The full text on this page is automatically extracted from the file linked above and may contain errors and inconsistencies.
October State and Local Government Activities in the Tenth District . . . . . . . . page 1960 3 Recent Adjustments in Petroleum Refining . . . . . . . page 10 Current Charts and Statistics . . . . . page 16 FEDERAL RESERVE BANK OF KANSAS CITY Subscriptions to the Mo TIILY REvrnw are available to the public without charge. Additional copies of any issue m,ay be obtained from the Research Department, Federal Reserve Bank of Kansas City, Kansas City 6, Missouri. Permission is granted to reproduce any material in this publication. State and Local Government Activities in the Tenth District homa, and Wyoming. 1 Most of the basic data used in this article were obtained from the U. S. Census of Governments: 1957-the most comprehensive recent source of information on state and local government activities. Education I THE AucusT 1960 Monthly Review, some aspects of the significance of state and local governments in the national econ01:1y were considerccl. ln addition, the kvel of genera] stale and local gov<'rnmcnl cxpc11ditu1Ts was compared among six stales in the Tenth Federal He.serve Distrid and the United States. State and local government outlays per person in District states were found generally to be high relative to the national average and to vary considerably among the states. Disparity in state per capita personal income levels was found to be associated in a general way with differences in per capita state and local expenditures, both in the Nation as a whole and within the District. However, income differences did not explain the higher-than-national average level of outlays in District states, nor did they account fully for discrepancies among the states. In the present article, rnstrict per capita state and local government activities are viewed in terms of the major functions performed by such governments. This is done not only because information on these functions is of interest in its own right, but also because it permits a clearer identification of the fa:t~rs associated with differing levels of activity among states and regions. While the experience of governments across the Nation is often looked to for factors which influence state and local government functions, the primary focus is upon six District states-Colorado, Kansas, Nebraska, New Mexico, OklaN Monthly Review • October 1960 Across the Nation, education accounts for more variation in per capita state and local government expenditures than any other function, despite the fact that ontlays for it are relatively kss variable from slate lo state than those for most other major functions. This apparent anomaly results from the large absolute size of education expenditures. Nationally, they accounted for around 35 per cent of general state and local government outlays in 1957 - an amount equal to about 3 per cent of gross national product. Among the District states, education outlays were responsible for 34 to 39 per cent of the total. Despite this comparatively small variation in the proportion of general outlays devoted to education, differences in per capita dollar amounts were large. Only highway expenditures, among the other categories of outlays, showed about as large a range in the District. Indeed, the range of education expenditures was far greater than the total spent for any other function except highways and, in the case of two states, public welfare. 1 WhilP a significant share of the population of the Tenth District resides in Missouri, the largest influence upon state and local expenditures in that state is beyond the hounds of the District. Unfortunately, it is not possible to make meaningful comparisons of state and local activities in the District portion of Missouri with other District states. A number of conceptual and statistical problems involved in making interstate comparison of state and local government expenditures are discussed in the August 1960 Monthly Review article. 3 State and Local Government Activities Nationally, interstate differences in per capita education expenditures display considerable association with variation in per capita incomes. However, high per capita expenditures are not necessarily always associated with a proportionately larger amount of education services. Costs of such services differ from place to place and a large clement in this disparity is variation in school salaries. Ordinarily well over half of local school expenditures are for payroJls, of which by far the largest portion goes for teacher salaries. These salaries are comparatively about as variable as outlays and relatively more variable than the salari<'s or all ollwr public <'lllploy<'cs. ationally as lllight l><' c,pcdcd teacher '>alarics arc closely cm-r<.