Resource depletion and technical change: effects on U.S. crude oil finding costs from 1977 to...

INTRODUCTION

Finding oil, M.A. Adelman once wrote, reflects an "endless tug of war" between increasing knowledge and diminishing resources. Increasing knowledge leads to advances in technology, reducing finding costs over time.(1) At the same time, ongoing exploration and development depletes

Finding oil at a competitive cost is a fundamental requirement for survival in the petroleum industry. Securities analysts consider low-cost reserve replacement an important criteria in valuation of oil and gas producers (EIA, 1995c; John S. Herold, Inc., 1992; Oil and Gas Investor, 1997), and producers frequently identify minimization of finding costs as an important strategic goal (Amoco, 1988; Phillips, 1992; Fina, 1993; Unocal, 1994).

In the early 1980s high oil prices supported relatively expensive exploration and development projects. In 1981, when oil prices peaked at $59 (1996 dollars) per barrel, crude oil finding costs were $21.06 per barrel onshore and $33.76 offshore in the United States (Table 1). The price collapse of 1986 forced producers to reduce finding costs in order to stay competitive. Onshore finding costs declined to $7.84 per barrel by 1990, and to $5.82 by 1994. Offshore finding costs fell to $12.52 by 1990, and to $6.09 per barrel by 1994.

What portion of the substantial cost reduction can be attributed to technical progress? What was the magnitude of the countervailing effect of depletion? These issues have provoked intense interest among petroleum analysts. Recent research focuses on finding rates (Forbes and Zampelli, 1995; Cleveland and Kaufmann, 1997; Iledare, 1997), but a comprehensive analysis of the effects of technology and depletion on crude oil finding costs has yet to be produced. That is the purpose of this paper.

Additions, not Extraction

Finding costs reflect exploration and development, the activities directed toward adding to proved reserves. Additions to proved reserves can be thought of as 'flows from unknown resources into a reserve inventory" (Adelman, 1990), and are distinct from oil production, which represents extraction from a reserve inventory. The economics of extraction of exhaustible resources involves issues such as the empirical validity of the Hotelling Valuation Principle (Hotelling, 1930; Dasgupta and Heal, 1974; Adelman, 1990; McDonald, 1994; Watkins, 1992; Adelman and Watkins, 1995), which are outside the scope of this paper.

TECHNICAL CHANGE

The reduction in finding costs in the United States is widely credited to application of state-of-the-art knowledge in computer technology, geophysics, and drilling. Three-dimensional seismology has been one of the most important of these cost-saving applications (Haar, 1992). Three-dimensional seismic images are sharper than earlier two-dimensional images, and help delineate oil reserves hidden by complex geological structures (AAPG, 1991). Large parallel processing computers must be used to harness the otherwise unwieldy level of detail generated by 3-D surveys (CERA, 1996).

Table 1. Highlights of U.S. Oil Exploration and Development
(1981-1994)

Industry-Wide Indicators

                        1981      1986      1990      1994

Oil Price               59.04     19.93     26.21     16.39
(19965/bbl)

Industry Well Drilling Cost (oil wells, 19965 per foot)

Onshore                 96.13     78.05     74.57     60.68
Offshore               519.08    454.27    365.35    324.14

U.S. Exploration and Development Efforts and Results, Large
Producers

Exploration and Development Spending (billion 19965)(*)

Onshore                  19.6       9.9       5.2       3.0
Offshore                  8.9       3.8       2.4       1.7

Oil Wells Completed (number of wells)

Onshore                  7387      8362      3880      2157
Offshore                  381       364       241       201

Finding Rate (thousand bbls/well)

Onshore                   140       115       332       356
Offshore                  925       519       981      1678

Finding Cost (three-year weighted average 19965/bbl)

Onshore                 21.06     12.98      7.84      5.82
Offshore                33.76     23.86     12.52      6.09

* Estimated for oil only, excludes natural gas.

