Past and Future of the Petroleum Problem: The Increasing Need To Develop Alternative Transportation Fuels

Transcription

1 TRANSPORTATION RESEARCH RECORD 1175 Past and Future of the Petroleum Problem: The Inreasing Need To Develop Alternative Transportation Fuels D. J. SANTINI An examination of the interations of petroleum supply and demand patterns suggests that the depletion of U.. oil reserves and the inreasing importane of oil onsumption in U.S. transportation are auses for greater attention to the development of alternative (nonpetroleum-based) transportation fuels. Absent suh development, the historial lessons presented l.n this paper suggest tltat another period of rising oil pries, errati oil market behavior, and subsequent eonomi dlffi. ulty is probable within the next two deades. Methanol is argued to be the most likely and most desirable substitute transportation fuel beause it an be produed more eonomially and used more effiiently than gasoline when derived from the rapidly expanding worldwide supplies of natural gas rf 4 The basi argument of this paper is that it is imperative that the United States now devote inreasing attention to developing substitutes for its petroleum-based transportation fuels. As Figure 1 shows, U.S. transportation now onsumes more oil than is produed in the entire ountry. At the urrent level of oil pries, this situation is being exaerbated by a utbak in U.S. oil prodution and an inrease in oil use in transportation. Contrary to popular opinion, the main ause of the post-1981 deline in oil pries and the reent oil prie ollapse is not inreased fuel effiieny of vehiles. Instead, the primary ause is the adoption-in less than a deade--of substitutes for oil by eletri utilities, industry, other businesses, and households. This onlusion is supported by data from Japan and Europe, as well as from the United States. Beause transportation is using an inreasing share of oil in the United States and other industrialized ountries, the ability to suessfully introdue substitute transportation fuels will be far more important when the next oil prie run-up begins, possibly as early as the 199s. In the ase of rhe oil prie run-up, pries were not driven down until transportation belatedly began to redue oil onsumption, reinforing the onsumption drop already under way in other setors. History shows that substitution in transportation fuel in the United States has typially taken deades (1 ). If real oil pries are to be effetively apped below 1981 levels, the United States annot risk waiting until after the start of another prie run-up to begin introduing substitutes for petroleum in transportation. Energy and Environmental Systems Division, Center for Transportation Researh, Argonne National Laboratory, 97 South Cass Avenue, Argonne, Ill M 44 M Average for Five- Intervals FIGURE 1 Petroleum problem: inreasing role of transportation, The 2th-entury relationship between petroleum market behavior and eonomi behavior in the United States and the world is examined in this paper. The history of U.S. petroleum markets and their past interations with patterns of U.S. eonomi growth are disussed, and Lh impliations of this history for transportation fuel markets through the end of the entury are examined. More reent international trends in petroleum onsumption and prodution are examined in detail, with speial a!lention given to those multiyear periods when trends (i.e., growth rates) in domesti U.S. onsumption and prodution temporarily diverge. These periods are referred to as "gaps." Suh gaps are shown to preede sharp oil prie movements. The oil prie movements themselves are assoiated with periods of relatively severe U.S. eonomi diffiulties. Sharp oil prie inreases "supply shoks" in maroeonomi terms-theoretially (2-4) and empirially (5, 6) are auses of reessions. Contrary to theory, 2th-entury statistial evidene, and simple intuition, 19th- and 2th-entury U.S. history does not indiate that energy prie ollapses ause eonomi booms (1 ). Thus both reent experiene and longer looks at historial evidene indiate that sharp energy prie movements are not desirable.

2 2 This paper and its supporting report (7) show that severe movements in petroleum prie-both upward and downwardan be antiipated beause they are preeded by a multiyear pronouned widening of the gaps just desribed. To redue the severity of petroleum prie swings, the gaps themselves logially must be redued in magnitude. It is argued here that lhe petroleum market exhibits errati prie behavior beause of diffiulties in developing substitutes for oil. It is suggested that preparation of suh substitutes must begin before prie signals indiate the desirability of suh preparation. If gaps an be "managed"-kept to a narrow range or substantially shortened in duration-the eonomy-wide onsequenes of subsequent wide swings in petroleum prie should be avoidable. This, of ourse, is the same type of argument used by those who would have government intervene to redue the ineffiienies of the business yle (2). However, the argument is limited in the sense that it applies only to one Sl:(.;tur of the eonomy and unique in the sense that wide gaps are early indiators of the need for poliy ation, allowing reasoned, areful development of poliies designed to narrow the gaps. THE "PETROLEUM PROBLEM" The "petroleum problem" faed by the United States and its allies has two faets: (a) the eonomi exhaustion of petroleum and (b) the inreasing dominane of petroleum onsumption by the fuel-inflexible transportation setor. Eonomi Exhaustion of Petroleum Undeniably, the United States is running out of eonomially reoverable petroleum. Further, aording to a statement attributed to OPEC President Rilwanu Lukman, it is a goal of OPEC to keep world oil pries low enough to disourage further development of Alaskan and North Sea reserves (8). Jn and of itself, this is not a theoretial problem. Many nations have maintained robust eonomi growth despite a lak of indigenous petroleum.. Exhaustion of domesti U.S. petroleum does, however, mean that the United States will no longer be able to lose onsumption-exess growth gaps by inreasing domesti petroleum prodution. Therefore the nation must rely inreasingly on measures that Inrease exports to earn the neessary foreign exhange to purhase imported oil, or Improve the thermodynami effiieny of equipment that uses petroleum produts, or Develop domestially produed substitutes for petroleum produts, or Import substitute fuels that are less ostly, or All of the preeding. Inreasing Dominane of Petroleum Consumption by the Fuel-Inflexible Transportation Setor The ability of the United States and its allies to rapidly redue onsumption of petroleum produts is diminishing beause the transportation setor has inreased its share of the industrialized nations' petroleum produts onsumption (Figures 1, 2) (9-14). Not only does transportation aount for a histor: Cl> Cl> a. 3 TRANSPORTATION RESEARCH RECORD 1175 E Exloong U.S FIGURE 2 Shares of oil