·latcd with per capita i11conws. Thcrdore, the national relationship hetw(•cn per capita incomes and per capita expenditures for education is to some degree attributable to teacher salaries. Since such a large share of education outlays is for payrolls, school employment provides one measure-albeit a very crude one-of the actual level of services rendered. Per Person Education Activity High in District \Vhethcr measured in terms of employment or - with one cxc('ption - in terms of outlays, each District state purchases more education services per person than the national average. They do this despite the fact that their per capita incomes arc generally below the national average. However, several factors can be identified which indicate that the "need" for education services in District states is unusually large. Possibly the most obvious of these is the high proportion of the population enrolled in elementary and secondary schools. In this regard, each state-except Nchraskacxcceds national cxp('ricncc hy a su hstantial margin. This is in considerable measure attrihutahlc to an age pattern which in most instances contains an unusually great number of school age persons. An additional factor which tends to raise demands upon Dish·ict 4 education systems is comparatively low density of population. While at least one recent study has indicated that economies of enlarged size are rather quickly exhausted in education plants, extremely low density probably does contribute somewhat to inefficient plant size. Furthermore, the cost of assembling the student body tends to be high in a sparsely settled area. The high level of per capita education outlays in the District is partly accounted for by advanced ( college and university) education upon which each state spends more than the national average. This is true of both total outlays and of operating <'Xp<'nclitmcs. Inch-cl, i11 19S7, Colorado spent 011ly slightly ll'ss pn capita for higher education than the lop stat<.' in the Nation a11cl in the same year New Mexico incurred higher per capita current operating outlays than any other state. Each of the District states also employed more people in higher education relative to population than the United States average, with New Mexico ranking second in the Nation. While various reasons for this play a part in each of the states, at least one factor is common to all of them. In each case the percentage of the population enrolled in public institutions of higher cclucation exceeds that in the Nation generally. This results both from high over-all advanced education enrollment and from the fact that a large percentage of it is in public institutions. In some states outside the District, where advanced education expenditures and employment are comparatively low, a large portion of enrollments is in private institutions. This is especially notable in several New England states. In som District states-Colorado is a striking example - higher education enro1lments are inflated by unusually large numbers of nonresident students. In terms of financing requirements, large enrollments of such students are at least in part offset by the comparatively high nonresident fees they pay. in the Tenth District Table 1 Education Activities In District States In 1957 General Expenditures Per Capita All Functions Rank Wyoming New Mexico Colorado Kansas Oklahoma Nebraska U. S. Average $ 328 279 280 271 248 202 237 1 3 2 4 5 6 Education Expenditu res Per Capita $ 120 109 102 92 88 78 83 Rank 1 2 3 4 5 6 Per Capita Expenditures for Local Schools $ 94 82 76 73 67 61 70 Rank 1 2 3 4 5 6 Per Capita Capital Outlays for Education $ 25 19 25 19 18 14 19 Rank 1-2 3-4 1-2 3-4 5 6 Enrollment in Local Schools Per Full-Time Education Employee 15 17 16 15 17 15 17 SOURCE : U. S. Census of Governments: 1957. Local School Expenditures in Individual States The number of public elementary and secondary school students per education employee dicl not differ greatly from stale to slate in the District in HJ.57. Thus, in looking at differc.·uces in local education expenditures among the individual District states, an interesting question presents itself. Why do per capita education expenditures at the local level vary so much when the amount of educational services rendered per pupil-measured by the rough standard of employment-is for the most part not radically different? In each state a pattern can be discerned which probably goes a long way toward an answer. Wyoming, which in 1957 ranked next to the highest in per capita local school expenditures nationally and highest in the District, did so largely as the consequence of a series of reinforcing factors. To begin with, the state has a high per capita income and the rate of pay for teachers is correspondingly high. In addition, a large percentage of the state's population is enrolled in local public schools and comparatively rapid growth in enrollment has given rise to large capital outlays. Furthermore, Wyoming is among the more sparsely populated states in the union. The latter fact may well help to account for a somewhat smaller number of pupils per employee than in other District states. In addition, low density tends to increase per pupil costs incurred outside the classroom. Monthly Review • October 1960 Nebraska, at the other end of the expenditure spectrum , presents a more complicated picture. Its per capita ineom<' is around the average for the District hut per capita expenditures for education in 19.57 W<'rc ,wll l)('low even those in Oklahoma - the lowest income state. Partly accounting for this is the fa ct that enrollment of students in primary and secondary schools was substantially lower in regard to population than elsewhere in the District. Similarly, the median pay of teachers in Nebraska was well below that of any other District state. Reinforcing these factors was a comparatively low rate of increase in enrollment, with correspondingly small capital outlays