Sources: Oil price, Energy Information Administration, Annual
Energy Review; industry drilling cost, American Petroleum Institute
Basic Petroleum Data Book; all others, Energy Information
Administration, Financial Reporting System (FRS).

The industry trade press is full of examples of dramatic cost reductions due to the use of 3-D seismology. In 1981, after drilling eight dry holes in a prospect offshore Louisiana, Chevron acquired one of the first 3-D surveys in the Gulf of Mexico. After the 3-D mapping, 28 of the next 29 wells found hydrocarbons (AAPG, 1991). More recently, Amoco (1995) noted that "(i)mprovements in 3-dimensional seismic technology have allowed Amoco to cut its cycle time for exploration surveys dramatically; what used to take from three months to one year may now take two weeks to two months, saving $200,000 to $1.2 million per month."

Another cost-saving technology, adopted in the later part of the 1980s, was horizontal drilling. Most reservoirs are wider than they are deep, so wells drilled horizontally to follow the reservoir have more well bore exposed to hydrocarbons (EIA, 1993). Some geological formations, like the Austin Chalk in Texas, trap hydrocarbons in vertical fissures which are easy to miss when drilling vertically. Drilling horizontally makes it more likely to hit the trapped oil or gas, and permits more than one fissure to be drained with a single well. The industry has adopted other sophisticated drilling technology such as "measurement while drilling" based on complex controls systems that sniff out oil deposits while the well is drilled.

Recent advances in information technology foster the search for oil in promising but previously inaccessible deepwater areas. Before widespread adoption of CAD/CAM programs, use of hand-drawn designs and rudimentary PC programs required that large safety factors be built into offshore platforms to accommodate possible inaccuracies (CERA, 1996). Platforms intended for 10-15 years might be designed and built to last for 30 years, with considerable impact on their cost. In 1989, the Auger platform, at $1.2 billion, accessed 220 million barrels of oil and gas reserves (Stuart and Townsend, 1994), at a cost of $5.45 per barrel (Craig and Hyde, 1997). Auger was the largest find in the Gulf of Mexico in 25 years, until four years later when the Mars platform was completed (Energy Report, 1996). Rapid development of CAD/CAM applications allowed Mars to be built in deeper water with 20 percent less steel than Auger. With the same $1.2 billion price tag as Auger, Mars reached more than twice the reserves, reducing the cost of the reserves accessed to $2.40 per barrel (Craig and Hyde, 1997).

PROSPECT HIGHGRADING

With or without the help of new technology, the oil price collapse of 1986 and the following period of lower prices made it imperative to cut costs. In 1981, the year oil prices peaked, the majors spent an estimated $28 billion (1996 dollars) on U.S. oil exploration and development (Table 1). They drilled 7,387 onshore oil wells, and 381 offshore oil wells. By 1990 the largest producers had cut oil exploration and development spending to under $10 billion, completing 3,880 onshore oil wells, and 241 offshore oil wells. Drilling only the most promising prospects, or "prospect highgrading," increased the onshore finding rate (reserves added per well completed, including dry holes) from 115,000 barrels per well in 1986 to 332,000 barrels per well by 1990, although continued cutbacks in onshore drilling after 1990 appear to have had very little effect on onshore finding rates. Offshore, finding rates went up from 519,000 barrels per well in 1986 to 981,000 barrels per well in 1990, and continued to increase to 1.7 million barrels per well in 1994. Because higher finding rates are expected to correspond to lower finding costs, prospect highgrading was widely adopted in the struggle to keep finding costs in line with falling oil prices.

RESOURCE DEPLETION

From 1977 through 1994, the majors drilled 75,000 successful oil wells onshore, and 4,000 successful oil wells offshore. The cumulative result of the oil and gas industry's exploration and development efforts (for the United States as a whole) was the addition of 93 billion barrels of oil to onshore proved reserves, and 12 billion barrels to offshore proved reserves.