onsumed by transportation In the United States and other OECD member nations, ially high share of U.S. oil onsumption, it also aounts for a far greater share of petroleum onsumption than it does in ompeting industrialized nations (Figure 2). Further, among member nations of the Organisation for Eonomi Cooperation and Development (OECD), the United States is far and away the most dominant user of oil in transportation (Figure 3). This is a problem beause eah of the other : Cl> 3 Cl> a. 2 1 U.S. Transportation OECO U. S. Europe Nontransportation FIGURE 3 Shares of total OECD petroleum produt onsumption aounted for in 1981, by setor and loation. petroleum using setors of the U.S. and other OECD eonomies managed to redue peu-oleum onsumption earlier and more rapidly than did the U.S. transportation setor after the two 197s oil prie shoks (Table 1) (12). Jn spite of legislation that mandated rapid gains in automotive fuel effiieny in the United States, the U.S. transportation setor had the least suess in reduing petroleum onsumption through 1985 (1). This pattern also held in Japan and Europe (12, 14). Oil Pries, Oil Imports, and Maroeonomi Ativity On another oasion in the United States ( ), there was a doubling in real rude oil prie during a period that was aompanied by inreasing oil imports and immediately followed by a severe reession (7, 13). Although maroeonomi theory now reognizes the oil "supply shoks" as auses of reent reessions (2-4), very little reexamination of the role of suh supply shoks as auses of pre-world War II reessions has taken plae. Jn this paper it is assumed that if oil prie

3 Santini 3 TABLE 1 PERCENTAGE DECLINE IN OIL CONSUMPTION FROM PEAK YEAR THROUGH 1981 FOR MAJOR OIL-CONSUMING LOCATIONS, BY SECTOR, POST OIL PRICE SHOCKa Industry Residential, Commerial, and Other Oil Consumers Peak Peak Loation Deline Deline United States United Kingdom Federal Republi of Germany OECD-Europe Japan For year in whih oil onsumption peaked through Eletri Utilities Transportation Overall Peak Peak Peak Deline Deline Deline shoks an be ited as auses of reent reessions in maroeonomis texts, other similar oil prie shoks may also be assumed to have ausally ontributed to earlier U.S. reessions if statistial analysis supports suh an assumption. Separate statistial analyses of the pre- and post-world War II periods do support this position (5, 6). Thus, if a reession-induing rude oil prie inrease (6) omparable to that of or [real prie inreases more than doubled in both ases (7, 9, 13)] ourred in the future, the United States, all other things being equal, would be less able to redue oil import purhases and osts than in the past. The nation would also suffer more than its industrial ompetitors, who should have less diffiulty reduing their oil onsumption beause the inflexibility of the transportation setor is relatively less of a problem for them (Figure 3). This appears to have been the ase in the years immediately after the large oil prie jump, beause Japan and Germany ahieved greater immediate suess than did the United States in reduing oil onsumption (Table 1). Similarly, sine the first oil prie shok in 1973, total oil onsumption as of 1986 had dropped by 19.1 perent in Japan, 17.7 perent in the Federal Republi of Germany, only 7.8 perent in the United States, and 16.5 perent elsewhere in the OECD (14). These delines in total oil onsumption ourred in spite of an 11 perent inrease in U.S. oil onsumption in transportation from 1975 to 1985 and an even greater 23 perent inrease during the same period in other OECD ountries (1, 14)! The purpose of this paper is to note and emphasize the possible importane of these fats in order to support the argument for the inreasing need for well-oneived transportation strategies that an quikly redue petroleum onsumption. The paper is not intended to prove that the desribed energy-tomaroeonomy relationships are orret. It is intended to raise issues and provoke thinking about the importane of advane planning for quik introdution of oil substitutes when oil pries rise. MANAGING THE NEXT GAP Need in Transportation for a Petroleum Substitute Introdution of substitute fuels in transportation will prove to be one of the most diffiult transitions ever to our within the U.S. eonomy. Beause of their physial properties, petroleumbased fuels make ideal transportation fuels. Potential substitutes are ostly and often inferior in key respets. Nevertheless, signifiant use of substitutes will eventually be needed, given the inevitable depletion of eonomial oil resoures. Beause the United States pioneered in the extensive use of oil in transportation (11, 15), it should logially be one of the first, if not the first, to fae the need for developing a widely used substitute. Methanol: A Possible Transportation Fuel Based on Natural Gas Many urrent analysts onsider methanol the most likely widespread substitute for petroleum-based gasoline (J 5-22). Methanol is a liquid fuel suitable for automobiles, and it an be made more heaply from natural gas than an gasoline (18). Although reent worldwide disovery rates of rude oil have been disappointing, natural gas has been added relatively rapidly to world reserves (9, 14). Drilling for oil has taken oil ompanies to inreasingly remote loations and greater depths. However, at greater depths the probability is higher that gas rather than oil will be found (21 ). Further, gas prie inreases of the 197s aused inreases in the searh for natural gas for its own sake. Previously, most natural gas was an unwanted produt assoiated with oil reovery. Reent hanges in drilling patterns have resulted in far greater suess in finding natural gas than oil. Sine the oil prie shok of 1973, world rude oil reserves have inreased by 5 perent while gas reserves have inreased by 93 perent (14). Thus, as oil beomes inreasingly more diffiult to find, the issue will be the kind of transportation fuel to make from the more readily available supplies of natural gas. Automakers, the Environmental Protetion Ageny (EPA), the state of California, and numerous other entities are spending signifiant amounts of money on experiments to assess and improve the ability of U.S. automobiles to bum methanol leanly and effiiently. The degree of suess of these and other experiments and the timing of future market development efforts will have a major effet on when or whether the world and the United States experiene another rude oil prie run-up as high as those of or