per person. This combination of factors helped lo produce low per person education expenditures despite a relatively high ratio of employees to students in elementary and secondary schools. New Mexico also presents an interesting case. In 1957 its per capita income was near the lowest in the District and ranked 31st in the Nation. Despite this, more was spent per person for local schools than in any other District state except Wyoming and more than in all hut six other states in the Nation. Important elements in the explanation for this situation were New Mexico's large student enrollment relative to population, which was greater than in any other District state, and the fact that it ranks next only to Wyoming in low density of population. However, an im- s State and Local Government Activities portant additional aspect is the fact that the median pay of teachers in New Mexico was well above that of any other District state. This is rather puzzling in light of New Mexico's comparatively low per capita income. However, it may well be related to another factor which helps to explain the state's high expenditures per person, i.e., the rapid increase in school enrollment in recent years. This growth required the state to engage substantially in capital outlays and may have made it necessary, as well, to attract teachers from outside the state. Colorado has a smaller percentage of its population enrolled in local schools and is more densely populated than New Mexico or Wyoming- the District states which cxcccclecl it in per capita outlays. On the other hand , elementary and secondary school enrollment in Colorado has risen faster than in any other District state. This was reflected in unusually heavy capital outlays for local schools. The balance of these factors combined with somewhat lower teacher salaries than in New Mexico and Wyoming resulted in local school outlays below the level of the other Mountain States but above the remaining District states. Kansas and Oklahoma- whose 1oca1 school outlays fell well below those of Wyoming, New Mexico, and Colorado, but were above the national average-have several features in common. Both have a low density of population by national standards but considerably higher than any of the other District states. While school enrollment in Oklahoma has tended to decline somewhat in recent years and Kansas has shown some increase, the rise in the latter was far less than that experienced in the Mountain States. At least partly as a consequence, neither spent as much per capita for capital outlays in 1957 as did the Mountain States. In addition, Kansas and Oklahoma differed relatively little in median annual teacher salaries and were well below the Mountain States. In some respects, how- 6 ever, they were highly dissimilar. While Kansas was about equal to the District average in terms of per capita income in 1957, Oklahoma fell considerably below any of the other states. On the other hand, Oklahoma ranked comparatively high in enrollment relative to population while Kansas fell near the District low. Furthermore, Oklahoma had one of the highest ratios of students to full-time education employees in the District, while Kansas ranked much lower in terms of this measure. The various factors influencing education expenditures in the two states resulted in a slightly higher per capita outlay in Kansas than in Oklahoma. Highways Spending on highway construction and maintenance accounts for roughly 20 per cent of the general outlays of state and local governments in the United States and typically for a somewhat higher percentage in the District. In 1957, per capita highway expenditures ranged from $25 in South Carolina to $107 in Connecticut and they were relatively the most variable of the major categories of state and local outlays. A high degree of variability is not surprising, perhaps, since over two thirds of highway exp enditures were for capital outlays and of these more than four fifths were for construction. Construction tends to be undertaken in large projects and, therefore, shows considerable variation, not only from state to state but also at different times in the same state. In 1957, variation in highway expenditures in District states, on balance, tended to reinforce the differences in per capita expenditures attributable to education. Consequently, by far the greatest portion of total variation in per capita outlays was du e to those two functions. Variability and the important role played hy Federal financing tend to obscure whatever correspondence there may be b etween highway expenditures and some of the factors that in the Tenth District Table 2 Per Capita State and Local Government Expenditures On Functions Other Than Education In 1957 Wyoming Kansas New Mexico Colorado Oklahoma Nebraska U. S. Average Highways Rank $ 93 81 66 54 54 51 46 1 Public Welfare 2 3 4-5 4-5 6 $ 17 22 20 44 46 14 20 Rank 5 3 4 2 1 6 Health and Hospitals $ 25 18 17 17 12 14 19 Rank 1 2 3-4 3-4 6 5 Urban Type Services $ 17 20 20 25 17 16 30 Rank Other Outlays 4-5 2-3 2-3 1 4-5 6 $ 58 39 46 39 31 29 40 Rank 1 4 2 3 5 6 SOURCE: U. S. Census of Governments: 1957. might be expected to influence them. For example, in 1957 there did not appear to be any straightforward relationship between per capita income and highway expenditures across the Nation other than near the ]ow end of the state income <lis