It is widely recognized by geologists and exploration economists that empirical modeling of oil supply should account for the effects of cumulative exploration and development. M. King Hubbert (1967) was the first to describe the effects of depletion in a formal model, specifying an exponential decline in incremental reserves added as cumulative exploration and development increases. Economic studies of oil supply often incorporate such a "discovery decline" relationship into models of economic optimization (Pesaran, 1990; Walls and Jones, 1990; Walls, 1991). Cumulative effort is typically measured as the cumulative number of wells drilled or cumulative drilling footage.

A COST-MINIMIZATION MODEL OF EXPLORATION AND DEVELOPMENT

The producer is assumed to minimize the cost of finding oil. The parameters of the total cost minimization problem can be estimated using the translog cost function (Christensen, Jorgenson, and Lau, 1973), and the average cost function can also be represented empirically by the translog (Berndt and Wood, 1982). Factor shares can be derived from the average cost function via Shephard's Lemma (Berndt and Wood, 1982). The data used in this paper are sufficiently disaggregated to identify three factor shares: wells, labor, and acreage. Second-order terms for factor prices and output allow for flexibility in representing non-homothetic production technology and non-constant returns to scale. Allowing exogenous shifts due to depletion, highgrading and technical progress results in:

[Mathematical Expression Omitted] (1)

where

c = Average finding cost

CW = Cumulative wells [drilled.sub.(t-1)]

Y = Additions to proved reserves

[P.sub.W] = Industry drilling cost per foot, oil wells

[P.sub.L] = Industry price of labor

[P.sub.A] = Industry price of acreage

t = Technical change index

H = Highgrading (number of wells drilled).

Linear homogeneity requires imposing the following restrictions on the translog average cost function:

[summation over i] [[Gamma].sub.i] = 1

[summation over i] [[Gamma].sub.ij] = 0

[summation over j] [[Gamma].sub.ij] = 0

[summation over i] [[Gamma].sub.ij] = 0 (2)

The cost-minimizing share equations are obtained by logarithmically differentiating the translog, assuming output quantities and factor prices are fixed, and by using Shephard's Lemma:

[S.sub.W] = [Delta]lnc/[Delta]ln[P.sub.W] = [[Gamma].sub.W] + [[Gamma].sub.WA]ln[P.sub.A] + [[Gamma].sub.WW]ln[P.sub.W] + [[Gamma].sub.WL]ln[P.sub.L] + [[Gamma].sub.WY]lnY + [[Gamma].sub.Wt]t

[S.sub.L] = [Delta]lnc/[Delta]ln[P.sub.L] = [[Gamma].sub.L] + [[Gamma].sub.AL]ln[P.sub.A] + [[Gamma].sub.WL]ln[P.sub.W] + [[Gamma].sub.LL]ln[P.sub.L] + [[Gamma].sub.LY]lnY + [[Gamma].sub.Lt]t (3)

[S.sub.A] = [Delta]ln[c/[Delta]ln[P.sub.A] = [[Gamma].sub.A] + [[Gamma].sub.AA]ln[P.sub.A] + [[Gamma].sub.WA]ln[P.sub.W] + [[Gamma].sub.LA]ln[P.sub.L] + [[Gamma].sub.AY]lnY + [[Gamma].sub.At]t

where

[S.sub.W] = Cost share of wells

[S.sub.L] = Cost share of labor

[S.sub.A] = Cost share of acreage.

To assess the effect of technical change on average cost differentiate (1) with respect to t (Berndt and Wood, 1982):

[Delta]lnc/[Delta]t = [[Gamma].sub.t] + [[Gamma].sub.Wt]ln[P.sub.W] + [[Gamma].sub.Lt]ln[P.sub.L] + [[Gamma].sub.At]ln[P.sub.A] + [[Gamma].sub.tY]lnY + [[Gamma].sub.tt]t (4)

The negative of this result is often referred to as the rate of cost diminution. The rate of cost diminution accelerates if [[Gamma].sub.tt] is negative, and decelerates if [[Gamma].sub.tt] is positive. It will vary over time if input prices vary over time. It may also vary with respect to the level of reserve additions ([[Gamma].sub.ty]). Equation (4) is expected to take a negative sign.

