used above could be used to justify a
higher royalty rate on drainage leases. Furthermore the reduced level
of risk for such leases implies that the seller may benefit from
specifying a higher royalty rate on these tracts in order to bear the

same amount of risk as on non-drainage leases.

97

Reservation Prices

In many auction situations the seller will specify a reservation
price, a minimum price below which the object will not be sold. The
reservation price is generally thought to represent the seller's valua-
tion of the object although there may be strategic reasons for setting
it at other levels. The State of Michigan recently changed their reser-
vation price for oil and gas leases from two dollars with a minimum of
three bidders required to ten dollars with no requirement on bidder
numbers. The reservation price in this state is based on the belief
that competition was not sufficient to ensure a fair price rather than
any knowledge of possible opportunity costs of development.

The work of Robinson (1984) demonstrates, in the common values
model, that the seller's revenue will be increased if the reservation
price is set higher than the seller's valuation. In this model the
optimum reservation price will depend on the bidder's valuation of the
object. Since this valuation is a function of information and some
auction methods convey more information than others the reservation
price will depend, in part, on the auction type as well. Bidder's valua-
tions will also be affected by the number of bidders so this, as well,
enters the function that determines the optimal reservation price.

In setting a reservation price that is higher than his valuation
the seller risks not selling an object that a prospective buyer values
more highly. While this represents a loss to the seller it may be
recovered when a lone bidder places a value on the object that is higher
than the reservation price. When several bidders place a value on the

object that is greater than the reservation price competition among them

98

will result in a price higher than the minimum. When only one bidder
places a value at least as high as the reservation price on the object,
however, the seller must rely on the reservation price to increase the
bid.

As the number of bidders increases so does the expected level of
the high bid and the potential gains from a high reservation price are
reduced while the potential losses are unaffected. The optimal reserva-
tion price, therefore, is negatively related to the number of bidders.
In Michigan's lease sales many leases are sold with only a few, and in
many cases only one, bidder active. Few bidders on many sales would
indicate that a reservation price should be set that is substantially
higher than the state's valuation. While the level should be set higher
than the valuation, however, the theory does not indicate at what level.

Since bidder's valuations will be affected by auction type and the
optimal reservation price is a function of bidder's valuations, the
choice of a reservation price will depend on the type of auction. The
English, or oral, auction conveys more value information than other
auction types, weakens the winner's curse and leads to more aggressive
bidding and higher expected prices. The optimal reservation price in the
English auction would be expected to be lower than with the other auc-
tion methods.

In summary a reservation price set higher than the seller's valua-
tion of the object would be expected to yield the greatest results when
the number of bidders is small. For many of the leases sold in
Michigan's sales there are few bidders and a high reservation price
would be expected to increase revenues to the state. The actual level

for the reservation price would have to be determined by when losses

99

from a high price begin to exceed gains. Because of the effects of
auction method on bidder's valuations the optimal reservation price
should be higher under a sealed bid auction than with the English
method. The reservation price is seen to be effective in increasing
expected prices when bidder numbers are low and there is little effec-
tive competition. The use of the reservation price should also be
considered when effective competition is reduced by other means such as

collusive behavior.

Summary

The purpose of this chapter has been to suggest some performance
outcomes of alternative institutional structures in Michigan's oil and
gas leasing program. The analysis has applied the results of the
theoretical models of auctions and bidding behavior with special atten-
tion to the situational characteristics of the State of Michigan. It is
hoped that the results of this study will assist the DNR in designing
the state's auction market so that the desired performance can be ob-
tained.

Given the nature of the market for oil and gas leases in Michigan
the English auction has some advantages over the sealed bid methods.
First, the expected revenue from the English auction exceeds that from
the sealed bid auction when bidder's valuations are dependent, which is
a reasonable assumption in mineral lease markets. This is a result of
the additional value information provided in the English auction and the
comparative advantage will be even greater when bidders are risk averse.

Secondly, in a market characterized by a disparity in firm size and

lOO

bidding budgets, which is the case in Michigan, the English auction
eliminates much of the advantage of the larger, wealthier firms in
winning leases. The allocation of leases among bidders may not be a
performance end in itself but may have consequences for other perfor-
mance measures. The primary disadvantage of the English auction is that
collusive behavior, which can result in reduced revenue to the seller,
is difficult to control. Collusion is not seen as a’ problem in
Michigan's sales but is often difficult to detect and the possibility
should not be ignored. The English auction can also encourage free
riders but this would be a greater problem when exploration costs are
high or where the object of bidding makes up a large share of the total
payment, neither being the case in Michigan.