trihulion, where difficulties in obtaining matching funds may be particularly important. Tl1e District states fell rather conveniently into three classes in regard to their per capita highway expenditures in 1957. Wyoming and Kansas were considerably above the other states and the national level. New Mexico fell somewhat lower but remained well above the national average. Colorado, Oklahoma, and Nebraska diverged little and were more moderately above the national average. While individual state expenditure levels are associated with specific and often transitory factors, a generally higher-than-national level of highway activities per capita has been characteristic of District states in recent years. This is true whether activities are measured in terms of operating expenditures, capital outlays, or employment. To a large degree the high level of activity probably is associated with less-than-average density of population combined with a location athwart major eastwest throughways. Thus, the number of persons per unit of surfaced road is much lower in each of the District states than in the Nation generally and the number of motor vehicles relative to population is appreciably higher. In addition, a significant portion of Monthly Review • October 1960 the District's highways are in mountainous terrain where costs of maintenance and construction are high. All these factors were reflected in Wyoming's unusually great - next to the highest in the Nation- per capita highway expenditures in 1957. For the most part, similar factors are involved in accounting for New Mexico's high rate of expenditures. Furthermore, both New Mexico and Wyoming were among the early large participants in highway funds allotted under the Federal Aid Highway Act of 1956. On the other hand, the high rate of spending in Kansas in 1957 cannot be accounted for in this manner. By District standards, Kansas is not particularly sparsely populated, Federal participation was not extraordinarily large, and the state's highways are not in unusually difficult terrain. The major reason for the state's high outlay was a heavy investment in the state-operated Kansas Turnpike in that year. Oklahoma also made a large toll road investment-in the Will Rogers Turnpike-but that in Kansas was larger, accounting for almost a third of total highway expenditures. Other State and Local Government Activities The functions already discussed, plus expenditures for public welfare and health and hospitals, account for about 70 per cent of all state and local government outlays. Outlays on the latter two categories are on the average of comparable size and considerably smaller than those for education and highways. The remaining expenditures are for provision of 7 State and local Government Activities a number of services associated with urban areas and for a variety of other purposes. Public Welfare Public welfare expenditures show virtually no positive relationship to per capita income. This is illustrated by the fact that in 1957 per capita outlays for this function ranged from about $7 per p rson in Virginia to about $46 in Oklahoma, both of which stand quit low on the scale of per capita incomes. Lack of association with income is the joint consequence of a variety of factors. For one thing, the Federal Government participat s unusuaJly heavily in th' w •Hare field. ationally, almost half of stat' and local expenditures on wdfar , have hecn finaHc cl from Federal con tributions in re '<'nt y<'ars. S ·cond, at least in some of the highest 'xp 'nditure stat 'S, lib •ral old-age assistance influences xp nditures heavily and per capita outlays on this function show no positive relation to income. While old-age assistance is, of course, related to the age structure of the population, there is in addition wide variation from state to state across the Nation, not only in payments per beneficiary but also in the proportion of the population over 65 r ceiving assistance. This variation r sults from diff renc s in law , interpretation of laws, and th in om' status of the aged. Third, whil high income is oft n associated with urbanization which, tak n by itself, tends to increas welfare case loads, high income as such would probably tend to reduce them. Consequently, the lowest per capita public welfare costs might be expected in high income states with low urbanization, relatively youthful population, and low oldage assistance payments. Thus Wyoming, which di plays most of the features that might be xpected to produ e moderate levels of w Hare paym nts, ranks near the District low in per capita outlays for the function. This is particularly impr s ive in light of the fact that the state has rank d highest in both of the other functions thus far 8 considered. On the other hand, the influence of old-age assis tance payments on per capita w lfar outlays is notable in Oklahoma, Colorado, and Kansas, but most striking in the former two. Oklahoma, which ranked lowest in per capita income among the District states and 36th in th Unit d States, spent more per capita on public w lfar in 1957 than any other state in the ation. While aid to cl pendent children is quite a large welfare factor in the state, about three fourths of the total expenditure on welfare was for old-age assistanc . Colorado, though spending less per capita on old -age assistance, used an even largc•r portion of its public