The bias in technical change for factor i is defined (Berndt and Wood, 1982) as:

[B.sub.i] = d[S.sub.i]/dt [multiplied by] I/[S.sub.i]

= [[Gamma].sub.it] [multiplied by] I/[S.sub.i] (5)

where [S.sub.i] = share of input i.

If [B.sub.i] [less than] 0, technical change is i-saving, if [B.sub.i] [greater than] 0, technical change is i-using, and if [B.sub.i] = 0, technical change is i-neutral. Since factor shares are positive, the sign of [B.sub.i] depends on the sign of [Gamma].sub.it]. [B.sub.i] represents the relative (but not absolute) effect of technical change on demand for the factors of production (Berndt and Wood, 1982). In other words, labor-using technical change does not necessarily increase demand for labor, but it does not reduce demand for labor as much as demand for another factor. The bias in technical change is reflected in [[Gamma].sub.Wt], [[Gamma].sub.Lt], and [[Gamma].sub.At] in the translog average cost function (Moroney and Trapani, 1981; Berndt and Wood, 1982).

As mentioned earlier, when oil prices fall firms highgrade their exploration and development plans. They drill fewer prospects and larger prospects, with higher finding rates and lower finding costs. Decisions to go ahead with specific projects are usually made on a yearly basis, during the firm's planning cycle, so that the number of projects (which could include multiple wells in a single project) can be considered fixed for a given year, and the firm will adjust all other inputs with respect to its yearly plan. Ideally, to represent the effect of highgrading decisions, we would use the number of planned projects for a firm in a given year. However, that data is not available, so for this paper, the number of wells drilled is used as a proxy for the number of projects, and included in the translog average cost function. This specification of the cost function captures the nature of the planning horizon in which highgrading decisions are made. The expected sign of [[Gamma].sub.H] is positive. The sign of [[Gamma].sub.CW] is expected to be positive, because cumulative drilling results in resource depletion, increasing average finding cost.

The average cost function is estimated with onshore data and offshore data separately, using iterative seemingly unrelated regression (ITSUR). A matrix of firm dummy variables is included because preliminary results indicated that a fixed-effects model is appropriate for both the onshore and offshore data sets (Fagan, 1995).

DATA CONSTRUCTION AND SOURCES

Average finding costs are defined as three-year weighted average oil exploration and development expenditures per barrel of proved oil reserves added, by firm and year. All expenditure and reserve additions data are from the Energy Information Administration, FRS data base, for 27 major oil producers from 1977 through 1994. Due to mergers, acquisitions, and divestitures some firms do not appear in some years. Exploration and development expenditures include expenditures on well drilling (well share), expenditures on unproved acreage acquisition (acreage share), and expenditures on geological and geophysical and all other activities (labor share). Except for well drilling expenditures, exploration and development expenditures are not reported for oil and gas separately, so the portion of these costs incurred by oil operations alone must be inferred. Following Adelman (1992) this is done by assigning expenditures to crude oil operations based on the percentage of total oil and gas wells completed in the year accounted for by oil. The cost of dry holes is assigned to oil on this basis, too.

Proved reserves are defined as the estimated quantities of crude oil and natural gas liquids which geological and engineering data demonstrate with reasonable certainty to be recoverable in future years from known reservoirs under existing economic and operating conditions (EIA, 1994). Additions to proved reserves are attributed to extensions and discoveries, revision to previous estimates, and improved recovery; all three categories are included in yearly reserve additions.