The choice of payment method determines the shares of risk borne by
the two parties to a lease. The optimal sharing of risk will depend on
the relative risk aversion of the buyer and seller. When the buyer is
more risk averse than the seller the largest expected revenues will
result from those payment methods that shift risk to the landowner, that
is the profit share and royalty as Opposed to the bonus. It seems
reasonable to assume, when the seller is a government agency that sells
many leases, that the risk aversion of the buyer exceeds that of the
seller.

The profit share payment has the disadvantage of being difficult to
monitor and while the royalty shares these problems they are not as
severe. When used as the object of bidding the profit share and royalty
methods suffer from a difficulty in comparing bids that is not a problem
with the bonus. Finally the royalty payment can affect development of a

lease through the abandonment decision and this may affect other

lOl

performance measures used by the state. In Michigan the bonus is the
object of bidding, avoiding the problem of offer comparability, while
the royalty, on a producing lease, makes up the majority of the actual
payment. Furthermore, when uncertainty is reduced, as in the case of
drainage sales, the royalty is increased. This behavior is consistent
with the goal of increasing revenue when the seller is less risk averse
than the buyer.

Finally the choice of an appropriate reservation price is important
to market performance in terms of expected revenue. The setting of a
reservation price above the seller's valuation in a mineral leasing
model would be expected to increase revenue to the seller. The optimal
reservation price will depend both on the number of bidders and the
auction type. The advantages of a high reservation price are most
pronounced when competition is least effective, this may be the result
of few bidders, different levels of information or collusive behavior.
In any of these cases a relatively high reservation price would be
expected to increase revenue. The optimal reservation price would be
expected to be smaller in the English auction than the sealed bid auc-
tions but is still greater than the seller's valuation. While the
theory states that reservation prices should be greater than the
seller's valuation the optimal level must be empirically determined.
Results of past sales in the state could be used to determine whether
the change in the reservation price has affected revenues.

In conclusion it seems as though the auction structure currently in
use in the State of Michigan is well tailored to the market situation to
obtain the desired performance. The English auction is expected to

generate a greater revenue than the sealed bid methods and yield a less

 

102

biased allocation of leases. These results are especially relevant in a
market, such as that in Michigan, where there is a large difference in
bidder size and wealth. A question that needs more attention in this
context is the possible performance consequences of the distribution. of
leases among different types of firms. The bonus as the object of
bidding eliminates the problems of offer comparability encountered with
a profit share or royalty while the royalty as an additional payment
shifts some of the risk to the state. Since it seems that the state
should be less risk averse than most of the buyers this is consistent
with theory. It may be desirable to explore the use of a sliding scale
royalty to avoid the impact of the payment method on lease development.
Finally the choice of a reservation price can affect revenue and some

effort in the empirical determination of an appropriate level could pay

off.

A Note on Empirical Work

The State of Michigan's lease sales and bidding experiments with
sealed bids have resulted in a considerable quantity of data. As in-
itialy conceived this study would have used these data to draw some
inferences about Michigan's leasing program. The data were, however,
found to be inadequate for this purpose. It may prove helpful for
future research to explore the deficiencies of the available data and to
discuss some possible methods of testing the hypotheses presented in
this paper.

The results of this study suggest that small firms may face con-

straints in sealed bidding relative to larger firms. In a market where

103

large and small firms compete with each other for leases these con-
straints would imply (l) greater competition for leases under the oral
bidding structure, (2) higher prices with oral bids and, (3) a different
distribution of leases among bidders under the two bidding structures.
Given the proper data these should be testable hypotheses concerning the
difference between oral and sealed bidding.

Empirical testing of the differences between oral and sealed bid-
ding may encounter several problems. First, because the oil and gas
market, and hence the situation under which leases are sold, can change
considerably from one sale to the next it would be desirable to collect
data on both oral and sealed bids at the same sale. Secondly, the
buyers should be familiar with both methods.

The DNR conducted sealed bid experiments in five lease sales. In
these experiments fifty leases were sold through sealed bids which were
opened at the start of the regular oral sale. Several attempts were
made to compare the results of the oral and sealed bids. First an
attempt was made to compare the mean prices of the sealed bid group with
the leases sold by oral bidding. Bid prices in both groups, however,
exhibited considerable variation and it was not possible to draw any
statistically significant conclusions from the data. Casual observation
suggests that tracts that are geographically close to each other have
similar sale prices. The second test compared the selling prices of
tracts sold by oral bids with adjacent tracts sold through sealed bids.
Again significant results were not obtained.

One possible solution to the difficulty in comparing oral and
sealed bidding in terms of the capture of rent may be to examine the

historical record of leases that have been sold and developed. Given

 

104

the selling price of a lease that has already been developed it may be
possible to determine whether the state has captured the economic rent
associated with the lease. This could be accomplished by comparing the
selling price with the eventual revenues. This method would be limited
to deteriming whether the state had captured the economic rent as-
sociated with a particular lease conditioned on the method of sale.
Unless different auction methods were used in the initial sale the
comparison of auction structures in terms of rent capture would not be
possible. The historical comparison method described above could also
be used to compare the bonus and royalty as methods of payment. To
yield useful results, however, there would have to be some variation in
actual methods of payment.