W<'lfare huclgct for th e purpose•. Colorado ranks ('Otnp,1ralivcly high in th<' Nalioll , both in terms of benefits paid per rec:ipi<'nt aiid case load relative to population. On the oth 'r hand, Oklahoma is somewhat less generous in awarding benefits but its case load is strikingly high both in regard to total population and in relation to the older members of the population. The other District states- ebraska, New Mexico, and Wyoming-differed comparatively little in per capita old-age assistance payments in 1957 and were som what below the national m an. In regard to old-age assi tanc and ducation expenditures, it i interesting to not that an age distribution of the population which tends to r quir' hug' expenditure for these purposes is also one which, other things being equal, tends to depress income. A recent study of state income differentials found that the proportion of the population above and below working age bears a significant negative relation to state per capita income. Thus, greater tax effort may be requir d in some states exhibiting th sc age haract ristics not only h cause ne('ds arc greater but also h caus the ability to m t them i 1 s. Health and Hospitals Health and hospital xpenditures ar financed almost exclusiv ly from state and local revenue sources, since F deral funds gener- in the Tenth District ally account for less than 5 per cent of the outlays for this function. At least partly as a consequence, per capita health and hospital expenditures are rather closely associated with income across the Nation. At first glance, such an association does not appear to exist among the District states. While Wyoming-with the highest incomedoes spend distinctly the largest amount, and Oklahoma- the lowest income state-spends the least, outlays by the other District states do not show a discernible relationship to income. However, if current operating costs are considered separately, a more distinct association appears . For example', hoth Kansas and New Mexico, which stood higher than OIH' would ex1wct on the basis of income, undertook large per capita capital outlays in 1957, and Colorado, which ranked lower than would have been expected, undertook relatively little investment in that year. lays of the District states. In most instances, per capita expenditures on these functions are comparatively small and, therefore, do not differ by substantial amounts from state to state. The most variable and perhaps the most interesting-from the point of view of the District-of the remaining outlays are those for natural resource activities. Each District state spends more per capita than the national average in the resources area, with the highest outlays occurring in states where natural resources assume a particularly important role. Other Functions range is of the median. The coefficients for the major types of state and local government expenditures across the Nat ion arc as follows: total education 17.6; highways 21.2; public welfare 1.15.0; health and hospitals 20.0; police and local fire protection 24.4; sanitation 30.6; and natural resources 49.1. Concluding Statements The foregoing analysis of the factors which appear to be associated with differing levels of per capita state and local activities cannot, of course, claim to he complete. For example, specific institutional arrangements in the various states may play an important role in limiting or accentuating expenditures. Mat"Urban Type" Services ters such as earmarked taxes, debt limits, and A group of expenditures which can be · property tax limits may lag in adapting to rather conveniently considered together are changed circumstances. To a degree, concepfor what might be termed "urban type" servtions of acceptable levels of services differ ices. This category includes expenditures for from state to state. These matters no doubt police and local fire protection, sanitation, play a significant- but hard to quantify-role local parks and recreation, housing and comin determining activity levels. Nevertheless, munity redevelopment, and local libraries. the factors which have been cited in this Since these arc services ordinarily rendered study are ones which previous investigations within urban areas, it is not surprising to find and perusal of the 1957 Census suggest are them smaller in the District states than in the significantly related to expenditure differNation generally. Per capita outlays for these ences at a national level and in most instances purposes were less than the national average they form a logical pattern in regard to the in each District state in 1957. Relative to other District states. types of expenditures they were even more TECHNICAL NOTE: In several places in the text, markedly below the national average. In the reference is made to the relative variation displayed District, outlays for these items ran between by particular variables. The measure used to indicate 5 and 9 per cent of the total, whereas nationrelative variation is a coefficient determined by computing the percC"ntage which the semi-interquartile ally they constituted over 12 per cent. The remaining expenditures of state and local governments are for a miscellaneous group of functions which, in 1957, absorbed between 12 and 17 per cent of the total outMonthly Review • October 1960 9 Recent Adiustments in Petroleum Refining has been centered recently on petroleum