4 4 SLOW PROCESS OF REVERSAL OF OIL PRICE INCREASES Comparison of the and Oil Prie Paths The two worst 2th-entury run-ups of rude oil prie in the United States ourred in and , after the two most prolonged and pronouned gaps in whih growth rates of oil onsumption exeeded those of domesti prodution. When these gaps were reversed and replaed by the 2th entury's most prolonged and pronouned gaps of the opposite type, in whih rates of prodution growth exeeded rates of onsumption growth, oil pries were driven bak down (Figure 4). This strong reversal of the onsumption-versus-produtiongrowth gaps was due in part to unusually severe delines in real national inome, although energy onservation and fuel substitution also undoubtedly played a role. Fifteen years after the steady real-prie rise that began in 1915, oil pries plummeted to their 1915 level. In 1986, 13 years after a prie rise that began in 1973, oil pries one more plummeted toward their preshok level (Figure 5). Dunkerley and Hoh (23) used international data to estimate that the prie elastiity of "road transport" oil onsumption is low (--.2), but that inome elastiity is high ( + 1.3); thus, the present g "C.. :_ if 8 a B..: 6.. g-i O> :> 4 (5 2. iii i5 g - 2 "8., ::::> CD...: - - : - 6 I- - 8 >( :Jl-.. '=! - :g 2.5.g 3: "-., 2. - "' 1.5 -g.. -- g 1. ;. -g u.o-.5 al fl a :lit Pries Well in Exess or the Simple Trend, Following Prolonged, Extreme Consumption-Exeeding-Prodution Gaps One Standard Deviation Band around Mean Note: 1915 was first year that Oil aounted for 1% or U.S. onsumption One Standard Deviation Band Comparable to Band in Graph Above FIGURE 4 Movements of U.S. oil onsumption-versusprodution-growth gaps against real oil pries, CJ) Ql Ql 3. (ii CJ) Ql 2.5.<::: 3': 2..s Ql 1.5 > :; Qi 1. : Ql.5 &. 2.5 CJ) Ql Ql (ii CJ) Ql 1.5.<:::.s Ql 1. > 'id :. 5 Ql Inrease Peak No. 1 Base '" Base TRANSPORTATION RESEARCH RECORD 1175 Oil Inrease Peak No FIGURE 5 Long-term rise and fall of real oil pries: omparison of the and periods. author infers that inome losses in reessions and larger prie responses in other setors do more to drive oil pries bak down t,;.an do prie responses by the transportation setor. One major purpose of this paper is to dery this unfortunate inferene and to argue, like Dunkerley, that ations to improve the prie response are desirable if governments are to avoid using inome losses to push down oil pries after prie shoks (23). That unusually severe U.S. reessions followed both of the worst U.S. real oil prie run-ups implies that balane in the energy setor must be maintained if eonomi stability is to be ahieved. If a prolonged and broad onsumption- exess growth gap emerges, history implies that it must be balaned by another prolonged and broad gap in whih growth rates of prodution exeed those of onsumption. During the 2th entury under the U.S. eonomi system, this has twie been aomplished by an initially sharp, and then prolonged, inrease in real rude oil prie that enouraged prodution, disouraged onsumption, and-onsistent with the "supply shok" theory (2, 3)---<:aused national output to be redued far below the "natural" level. Consistent with the theoretial proess of reovery from a supply shok, these ombined effets

5 Sanlini reversed the gap and depressed the rude oil prie. On the first oasion ( ), this proess was followed by a period during whih onsumption and prodution were more or less in balane, and, aside from one mild reession, steady, robust eonomi growth (the Roaring Twenties) was experiened from the reession trough in 1921 until the end of the deade. In that ase, the United States had the benefit of domesti oil disoveries but later paid the osts of writing down the value of those oil assets in 1931, during the ollapse into the Great Depression. On the first oasion, an initial sharp oil prie run-up started in 1915 and was followed by a seond oil prie shok and subsequent ollapse 15 years later. On the seond oasion ( ), not muh new domesti oil was found after the first prie shok, and a seond shok, whih was more severe than that of the mid-192s, ensued. On this oasion, oil and transportation aounted for more of U.S. total energy onsumption than in the ase, and the start-to-peak real oil prie inrease was greater (Figure 5). When oil pries ollapsed in 1986, the absene of domesti disoveries was an advantage in that other nations arried a greater part of the burden of writing off oil assets. Nevertheless, reent events in the stok market raise the possibility that the proess of writing off domesti oil assets may one again have ontributed to finanial diffiulties in the banking system. Low oil pries have been refleted in low stok values for energy-enter banks in partiular, and the banking system in general, as well as reord post-world War II bank failures. The oil market's behavior in and was similar in several respets. Both periods saw dramati real oil prie run-ups, with two distintly dillerent prie shoks peaking 6 years apart. After the two prie shoks, both periods witnessed prolonged and severe erosion of rude oil pries. In both ases, a ollapse in rude oil pries ourred about 5 to 6 years after the seond oil prie shok (see Figure 5). Current evenls also imply that a rebound from the ollapse, followed by another ollapse, may also be a similarity of the two periods (see Figure 4). A key finding of this analysis is that prie movements within the petroleum market balane out (ome to equilibrium), as basi eonomi theory would lead one to expet (2, 3). However, the proess by whih balane isreestablished (after a onsumption...:.exess growth boom) has historially been quite long in duration and has involved reessions after the upward prie movements. The analysis thus implies that the initial gaps themselves preede prie shoks that ause subsequent deelines in eonomi ativity. When the gaps are reated through delines in rude oil prodution, this sequene of events is onsjstent with "supply shok" textbook theory (2, 3). The widening of the gaps that preedes prie shoks appears to be identifiable well in advane of the shok and the national inome delines that follow (Figure 4). Given these harateristis, it would appear to be both possible and desirable to take steps to redue the size of gaps before they ontribute to the onsequenes identified here. Thus it is argued that the auses of gaps should be onsidered in advane so that the size of gaps an be limited (i.e., gaps should be "managed"). Statistial Evidene of Setoral Responses to Oil Pries from 1973 to 1986 The desriptive examination of the and periods showed that a prodution-in-exess-of-onsumption growth gap eventually drove real oil pries down in response to two prior large real oil prie inreases that were separated by about 6 years. Pries were not pushed bak to the level that existed at the beginning of eah period until this prodution-inexess-of-onsumption growth gap emerged. One question that arises from observation of this pattern of market behavior is "why was the initial prie rise insuffiient to ause enough substitution to drive the relative prie bak to its original level?" Although the emphasis here is on the relative diffiulty of introduing substitutes for petroleum in transportation, it is also true that it is diffiult, in an absolute sense, to introdue substitutes for petroleum in the short run in any major petroleum-using setor of the eonomy. The proess of substitution an be desribed as ourring in two phases logially separated by a few years. First, immediately after a prie shok, onsumers and investors