While this method may hold some promise it too has limitations.
Probably the greatest obstacle to implementation will be the data re-
quirements. In addition to original selling price, historical data on
development costs, production, and firm revenues will be required. Since
the development takes place over time it will also be necessary to
discount revenues and expenditures and the question of an appropriate
discount rate will have to be addressed. Finally, the calculation of
rent must account for the cost of risk bearing to the developer and a
normal return to his investment. These may be difficult to determine
and some simplifying assumptions may be needed.

Given the proper data it may also be possible to test the
hypotheses concerning competition and lease distribution. In the sealed
bid setting the collection of information on the number and identity of
bidders on each lease is straightforward. If similar information could

be generated from the oral auctions it may be possible to determine

105

whether there is a difference between the two methods. The nature of
the oral auction, however, makes it difficult to count and identify
bidders.

While there are difficult problems associated with empirical tests
of these hypotheses, the information gained from such tests could be
very useful. Feasible empirical tests of auction markets may provide a

fruitful area for future research.

Bibliography

Arrow, Kenneth J., Essays in the Theory of Risk-Bearing, Chicago,
Markham Publishing Company, 1971

 

Bernoulli, Daniel, "Exposition of a New Theory on the Measurement of
Risk", Econometrica, 1951, pp. 23-36

 

Capen, E. C., Clapp, R. V. and Campbell, W. M., "Competitive Bidding in
High Risk Situations", Journal of Petroleum Technology, June 1971, pp.
641-653

 

Deaton, Angus and Muellbauer, John, Economics and Consumer Behavior,
Cambridge University Press, 1980

 

Department of Natural Resources, Policy for Oil and Gas Leases on State
Lands, Unpublished Department Document, State Of Michigan Department of
Natural Resources, Policy No. 2306, April 16, 1982a

 

Department of Natural Resources, Geological Survey Division, Annual
Statistical Summary #88, Michigan Oil and Gas Fields, 1982b

 

 

Engelbrecht-Wiggans, Richard, "Auctions and Bidding Models: A Survey",
Management Science, 26(1980), pp. 119-142

 

Engelbrecht-Wiggans, Richard, "An Introduction to the Theory of Bidding
for a Single Object", Auctions, Bidding and Contracting, Richard
Engelbrecht-Wiggans, Martin Shubik and Robert M. Stark editors, New York
University Press, 1983, pp. 53-104

 

 

Friedman, Milton and Savage, L.J., "The Utility Analysis of Choices
Involving Risk", Journal 9f Political Economy, August 1948, pp. 279-304

 

Gilley, Otis W. and Karels, Gordon V., ”The Competitive Effect of Bonus
Bidding: New Evidence, Bell Journal, Autumn, 1981

 

Hardin, Garrett, "The Tragedy of the Commons”, in Managing the Commons,
Garrett Hardin and John Badden, editors, W. H. Freeman, 1977

Holt, Charles A. Jr., "Competitive Bidding for Contracts Under Alterna-
tive Auction Procedures", Journal 2f Political Economy, 3, 1980, pp.433-
445

 

Howe, Charles W., Natural Resource Economics, John Wiley and Sons Pub-
lishers, 1979

 

106

107

Hughart, David, "Informational Asymmetry, Bidding Strategies, and the
Marketing of Offshore Petroleum Leases", Journal gf Political Economy,
1975, pp. 969-985

 

Knight, Frank H., Risk, Uncertainty and Profit, Augustes M. Kelley,
Bookseller, 1964

 

Kunreuther, Howard and Slovic, Paul, "Economics, Psychology,and Protec-
tive Behavior", The American Economic Review, Papers and Proceedings;
May 1978, pp. 64-69

 

Landsberg, Hans H., Energy: The Next Twenty Years, Report by a study
group sponsored by the Ford Foundation and administered by Resources for
the Future, Hans H. Landsberg, Chairman, Ballinger Publishing Company,
1979

 

 

Leftwich, Richard H., The Price System and Resource Allocation, The
Dryden Press, 1976

 

 

 

Leland, Hayne E., "Optimal risk Sharing and the Leasing of Natural
Resources, with Application to Oil and Gas Leasing on the OCS", Th2
Quarterly Journal of Economics, August, 1978, pp. 413-437

McDonald Stephen L., The Leasing 2f Federal Lands for Fossil Fuels
Production, Published for Resources For the Future, by Johns Hopkins
University Press, 1979