The price of labor is hourly earnings in SIC 138, Oil and Gas Field Services, from the U.S. Bureau of Labor Statistics. The cost of wells is the industry cost per foot for oil wells, (for onshore and offshore separately) from the American Petroleum Institute (1995). The price of acreage is the cost per acre of Federal oil and gas leases, calculated by dividing total yearly lease bonuses paid on Federal land by yearly acreage awarded by the Federal government, for the industry as a whole. Onshore lease bonuses and acreage awarded are from the U.S. Department of the Interior, Bureau of Land Management. Offshore lease bonuses and acreage awarded are from the U.S. Department of the Interior, Minerals Management Service (MMS). For offshore acreage, an adjustment is made to account for MMS's switch to area-wide leasing in 1984.

Cumulative exploration and development effort is measured as cumulative FRS exploration and development oil wells drilled (starting in 1977), onshore and offshore separately. Highgrading is represented by the total number of oil wells completed by a firm in a given year, onshore and offshore separately.

ESTIMATION RESULTS

Resource Depletion Significantly Increases Finding Cost

Results of parameter estimates for [[Gamma].sub.CW] for both the onshore and offshore data are positive and significant, reflecting the expected cost-increasing effect of depletion (Table 2). Onshore, a one percent increase in cumulative wells corresponds to a 0.45 percent increase in average finding cost. On an average yearly basis, this corresponds to a 7 percent increase in finding costs due to depletion. The depletion effect offshore is larger, 0.74 percent for every one percent increase in cumulative offshore wells, or an average yearly rate of 12 percent. Offshore operations in the United States were begun only about 50 years ago, and thus offshore fields tend to be newer than onshore fields. The smaller depletion effect onshore versus offshore is consistent with the shape of the discovery decline curve, which tends to flatten as fields mature (Uhler, 1979).

[TABULAR DATA FOR TABLE 2 OMITTED]

Prospect Highgrading Reduces Costs Slightly

The coefficient on prospect highgrading, [[Gamma].sub.H], takes the expected positive sign in both the onshore and offshore models. Onshore, a one percent cutback in the number of wells completed by a firm corresponds to a 0.06 percent decline in average finding cost. However, the result is not statistically significant at 95 percent, reflecting the relationship noted earlier - the large cutback in drilling activity in the 1990s corresponded to only small gains in onshore finding rates (Table 1). Offshore, a one percent cutback in wells drilled by the firm corresponds to a small but statistically significant 0.14 percent decline in finding costs.

Technical Change Accelerates from 1979 to 1994

The results of the estimation confirm and quantify the accelerating pace of technical change in the oil industry, where "As recently as the late 1980s, the U.S. Gulf of Mexico was being referred to as the dead sea..." (Skrebowski, 1997). Small cost savings from technology in the early 1980s accelerated to nearly 18 percent in 1994 for the U.S. offshore (Table 3). Onshore, the increase in yearly cost reductions due to technical change was less steady, although the rate of cost diminution reached nearly 15 percent in 1994.

Table 3. The Effect of Technical Change on Average Finding Cost(*)

                      [Delta]lnc/[Delta]t (in percent)
Year                     Onshore             Offshore

1979                         7.2             1.4
1980                        -1.6             0.3
1981                        -6.2            -0.8
1982                         0.6            -3.3
1983                         0.7            -4.7
1984                        -7.2            -5.9
1985                        -2.7            -7.7
1986                        -2.9            -8.6
1987                        -7.6            -9.2
1988                        -7.4            -9.9
1989                       -11.9           -11.1
1990                       -12.4           -12.4
1991                       -12.4           -13.4
1992                        -8.9           -15.0
1993                       -11.5           -15.8
1994                       -14.5           -17.6

* The positive values for [Delta]lnc/[Delta]t in some early
years reflect the positive coefficients for the bias in technical
change.

Technical Change Favors Labor

The positive signs on [[Gamma].sub.Lt] indicate that technical change was relatively labor-using both onshore and offshore. Although the U.S. oil industry endured a protracted period of downsizing and layoffs after 1986, the estimation results indicate that things might have been worse without the help of technology. Industry analysts point to continuing viability (due to new technology) of mature provinces as a boon to employment in the oil and gas industry: "The oil industry will face a challenging shortage in the years ahead of its greatest asset: people" (John S. Herold, Inc., 1996).