substitute existing tehnologies to redue oil onsumption as muh as possible. Some industrial and business onsumers take advantage of a built-in ability to immediately swith fuels. In the seond part of the substitution proess, produers of energyintensive produts begin programs of researh, development, and produt introdution to implement new tehnologies that are more thermodynamially effiient or use more abundant, less expensive fuels, or both. The lag between the start of suh programs and the first ommerial sale of new produts is typially several years. For example, flow harts (24-26) and publi statements (2) by manufaturers indiate that 5 years are needed lo omplete engine development programs. If suh lags also exist for furnaes, air onditioners, boilers, eletri generators, and proess equipment, it might be reasonable to expet a lagged reation to sharp oil prie inreases after about half a deade. If this desription of the substitution proess is orret, then on those oasions when the first round of substitution- using existing produts-is suffiient to push the prie down to its original level, produt development programs might well be shelved in favor of ontinuing prodution of existing tehnologies. However, if the initial substitution of other ommerially available tehnologies does not push pries bak down, a seond prie shok ould be expeted to ause ommerial introdution of new tehnologies. If this is an aurate desription of the proess, the ombined short- and long-term substitution responses to the pair of prie shoks should be suffiient to push oil onsumption below prodution and indue the needed prie deline. The problem with this sequene is that it implies that several years of sustained high pries must our to ause the widespread adoption of new oil-substituting and oil-onserving tehnology. A. J. Sobey of General Motors has argued that it would take several years of gasoline pries well above reent levels to indue widespread substitution of methanol for gasoline (J 6). More reently, Sobey indiated that the proess of substitution based on some low-ost non-u.s. gas soures ould start at an oil prie of $2 per barrel (J 7). If the proess just desribed is aurately depited, it is legitimate to expet the temporary but severe eonomi diffiulties assoiated with petroleum prie shoks to one more our if large-sale development and implementation of oil-substituting tehnology in the transportation setor are not begun until after another substantial rise in oil pries. To test for the existene of these effets and to statistially desribe the nature of this proess, the setor-speifi short- 5

6 6 and long-run prie elastiities of U.S. oil (and energy) onsumption rates per unit of output for the period were estimated using an individual, 13-observation, ordinary least squares regression model for eah of the four major petroleum-onsuming setors, as ategorized by the Energy Information Administration. The oil onsumption data series that was used begins with 1973 (JO). Annual oil onsumption per unit of output was estimated by dividing setor-speifi oil onsumption by a measure of output for that setor. The output index for eletri utilities was net generation (JO), for industry it was the Federal Reserve Board industrial output index (27), for residential and ommerial oil onsumption it was onstant dollar personal onsumption expenditures (28), and for transportation it was vehile miles of travel (VMT) (29). The advantage of measuring hanges in energy use per unit of output is that the short-run elastiity estimates do not suffer from the upward bias (of the absolute magnitude of the negative elastiity estimate) resulting from the output losses that immediately follow an energy prie shok. This short-run bias showed up in unreported regression models in whih the author tested for its existene. In the ase of the residential/ommerial setor, it was reasoned that the best measure of "output" would be a proxy for onsumer utility, and real personal onsumption expenditures were seleted as the best proxy measure. In the ase of transportation, the VMT series ended in 1985, so transportation estimates do not inlude the immediate response to the 1986 oil prie ollapse. A hek of the effet of deleting 1986 in the TRANSPORTATION RESEARCH RECORD 1175 other models of oil onsumption showed no sign hanges in otherwise signifiant variables, as well as onsistently larger hanges of elastiity estimates for short-run effets than for long-run effets. Changes in short-run elastiity estimates were -26 to +8 perent whereas hanges in long-run estimates were -11 to + 18 perent. The largest oeffiient shifts ourred in the residential/ornrnerial equation, in whih the output index is most questionable in any ase. On the whole, given the rather large differenes in elastiity estimates aross setors, these experiments suggest that the transportation estimates an be reliably ompared with those for the other setors in spite of the exlusion of the "outlying" 1986 observation in the transportation model. The rate of hange of oil onsumption per unit output ( ), the rate of hange of energy onsumption per unit (E ), and the rate of hange ofreal rude oil prie [ (P ;), in 1967 dollars, from the Amerian Petroleum Institute (9)], were omputed by logging (using base 1) the ratio of this year's value to last year's value. Short-run elastiities were initially estimated by regressing this year's - and E-values against present (P ) and past oil prie hanges (P 1, P 2,. Pi). Statistially signifiant prieindued redutions in ourred immediately in the eletri utility, industrial, and residential/ommerial setors, but not until the year after the oil shok for transportation (see Table 2). After a lag of 1 year, statistially signifiant effets onsistently disappeared in regression models with two or more P; lags inluded. For omparability aross all setors, elastiities for ontemporaneous reations and for reations after 1 year TABLE 2 SECTORAL ESTIMATES OF SHORT-RUN AND LONG-RUN ELASTICITY OF OIL AND ENERGY CONSUMPTION PER UNIT OF OUTPUT WITH RESPECT TO OIL PRICEa Elastiity Estimates Short Run 2- Sum Long Run Adjusted Setor Po P1 (Po+ P1) (l) If Oil Eletriity (-3.71)b (-.5) (-3.74)b (8.9l}b Industrial (-2.69f (-.66) (-2.96)b (3.97) Residential/ ommerial (-l.85f (-1.36) (-i.22) (2.27) Transportation (-.57) (-2.58f (-3.2l)b (8.7)b Energy Eletriity d (.22) (.5) (-.36) (.9) Industrial - O.Q (-1.7) (-.16) (-3.14)b (3.43f Residential/ :: :: :: d ommerial NEe NE NE (-1.65) (2.73) Transportation (.21) (-2.48f (-2.31f (6.15) al-statistis for oeffiients are shown below the oeffiient in parentheses; F-statistis for model K are below the If- value in parentheses. bstatistially signifiant at the 1 perent level (two-tail). statistially signifialll llt the 1 perent level (two-tail). d A property of the small sample adjustment of the R 2 statisti is that a negative value an result from the adjusunnt. ene = Not estimated in the "best" model presented in the table; approximately rero in other estimated equations.