 

 

Mead, Walter J., Competition and Oligopsony in the Douglas Fir Timber
Industry, University of California Press, Berkely, California, 1966

 

Mead, Walter J., "Natural Resource Disposal Policy - Oral Auction Versus
Sealed Bids", Natural Resources Journal, April, 1967, pp. 194-224

Meyerson, Roger B., "The Basic Theory of Optimal Auctions", Auctions,
Bidding and Contracting, Richard Engelbrecht-Wiggans, Martin Shubik and
Robert M. Stark editors, New York University Press, 1983, pp. 149-163

Milgrom, Paul R. and Weber, Robert J., "A Theory of Auctions and Com-
petitive Bidding”, Econometrica, September, 1982, pp. 1089-1122

Patric, William C. and Kakela, Peter J., "Michigan's Oil and Gas
Regulations", Extension Bulletin E-l648, Cooperative Extension Service,
Michigan State University, August, 1982a

Patric, William C. and Kakela, Peter J., “Michigan's Oil and Gas In-
dustry: Past, Present and Future", Extension Bulletin E-l647, Coopera-
tive Extension Service, Michigan State University, August, 1982b

 

Patric, William C. and Kakela, Peter J., "Compulsory Pooling", Extension
Bulletin §-1647, Cooperative Extension Service, Michigan State Univer-
sity, August, 1982c

Pratt, John W., "Risk Aversion in the Small and the Large",
Econometrica, 1964 pp.122-l36

108

Ramsey, James B., Bidding and Oil Leasing, Jai Press Inc., 1980

 

Randall, Alan, Resource Economics, Ag Economic Approach E9 Natural
Resources and Environmental Poligy, Grid Publishing Inc., 1981

 

 

Reece, D. K., ”Competitive Bidding for Offshore Petroleum Leases", Bell
Journal pf Economics, Autumn 1978a, pp. 369—384

Reece, Douglas K., "An Analysis of Alternative Bidding Systems for
Leasing Offshore Oil", The Bell Journal pf Economics, 1978b, pp. 659—669

 

Riley, J. G. and Samuelson, W. F., "Optimal Auctions", American Economic
Review, 1981, pp. 381-392

 

Robinson, Marc S., Oil Lease Auctions: Reconciling Economic Theory with
Practice, Discussion Paper #292, Department oF Economics, University of
California, Los Angeles, Revised September, 1984

Rothkopf, Michael H., “A Model of Rational Competitive Bidding”, Manage-
ent Science, March, 1969, pp. 362-373

Rothkopf, Michael H., “Bidding in Simultaneous Auctions with a Con-

straint on Exposure", Operations Research, July-August, 1977, pp. 621-
629

 

Scherer, F. M., Industrial Market Structure and Economic Performance,
Houghton Mifflin Company, 1980, Second Edition

 

 

Schmid, A. Allan, Property, Power and Public Choice, Preager, 1978

 

Schmid, A. Allan, and Schaffer, James D., Community Economics, A
Framework for Analysis pf Community Economic Problems, Unpublished
document, Michigan State University, 1983

 

 

Simon, Herbert A., Models 2f Man, Social and Rational, John Willey and
Sons Inc., 1957

 

Simon, Herbert A., "Rationality as Process and Product of Thought", Th3
American Economic Review, Papers and Proceedings, May 1978, pp. 1-16

Simon, Herbert A., "Rational Decision Making in Business Organizations",
The American Economic Review, September 1979, pp. 493

Smith, James L., ”Non-Aggressive Bidding Behavior and the "Winner's
Curse" ", Economic Inquiry, July, 1981, pp. 380-388

Stark, Robert M. and Rothkopf, Michael H., ”Competitive Bidding: A
Comprehensive Survey”, Operations Research, 1979, pp. 364-390

 

Tversky, A. and Kahneman, D., "Judgment under Uncertainty: Heuristics
and Biases", Science, 1974, 1124-1131

Vickery, W., "Counterspeculation, Auctions, and Competitive Sealed
Tenders", Journal pf Finance, March, 1961, pp. 8-37

 

109

Von Neumann, John and Morgenstern, Oskar, Theory pf Games and Economic
Behavior, Princeton University Press, 2nd ed., 1947

 

Weber, Robert J., ”Multiple-Object Auctions", Auctions, Bidding Egg
Contracting, Richard Engelbrecht-Wiggans, Martin Shubik and Robert M.
Stark editors, New York University Press, 1983, pp. 53-104

Wilson, R. B., "A Bidding Model of Perfect Competition“, Review pf
Economic Studies, October 1977, pp. 511-518

  

   

L.,juiflm‘lm

Hill"IIIUMIIIIHIHI‘MI“

31293 03

  
  

n’iifitlfinflu