The estimation results for [[Gamma].sub.Wt], indicate that technical change was well-saving offshore, but well-neutral onshore. One would expect use of technology such as three-dimensional seismology, which prevents drilling of costly dry holes, and horizontal drilling, which allows one well to access several traps, to save on well drilling both onshore and offshore. However, technical improvement also allows drilling of prospects that might not otherwise have been profitable. The countervailing effect of a greater number of profitable prospects to drill may have offset the reduction due to efficiency gains for individual wells onshore.

Technical change was acreage-saving onshore, but relatively acreage-using offshore. Better well placement through the use of 3-D seismology, and better access through advanced drilling may have saved on acreage acquisition onshore. Offshore, the use of sophisticated platform construction techniques allowed drilling of deepwater acreage which may not have been considered viable without the existence of the technology.

The estimation results also indicate that technical change varied with respect to the level of reserve additions in both onshore and offshore regions. In the onshore, [[Gamma].sub.ty] is positive, ([Delta]lnc/[Delta]lnY is less negative over time), indicating that technical change reduced the cost of smaller finds more than it reduced the cost of larger finds. This is consistent with technology such as 3-D seismology and horizontal drilling, which make smaller finds profitable. Offshore, the negative sign on [[Gamma].sub.ty] indicates that technical change reduced the cost of larger finds more than the cost of smaller finds. This is consistent with the nature of technical change in the offshore, facilitating exploration and development of huge finds such as Auger and Mars.

Economies of Large Finds

The coefficients of the translog cost function allow ready insight into scale effects. The percentage change in average cost corresponding to a percentage change in the size of reserve additions is:

[Delta]lnc/[Delta]lnY = [[Gamma].sub.Y] + [[Gamma].sub.YY]lnY + [[Gamma].sub.WY]ln[P.sub.K] + [[Gamma].sub.LY]ln[P.sub.L] + [[Gamma].sub.AY]ln[P.sub.A] + [[Gamma].sub.tY]t (6)

The value of [Delta]lnc/[Delta]lnY varies over time and with input prices, but is negative for every onshore and offshore observation in the data. In a study of manufacturing industry, this would indicate economies of scale. But in oil exploration and development, "scale" must be found rather than built, so the decline in average finding costs at increasing levels of reserve additions might be better thought of as economies of large finds. In order to exploit such economies, the search for large finds (known as "elephant-hunting") is a key exploration and development strategy for the major oil companies (Solomon, 1993; Oil and Gas Journal, 1995).

CONCLUSIONS

Cleveland and Kaufmann (1991) attribute much of the debate between resource economists (the "optimists") and physical scientists (the "pessimists") over the quantity of ultimately-recoverable resources to assumptions about the relative effects of technical progress versus depletion. The results of this paper give reason for optimism. Technology more than mitigated depletion and was more important than prospect highgrading in reducing finding costs. This justifies optimism for the future of finding costs, because even a decelerating rate of technical progress could continue to dominate depletion effects. Barring cyclical increases in finding costs that could come from higher factor prices, finding costs could stay at current levels, or even continue to decline, as the search for oil in the United States goes on.

This research is based on the author's dissertation, "Measuring Cost and Efficiency in U.S. Crude Oil Resource Development: A Frontier Translog Cost Function Approach," The American University, Washington, D.C., April, 1995.

1. Finding costs are defined as the cost of exploration and development (excluding expenditures on purchases of proved reserves) divided by additions to oil reserves (excluding net reserve purchases). Finding costs are calculated as 3-year weighted averages, to smooth out volatility in discoveries, and reduce the lag between drilling and the booking of associated reserve additions (Gaddis, Brock, and Boynton, 1992; Arthur Anderson and Co., 1994; Energy Information Administration, 1995b).

REFERENCES

Adelman, M.A. (1990). "Mineral depletion, with special reference to petroleum." The Review of Economics and Statistics 72 (1) (February): 1-10.