7 Santini were estimated for every equation. For short-run elastiities, the remaining disussion refers to the sum of the P and P 1 oeffiients (elastiities) for the equation for eah setor. These sums are given in Table 2, along with the separate Pi oeffiients. Long-run elastiities were estimated by an iterative proedure. The independent variable measuring long-run oil prie pressure (L) used the log of the following quantity: oil prie in a given year divided by oil prie in an earlier year. The interval between the injlial year and the earlier year was seleled by examining R 2 improvements and oeffiient t-values when the long-run variable was added to equations estimating short-run elastiities. Intervals as long as 7 and as short as 2 years were tested. The best interval proved to be 6 years. This interval was used throughout in the reported results. In addition to this iterative "optimization" proedure, the long-run prie variable itself was lagged from to 3 years (L, L 1, L 3 ) and a "best" lag was determined, also based on R 2 improvements and oeffiient t-values. These lags were 1 year eah for eletriity and transportation, 2 years for industry, and 3 years for residential/ommerial. The elastiity oeffiient for this longrun prie variable indiates the longer-run effet of a quasipermanent prie inrease, rather than the short-run shok effet indiated by the P-oeffiients. Other long-term lag formulations, suh as distributed lags, have not been tested at this time. Results of these experiments are given in Table 2. Eah regression model inluded a onstant term, whih is not reported. Short- and long-run elastiity estimates by setor are shown in Figure 6. The degree to whih substitution versus straight onservation is responsible for redued oil onsumption in a given setor an be approximated by omparing short- and long-run elastiities for total energy onsumption with those for oil-derived energy onsumption. The ratio of the oil prie elastiity of total energy onsumption to that for oil energy onsumption (times 1) approximates the perentage of the oil-energy-per-unit-of-output () redution ahieved by onservation. Subtrating this quantity from 1 approximates the perentage ahieved through substitution. The results of this latter substitution share approximation are shown at the bottom of Figure 6, diretly under the elastiity values from whih they were derived. The numbers of observations in these regressions are relatively small-smaller than those found in statistial tables for the Durbin-Watson d-statisti. On the basis of linear extrapolation of the d-values, appliation of the Durbin-Watson test (results not reported) did not allow onlusive determination of the presene or absene of autoorrelation. Beause of the small number of observations, the R 2 value presented is the adjusted value (3). The statistial results of these experiments tend to onfirm the present desriptive analysis of the relative diffiulty of substituting for oil in transportation. The estimates for both short- and long-run elastiity for transportation are substantially smaller in absolute magnitude than for the other setors (Table 2 and Figure 6). Further, supporting the earlier desriptive arguments, the perentage of redution in oil use estimated to have ourred through fuel substitution is far lower for transportation than for any other setor (Figure 6). These results also support the argument that there is a quite long lagged tehnologial response to sustained energy prie inreases, Elelrlity Industrial Short Long Short Long.OO T"""-R u n- -R u_n -Run -Run > ().2 C/J.3 "' IJJ t! () 4 ::i - <I> 2 a: Residential/ Commerial Short Run Long -Run Transportation Shorl -Run Long -Run - Long-Run Energy Consumption ::J Sheri-Run :::J Long-Run Estimated Elastiity of Energy or Oil Use per Unit of Output In Response to Short- and Long-Run Real Oil Prie Change Elelriily lnduslrial Residenlial/ Commerial Perenlage l Oil Use Redulin through Subs!l!ution - Shorl-Run - Long-Run Transporlalion Estimated Perentage of Redution of Oil Use per Unit of Output Ahieved through Substitution FIGURE 6 Setoral estimates of short- and long-run elastiity of oil and energy onsumption in response to real oil prie: omparison with estimates of substitutability for oil, whih takes hold only after years of effort to researh, develop, and introdue new and improved energy onversion tehnologies. LONG-TERM LIMITATIONS OF U.S. OIL SUPPLIES A fundamental problem faed by the United States is the delining domesti supply of eonomially reoverable oil (Figure 7). Sine 19, the growth rates of oil onsumption and prodution have moved in a lear, although slightly errati,... II> - ::.s::. -.._ (.:> < " o Consumption Growth Rate Prodution Growth Rate FIGURE 7 Deade averages of oil onsumption and prodution growth rates (history to 198, projetions to 25). 7

8 8 downward path. Ominously, in the 197s, domesti U.S. prodution "growth" rates turned negative. Although onsumption growth rates have not delined as rapidly as prodution growth rates, it is lear that onsumption growth rates have steadily delined in onert with domesti prodution limitations. The unsteady redution of onsumption growth rates until they shifted into deline in the early 198s was promoted by unsteady but lear long-term inreases in oil prie relative to pries of other goods (Figure 4). Domesti prodution of oil is learly onstrained by ost. The unpreedented prie inreases of rude oil of the 197s only redued the rate of deline of domesti prodution; they did not lead to an atual inrease in domesti prodution. Rates of onsumption growth, however, were turned into delines even greater than those in prodution. Thus the prodution-exess growth gap of the early 198s was far different than that of the early 193s (Figures 4 and 7). Reause growth rates of oil onsumption were onsistently above those of prodution during (Figure 4), imports were needed to make up the differene (Figure 8). On average, the differene between onsumption growth and prodution growth was relatively steady and slight during (Figures 4 and 9), leading to a slow inrease in the gross perentage of imported oil from about 13 to 21 perent. The United States found that imported oil was heap, and thus it maintained trade surpluses throughout the period; through 1968 this trend was readily manageable. Given the growth trends of oil prie and domesti onsumption and prodution for the period, this pattern probably ould have ontinued for many years. CAUSES OF THE RECENT GAP Regulations The steady trend of did not ontinue. Legislation was passed to improve safety in underground oal mines, most of whih were in high-sulfur eastern oal fields. At the same TRANSPORTATION RESEARCH RECORD 1175 time, sulfur dioxide emissions standards for power plants and industry were legislated and promulgated. The result was a sharp inrease in the ost of burning the dominant oal (i.e., high-sulfur eastern bituminous). This was refleted in a oal prie shok during Consistent with supply shok theory, a reession ourred in 197. Unable to bum oal leanly and eonomially under the new regulations, eletri utilities and industry began a massive swith to oil during (Figure 1). This led to an average inrease of more than 1 perent per year in the growth rate of oil onsumption during Absene of Transportation Influenes The reent gap was not, ontrary to popular opinion, reated by events in the transportation setor, where the growth rate of oil onsumption atually dropped below the average (Figure 1). This gap led to a sharp inrease in the need for, and the use of, imported oil (Figure 8)-espeially from the Middle East-making the United States highly suseptible to the imposition of an oil embargo. The ensuing rude oil prie rise (Figure 4) annot, aording to the logi of this paper, be attributed to anything other than the U.S. gap, given the absene of suh a gap worldwide (Figure 9). Oil Prodution Prodution of U.S. oil peaked during just as onsumption growth began its regulation-indued inrease. The ombination of inreasing onsumption growth and dereasing prodution growth reated a severe onsumption-exess growth gap in the United States (Figure 9) at the same time that free-world onsumption growth rates were atually below prodution growth rates. Oil Imports When growth of oil onsumption exeeds that of domesti oil prodution for a period of years, it is obvious that the extra oil -(II >. - CD :;::: 3.!!! VI.Q 2 a. E Prie Snok 2! :::J (II,,...,.... :'. -'. -'. -'.-'. > '.-: lnlmol 1 > :-:-: '.-> < : lnlmol 2 lntmol l -'. : lnlm I 4 '.- 1 U, : : : : : : : : : : : : : : : :.C.1_5.. :::::::::: : ::: : ::.(.1. :-.) : :. 1. '.9. :-.l.:: : _<1.. :: 'o ' ' ' o o ' ' o o ' " o I I I I > o ' o ' '" ' ' I o " o. o '. t t' o o' " '.' '. ' Snort-Term U.S. 11 Consumption Exess Growtn Gap Long-Term U.S. 11 Consumption Exess Growtn Gap FIGURE 8 Short- and long-term patterns of gross oil import shares In domesti U.S. oil onsumption,