Adelman, M.A. (1992). "Finding and developing costs in the United States, 1945-1986." Energy, Growth, and the Environment. Vol 7 of Advances in the Economics of Energy and Resources, John R. Moroney, (ed.) Greenwich, Connecticut: JAI Press, Inc.: 11-58.

Adelman, M.A. and G.C. Watkins. (1995). "Reserve asset values and the Hotelling valuation principle: further evidence." Southern Economic Journal 61(3) (January): 664-672.

American Association of Petroleum Geologists (1991). "Studies underscore value of 3-D." AAPG Explorer (November): 16-17.

American Petroleum Institute (1995). Basic Petroleum Data Book. VXII No. 2, Washington, D.C.: American Petroleum Institute, May.

Amoco Corporation (1988). Annual Report.

Amoco Corporation (1995). Annual Report.

Arthur Andersen and Company (1994). Oil and Gas Reserve Disclosures. Houston, Texas, July.

Berndt, Ernst R., and David O. Wood (1982). "The specification and measurement of technical change in U.S. manufacturing." Advances in the Economics of Energy and Natural Resources. Greenwich, Connecticut: JAI Press: 199-221.

Bureau of Labor Statistics (1991). Office of Employment and Unemployment Statistics. (Division of Monthly Industry Employment Statistics). Washington, D.C. Facsimile dated 28 June.

Cambridge Energy Research Associates (CERA) (1996). Quiet Revolution: Information Technology and the Reshaping of the Oil and Gas Business. Cambridge, Massachusetts: Cambridge Energy Research Associates.

Christensen, Laurits R., Dale W. Jorgenson, and Lawrence W. Lau (1973). "Transcendental logarithmic production frontiers." Review of Economics and Statistics 55(1) (February): 28-45.

Cleveland, Cutler J. and Robert K. Kaufmann (1991). "Forecasting ultimate oil recovery and its rate of production: incorporating economic forces into the models of M. King Hubbert." The Energy Journal 12(2): 17-46.

Cleveland, Cutler J. and Robert K. Kaufmann (1997). "Natural gas in the U.S.: how far can technology stretch the resource base?" The Energy Journal 18(2): 89-108.

Craig, Michael J.K. and Steven T. Hyde (1997). "Deepwater Gulf of Mexico more profitable than previously thought." Oil and Gas Journal. (March 10): 41-50.

Dasgupta, P.S. and G.M. Heal (1979). Economic Theory and Exhaustible Resources. Oxford, England: James Nisbet and Co., Ltd. and Cambridge University Press.

Department of Commerce, Bureau of the Census (1977, 1982, 1987). Census of Mineral Industries: Oil and Gas Field Services, Washington, D.C., MIC77-L-13C, MIC82-L-13C, MIC87-L-13C.

Energy Information Administration (1993). Drilling Sideways - A Review of Horizontal Well Technology and its Domestic Applications. DOE/EIA-TR-0565. Washington, D.C.: U.S. Department of Energy, Energy Information Administration. April.

Energy Information Administration (1994). U.S. Crude Oil, Natural Gas, and Natural Gas Liquids Reserves 1993 Annual Report. DOE/EIA-0216(93). Washington, D.C.: U.S. Department of Energy, Energy Information Administration. October.

Energy Information Administration (1995a). Annual Energy Review 1994. DOE/EIA-0384(93). Washington, D.C.: U.S. Department of Energy, Energy Information Administration. July.

Energy Information Administration (1995b). Performance Profiles of Major Energy Producers 1993. DOE/EIA-0206. Washington, D.C.: U.S. Department of Energy, Energy Information Administration. January.

Energy Information Administration (1995c). Oil and Gas Resource Development in the Early 1990s: An Expanded Role for Independent Producers. DOE/EIA-0600. Washington, D.C.: U.S. Department of Energy, Energy Information Administration. October.

Energy Report (1996). "Mars oil project sets records." (August 12).

Fagan, Marie N. (1995). "Measuring cost and efficiency in U.S. crude oil resource development, 1977-1990: a frontier translog cost function approach." Dissertation, The American University, Washington, D.C. (April).

Fina (1993). Annual Report.

Forbes, Kevin. and Ernest M. Zampelli (1995). "Technological progress, technical efficiency, and the search for oil and gas in the U.S. onshore." Proceedings of the 18th International Association for Energy Economics International Conference, July 5, Washington, D.C.

Gaddis, Dwight, Horace Brock, and Charles Boynton (1992). "Pros, cons of techniques used to calculate oil, gas finding costs," Oil and Gas Journal, 1 June: 93-95.

Haar, Robert (1992). "Three-dimensional seismology - A new perspective." Petroleum Supply Monthly, DOE/EIA-0109(92/12). Washington, D.C.: U.S. Department of Energy, Energy Information Administration. December.

Hotelling, Harold (1931). "The economics of exhaustible resources." Journal of Political Economy 39: 135-175.

Hubbert, M. King (1967). "Degree of advancement of petroleum exploration in the United States." American Association of Petroleum Geologists Bulletin 51: 2207-2227.

Iledare, Omowumi O. (1997). "The competing influence of technology, depletion, and economics on petroleum drilling productivity: how persuasive is the evidence in the Gulf of Mexico outer continental shelf?" Proceedings, SPE Hydrocarbon Economics and Evaluation Symposium. Dallas, Texas. Society of Petroleum Engineers: 57-66. (March).

John S. Herold, Inc. (1992). Herold's Reserve Replacement Cost Analysis. Greenwich, Connecticut. (July).

John S. Herold, Inc. (1996). The Coming Shortage of the Oil Industry's Greatest Asset: People. Stamford, Connecticut. (September).

McDonald, Stephen L. (1994). "The Hotelling principle and in-ground values of oil reserves: why the principle over-predicts actual values." The Energy Journal 15 (3): 1-17.

Moroney, John R., and John M. Trapani (1981). "Alternative models of substitution and technical change in natural resource intensive industries" in Modeling and Measuring Natural Resource Substitution. (eds). Ernst R. Berndt and Barry C. Field. Cambridge, Massachusetts: MIT Press.

Oil and Gas Investor (1997). 1997 Petroleum Analysts Yearbook (February).

Oil and Gas Journal (1995). "The routes to efficiency." (October 23) 31.

Pesaran, M. Hashem (1990). "An econometric analysis of exploration and extraction of oil in the U.K. continental shelf." The Economic Journal 100 (June): 367-390.

Phillips Petroleum Company (1992). Annual Report.

Skrebowski, Chris. (1997). "Shell leads rush to deeper waters." Petroleum Economist (January): 5.

Solomon, Caleb (1993). "The hunt for oil: petroleum industry pins future on finding large overseas fields." Wall Street Journal (August 25): 1, A4.

Stuart, Jan and Alan Townsend (1994). "Offshore technology: innovations test new frontiers." Petroleum Economist (April): 6-8.

Uhler, Russell S. (1979). "The rate of petroleum exploration and extraction." in Advances in the Economics of Energy and Resources. (ed). Robert S. Pindyck. Greenwich, Connecticut: JAI Press: 93-118.

Unocal Corporation (1994). Annual Report.

Walls, Margaret A. (1991). "Using an 'hybrid' approach to model oil and gas supply: a case study of the Gulf of Mexico outer continental shelf." Resources for the Future Discussion Paper ENR91-18, Washington, D.C. (September).

Walls, Margaret A. and Andrew S. Jones (1990). "From bad to worse: impacts of the 1986 oil price collapse." Resources for the Future Discussion Paper ENR90-08, Washington, D.C. (February).

Watkins, G.C. (1992). "The Hotelling Principle: Autobahn or Cul de Sac?" The Energy Journal 13 (1):1-24.