9 Santini 9 "' C'l... > "O ;: a. '-" ':... (.!) := U.S. -- I ---- Non-U.S. WOCA I -5 -r ,-, l,_.., ::,/. u.. If) Cl u 'L: a.. <'I ::,/. u.. If)., u.: a.. -,-.--, FIGURE 9 Differene between growth rates of oil onsumption and prodution for United States and world outside ommunist areas (WOCA), 195(}-1982 (period averages). r--. '-" C'l.L:. u C'l... CD > < a :I < D l'.zl (Post Shok 1) l'iji'{1j (Post Shok 2) Eletri Utlities Industrial and Transportation Total Household- Commerial FIGURE 1 Changes in setoral rates of growth of oil onsumption, 195(}-1984 (period averages). must ome from elsewhere. Importing oil allows the gap to emerge, but it also makes the U.S. oil supply and eonomy more suseptible to events beyond its borders. Eah of the most severe U.S. oil prie run-ups in the 2th entury ourred when oil import perentages were at all-time highs (Figure 11). It appears that in both ases the market deided that the inreasing dependene on imports had to be reversed, beause the ensuing oil prie inreases led to subsequent sharp drops in the perentage of imported oil. INCREASING NEED TO DEVELOP OIL SUBSTITUTES IN TRANSPORTATION Role of Transportation In Overall Oil Consumption As mentioned, transportation was less responsible for initiating the gap than were other setors (Figure 1). Although the other setors did inrease their oil onsumption during , they were able to rapidly reverse these

10 1 TRANSPORTATION RESEARCH RECORD 1175 "tj ti t: 4 - a "--'--'----'.._ CDO 2 z ti "tj 2 1 u Notes: Import flguru not given before Refined produt. not Inluded....., Ill t:;t",, m ; &;!:@;;;;; :, ;;i ;h; ;:' H il i'llillli -t"""""".:.:.:.:1'""";.:.:.:.'""'"i':.:.:.:."""'""f".;.:.:.;:.:.:.:.:.;.:.:.:.:.:.:.:.:.:.jw:.:.:.:.:o::.:.:.:.:.:.:.:."'i""'.:.:.:.:"""":,:.:.: FIGURE 11 Comparison of net perentages of imported rude oil with a rude oil prie Index, inreases so that they ontributed far more to the losure and reversal of the gap than did transportation. This was possible mainly beause it was far easier for these setors to swith fuels than it was for transportation. Indeed, transportation's growth momentum and its inflexibility in fuel substitution and onservation aused this setor to onsume more oil in 1985 than in 1973, while eah of the other setors redued oil onsumption enough to ause an overall derease in onsumption (Figure 12). The long-term pattern in whih transportation aptures an inreasingly large share of the petroleum produts market ontinues unabated in the United States and other industrialized nations (Figures 1 and 2). These nations an therefore expet the inflexibility of transportation to be an even greater problem during the next gap. Further, beause the United States has been relatively suessful in finding substitutes for oil in nontransportation setors, it uses a far larger share of its petroleum produts for transportation than do its industrial ompetitors (Figure 2). Thus the United States has been, and will ontinue to be, less able to quikly redue its oil onsumption in response to a prie shok than have been, and will be, its industrialized ompetitors. Summary of Basi Arguments The fundamental arguments of this paper an be summarized as follows: e Beause of its inability to eonomially and rapidly substitute nonpetroleum fuels (Figure 6), transportation historially has been less able to redue petroleum onsumption than have other setors of industrialized eonomies (Figures 1 and 2). Beause of transportation's fuel inflexibility, industrialized nations that devoted a greater share of their total petroleum Total O O O O o o o + O O o o I o f I - J -2 - t ii Change In 11 Consumption by U.S. Setors, FIGURE 12 U.S. setoral hanges in oil onsumed per year, 1985 versus 1973 (In quads; 1 quad = quadrillion, or 1 15, Btu). onsumption to transportation had greater diffiulty reduing oil onsumption after the rude oil prie run-up (Table 1). e The United States devotes a far larger share of its petroleum onsumption to transportation than do its industrial ompetitors (Figure 2). e The United States is by far the most dominant transportation oil onsumer among OECD nations, aounting for nearly 6 perent of all OECD onsumption (Figure 3). e The long- and short-term evidene indiates that the trend in the United States is toward use of a greater share of petroleum produts by transportation (Figure 1).

11 Santini The short-term evidene indiates that other industrialized OECD nations are also devoting inreasing shares of their petroleum onsumption to transportation (Figure 2). It is therefore lear that when the next oil onsumption-exess growth gap appears likely, the suess or failure of U.S. efforts to redue onsumption by transportation will be more important than ever before in preventing or reversing the gap. Beause the suess of suh efforts in other setors has been greatly enhaned by those setors' ability to substitute nonpetroleum-based fuels, logi suggests that this apability must be developed in transportation if even more alamitous eonomi onsequenes than those of the early 198s are to be avoided. Need for Gradual Introdution of a Petroleum Substitute The problems with developing a substitute fuel for transportation are the high ost and the great amount of time needed. If it were not for the high ost, the prie run-ups of the 197s would undoubtedly have brought on the widespread introdution of substitute fuels in transportation. Beause many years are needed to smoothly implement new and more ostly transportation fuel tehnologies (witness the slow phase-in of unleaded gasoline)-whereas rude oil prie run-ups our in only a few years or even a single year-the free market annot respond "effiiently" to rude oil prie signals. Consequently, enlightened nonmarket intervention in the transportation fuels marketplae will be neessary if yet another severe prie runup and reession are to be avoided. The key word here is "enlightened," beause intervention in the fuels marketplae is the rule rather than the exeption. The influene of environmental and safety legislation on fuels markets is as old as the markets themselves. Earlier intervention, however, more often ontributed to the reation of gaps than to their elimination [see Santini (J) for a detailed disussion of regulatory influene on fuel markets]. Relatively rapid fuel substitution in U.S. transportation has ourred several times, but never within a period of only a few years. Typially, substitution aelerates during a time of unusually depressed business ativity. Major transportation fuel shifts ourred in the late 183s and early 184s, the 187s, the 189s, and the period (J). A U.S. depression of several years ourred during eah of these periods (J ). If the transition from leaded to unleaded gasoline was not major, then no major shift ourred in , but, as has been seen, the need for a major shift beomes ever more pronouned as transportation inreases its share of petroleum produts use and as the United States and ultimately the world run out of eonomially reoverable oil. Methanol: A Possible Mass-Market Substitute Fuel As the world runs out of oil, both long-term and reent history suggest that natural gas will be the most ommon soure of substitute transportation fuels (16, 17, 31). From 1973 to 1987, gas reserves in nonornmunist ountries inreased by 67 perent, while reserves of free-world oil inreased by only 9 perent (14). Long-term trends show gas apturing an everinreasing share of the fuels market. Natural gas was originally found in unwanted assoiation with rude oil and still is often flared at the well or reinjeted into the field beause of the lak of a market. Beause rude oil relatively near the surfae has been depleted, many deeper wells drilled to find oil have found gas instead. At greater depths, gas is more likely to be found than is rude oil (21). Thus, as oil reserves are depleted, a greater proportion of gas is found. Further, reent high pries for natural gas led to more searhes for gas for its own sake. Although natural gas an serve as a soure of gasoline, it is a more ost-effetive soure of methanol (18). (Both gasoline and methanol an also be produed from oal, but, again, methanol would be the heaper of the two--although more ostly than methanol from natural gas.) Thus a swith from gasoline to methanol would pay dividends to the United States for many deades beause oal derivatives are likely to be the transportation fuels of the far future, following (and overlapping) the era of natural gas. Although gasoline or a gasolinelike fuel might be produed from natural gas, a large body of researh is emerging in favor of methanol on both ost and environmental grounds (17, 18, 32). Methanol an be burned "neat" (pure) or "near-neat" in a vehile speially modified for the purpose. It ontains less energy per gallon than does gasoline, but a properly designed engine an burn it with greater thermodynami effiieny. Thus a methanol-powered vehile designed to provide the same horsepower per pound as a gasoline- or diesel-powered vehile would have a larger fuel tank but a smaller engine. The use of neat or near-neat methanol has the potential, aording to urrent EPA researh, of reduing smog in urban areas, an environmental goal that is proving diffiult to ahieve (32). One diffiulty with near-neat methanol is its poor old-start properties in urrent vehile designs. This problem has been addressed in Canada by government-sponsored researh and is now being investigated more aggressively in the United States. Methanol ould be gradually introdued into the gasoline market beause of its ability to boost the otane of gasoline; indeed, this is now being done by some refiners beause of federal requirements to remove lead from gasoline. Although methanol is still too expensive to burn as a neat fuel, it has a muh higher dollar value than gasoline when used as an otane booster. Among otane boosters that an replae the heap but environmentally unaeptable tetraethyl lead, methanol is quite heap (33). Used this way, however, methanol does not have the desirable environmental properties that it has as a neat fuel; thus, from an environmental standpoint, its ultimate use as a near-neat fuel remains desirable (32). Materials problems, suh as orrosion, related to use of methanol blends have largely been solved by automobile manufaturers, but many existing vehiles ould experiene materials-related problems if they used methanol blends. Other Possible Substitutes Although it is the judgment of this author that methanol is the universal fuel of the far future, other fuel substitutes are possible, espeially for seleted market nihes. Compressed natural gas, liquefied natural gas, liquid propane, ethanol, and "syntheti" fuels made from oal, oil shale, and tar sands are all tehnially suitable and oasionally eonomial transportation fuels under speifi, limited market onditions. It must be 11

12 12 stressed that small hanges in fuel supply an make large differenes in fuel pries, so even small substitutions of fuels other than those based on rude oil an have a strong effet in holding down rude oil pries. Given urrent ost estimates, however, these fuels are far less likely than methanol to serve satisfatorily as a universal, internationally traded, inexpensive standard fuel. Questions of Timing and Effort Level Given the ultimate need for an oil substitute in transportation, questions remain about probable Liming and effort levels of needed government and industry researh and development (R&D), government regulation, and industry investment and marketing. Most leading foreasters projet an aeleration of rude oil pries in the 199s, although they differ on the de.gree. of inrease (34). Even the Atlanti Rihfield Corporation (ARCO), whih antiipated sharp delines in real oil pries in three of its four publily released strategi planning senarios, projets rising real rude oil pries in the 199s in eah of those ases (Figure 13) (35). Unlike reent projetions for the United States by the U.S. Department of Energy (DOE), however, these projetions assume a "ap" on future oil pries below the 1981 level. ARCO has more experiene with methanol than any other major oil ompany and assumed that a methanollike fuel would ap oil pries. Chevron published far higher ost estimates but nevertheless alulated that the prie of methanol would be lower than some of DOE 's earlier projeted oil pries (36). A. J. Sobey of General Motors has ited the lowest ost, quoting $2 per barrel of oil "equivalent" from some non-u.s. soures of natural gas (17). Presumably, one differene between the ARCO and Chevron projetions is the assumed amount of ost-utting innovation that an be expeted in future methanol prodution. The Sobey $2 estimate relies on TRANSPORTATION RESEARCH RECORD 1175 qui re low wellhead gas osts a.ud on nearby soures. Reent osl estimates and levels of R&D effort imply that ost-utting innovation may be rapid (17, 37, 38). In one Data Resoures, In. (DR.I) projetion and in two DOE projetions, the osts of petroleum-related imports (produts and rude) in the late l 99 or early 21st entury exeed the value of the worst U.S. 1rade defiit in history (1986), as well as that of the worst U.S. oil trade defiit in history (198) (7). These projetions were of interest beause they illustrated the U.S. dilemma that an be expeted beause of the inevitable deline in oil output. If the United States is to redue its oil use and oil imports, the prie of oil must rise dramatially-thus ausing its oil import bill to rise. If the United States has the benefit of low-ost oil imports, it will produe less domestially and inrease its imports dramatially-thus ausing its oil import bill to rise. Methanol ould redue the seriousness of this situation in two ways. If methanol were produed from indigenous natural gas soures (more probable with high pries for world rude), import osts would be lowered largely by reduing import quantity. If methanol were provided through imports from lowost international suppliers, it ould redue total transportation fuel import osts by holding down the prie of imported fuel. The ability to readily substitute methanol for gasoline is neessary to ahieve the prie restraining effet. Both of these positive effets ould be put into plat:e with a arefully rafted methanol transition strategy developed through ooperating industrial and government organizations. The most optimisti (lowest oil prie, highest eonomi growth) senario onsidered by ARCO ("eonomi renaissane") apparently assumed suh a transition. ARCO assumed, among other things, "[A] long-term view, "Major suesses with... alternative energy, and "Pratial energy and physial environment poliies.",, 4 a> b 3 I senarios ' , FIGURE 13 Three of four published ARCO oil prie senarios developed for strategi planning.

13 Santini Given the onsensus of experts on pressures for an oil prie rise in the 199s, "long term" for the petroleum problem may well be less than a deade. Given the lessons of history and the results given in Table 2 and Figures 4-6, preventing a severe oil prie run-up by developing substitute transportation fuels will take several years at best. fu this paper, it is argued that the time to develop substitutes is when the gap beomes severe. Thus the time to aelerate efforts to develop suh substitutes (ounterintuitively, given reent delines in oil pries) is now. If early 1987 trends in U.S. oil onsumption and prodution are indiative of the full year's behavior, 1987 will have been a year during whih the United States one more opened up an "extreme, outlying" onsumption-in-exess-of-prodution gap as defined in the topmost portion of Figure 4. If 1988 beomes what 1987 was, the gap in 1988 will be as large as it was at its peak in the 197s. If this gap remains wide, a sharp oil prie rise in the 199s is quite likely indeed. CONCLUSIONS It has been argued, by examining petroleum market harateristis for the first 85 years of the entury, that a long-term view, in whih pratial energy and physial environment poliies are followed to ahieve major suesses with alternative energy, is ruial if the goals of steady eonomi growth and ompetitive suess in the international marketplae are to be ensured for the United States through the rest of the 2th entury. It has been shown that if these goals are to be met, a far greater emphasis must be plaed on introduing alternatives to petroleum in the transportation setor. ACKNOWLEDGMENT This work was done under ontrat with the Offie of the Assistant Seretary for Conservation and Renewable Energy, U.S. Department of Energy. REFERENCES 1. D. J. Santini. Mlro- and Maroeonomi Responses to Energy Prie Shoks: A Disussion of Past, Present, and Future Eonomi Problems in Simultaneously Introduing Alternative Fuels and Major Effiieny-Enhaning Vehiular Engine Innovations. Presented at the Operations Researh Soiety of Ameria and Institute of Management Sienes Joint National Meeting, St. Louis, Mo., Ot 25-28, E. G. Dolan. Maroeonomis, 4th ed. The Dryden Press, Chiago, Ill., C. R. MConnell. Eonomis: Priniples, Problems, and Poliies, 1th ed. MGraw-Hill, New York, G. A. Kahn. Dollar Depreiation and Inflation. Eonomi Review of the Federal Reserve Bank of Kansas City, Vol. 72, No. 9, Nov. 1987, pp D. J. Santini. Verifiation of Energy's Role as a Determinant of Maroeonomi Ativity. Pro., 21st lntersoiety Energy Conversion Engineering Conferene, Vol. l, 1986, pp D. J. Santini. Interations of Energy and the Maroeonomy-95 s of Evidene. Papers and Pro., 8th Annual North Amerian International Assoiation of Energy Eonomists' Conferene, May 1987, pp D. 1. Santini. The Petroleum Problem: Managing the Gap. Report ANL/CNSV-58. Argonne National Laboratory, Argonne, DI., Aug. 1986, Vol Investor's Daily, Nov. 18, Petroleum Data Book: Petroleum Industry Statistis. Amerian Petroleum Institute, Washington, D.C., various issues. 1. Monthly Energy Review. Energy Information Administration, U.S. Department of Energy, various issues. 11. E. E. Dunstan et al. The Siene of Petroleum. Oxford University Press, London, England, Energy Balane of OECD Countries, I Organisation for Eonomi Co-operation and Development, Paris, Frane, Historial StaJistis, Colonial Times to 197. U.S. Department of Comm re Energy Statistis Sourebook. PennWell Publishing Co., Tulsa, Okla., H. F. Williamson and A. R. Daum The Amerian Petroleum Industry: Tlie Age of Energy, Northwestern University Press, Evanston, Ill., A. J. Sobey. Eonomis of Mandated Automotive Fuel Eonomy S1andards. Presented at 64th Annual Meeting of the Transportation Researh Board, Washington. D.C.. Jan A. J. Sobey. The Automobile Manufaturers' View of the Use of Compressed Natural Gas. Pro., On The Road with Natural Gas Conferene, Amerian Gas Assoiation, Washington, D.C., Swedish Motor Fuel Tehnology Co. Alohols and Alohol Blends as Motor Fuels. In1emational Energy Ageny, Paris, Frane, 1986, Vols. I, IIA, and lid. 19. C. Gray and J. Alson. Moving Ameria to Methanol. Vniversity of Mihigan Press, Ann Arbor, R. Bright. The Ford Flexible Fuel Vehile. Presented at Third Windsor Workshop on Alternative Fuels, Windsor, Ontario, Canada, June 24-26, B. Mossavar-Rahrnani and S. Mossavar-Rahmani. The OPEC Natural Gas Dilemma. Westview Press, Boulder, Colo., B. D. MNutt and E. E. Eklund. Is There a Government Role in Methanol Market Development? Soiety of Automotive Engineers Tehnial Paper Presented at International Fuels and Lubriants Meeting and Exposition, Philadelphia, Pa., Ot 6-9, J. Dunkerley and I. Hoh. The Priing of Transport Fuels. Energy Poliy, Aug. 1986, pp J. J. Amdall. Initial Operating Results from Caterpillar 36 Diesel Engine. Amerian Soiety of Mehanial Engineers Paper 87- ICE-3. Presented at Energy-Soures Tehnology Conferene and Exhibition, Dallas, Tex. Feb. 15-2, R. Sekar and L. Tozzi. Advaned Automotive Diesel Assessment Program. NASA-Lewis Report DOE/NASA/261-1 NASA CR CTR Lewis Researh Center, National Aeronautis and Spae Administration, Cleveland, Ohio, De C. F. Rossi. Researh in the Automotive Industry. Pro., Institution of Mehanial Engineers, 1986, pp Statistial Abstrat of the United States. Bureau of the Census, U.S. Department of Commere, various issues. 28. Survey of Current Business. Bureau of Eonomi Analysis, U.S. Department of Commere, Sept Highway Statistis: Summary to FHWA, U.S. Department of Transportation, W. C. Merrill and K. A. Fox. Introdution to Eonomi StaJistis. John Wiley & Sons, New York, C. Marhe1ti. The Future of Natural Gas: A Darwinian Analysis. Tehnologial Foreasting and Soial Change, Vol. 31, 1987, pp R. D. Wilson. New Fuels for Cleaner Air. Presented at Alohol Week/Inside EPA Conferene on New Fuels for Cleaner Air, Washington, D.C., July , G. H. Unzelman and G. W. Mie.halski Otane Improvement Update-Refinery Proessing, Antiknoks, and Oxygenaies. Presen1d at 1984 National Pell'oleum Refiners Assoiation Meetings, San Antonio, Tex., Marh 25-27,