refining as a result of the dramatic turnabout of conditions in that industry. After enjoying unprecedented prosperity in late 1956 and early 1957, petroleum refineries have since operated under considerably less auspicious circumstances. Whil · developments in the industry unc.louhtedly reflected the impact of the recession in 1958, there is some evidence that the problems facing the industry have persisted during the business recovery. For example, high levels of excess capacity and burdensome stocks of refined products still persist. This situation reflects comparatively small increases in demand for refined products and continued increases in total crude oil refining capacity. With these factors in mind, this article has several objectives. First, developments in the demand for refined petroleum products and changes in refining capacity in recent years are described. Second, inducements to invest arising out of changes in the nature of demand for petroleum products and technological developments in the industry are discussed. Third, the current level of capacity utilization is analyzed in terms of more prosperous years in the petroleum industry. Finally, some explanation of the continuing expansion of refining facilities in light of the low rate of utilization of existing capacity is undertaken. While this discussion is of direct relevance to the petroleum industry, it may also have some implications for other industries where capacity is more than adequate. Recent low rates of utilization in some manufacturing industries have given added interest to analysis of ONSIDERABLE INTEREST C 10 investment under conditions of excess capacity. Although the factors responsible for the problems facing the petroleum industry have become more evident in the past few years, some of them appear to have been developing for a considerable period. The influence of these factors was delayed and obs urecl by tlw occu1T('nce of extraordinary circumstances, such as the Korean War, the oil shutdown in Iran, and the closing of the Suez Canal. For the purpose of describing the turnabout of conditions in the petroleum industry, developments from 1946 to 1956 have been contrasted with those from 1956 through 1959. In the decade following the end of World War II, demand for refined products grew at a relatively rapid rate. Since then, the industry has had to cope with a number of influences - some short run and others reflecting longer-term factors - which have tended to depress demand growth. The increase in overall demand declined from an average annual rate of 6 per cent in the postwar decade to an average rate of less than 2 per cent in the past 3 years. Changes in the Rate of Increase and Structure of Demand Several factors have caused demand for refined products to grow at a slower rate in the past 3 years than in the previous decade. The sharp dip in business activity which reached a ]ow in 1958 undoubtedly had an adverse eHect on demand. Total energy consumption in the United States-measured in BTU equivalents of coal, petroleum, natural Recent Adjustments in Petroleum Refining gas, and water power-declined in both 1957 and 1958. This was the first time in the postwar period that energy consumption had declined 2 consecutive years. In the decade after World War II, growth of this measure of aggregate energy consumption proceeded at an average rate of 3.4 per cent annually. While information for 1959 is not available, energy consumption undoubtedly increased to some degree, as previous movements in this series have conformed roughly to movements in gross national product. Factors other than short-run business conditions in the U nilccl States were responsible, al least in part , for the slower rate of growth of demand for refined procl11cts sine<' H)56. CNlainly one of the most widely fluctuating influences on total demand was exports of refined products. After increasing at an average annual rate of 3.4 per cent from 1946 to 1956, exports registered an annual decline of 9.5 per cent on the average from 1956 to 1959. Exports of refined products in 1959 were at the lowest point since 1950. Thus, by 1959 foreign demand had sustained a sharp decline from the 1957 peak as the abnormal conditions responsible for the stimulation of demand during the Suez crisis swiftly moderated. On the other hand, longer-term developments in domestic demand also have been significant, although somewhat less striking than the reversal in exports. During the period from 1946 to 1956, consumption of petroleum products-excluding natural gas-increased at a rate of 6 per cent annually in contrast to a 3.4 per cent annual increase in total energy consumption. As a result, consumption of petroleum products accounted for 44 per cent of the energy market in 1956 as opposed to 34 per cent 10 years earlier. While there was no reason to conclude that the portion of the energy market which petroleum had captured by 1956 was the maximum penetration it would ever achieve, there was some reason to believe future gains at Monthly Review • October 1960 the expense of other fuels might be at a moderated rate. It does not appear likely that refined petroleum products will ever completely displace other energy sources, and indications are that the process of substituting petroleum products for coal in household heating and as railroad fuel have been largely completed. In addition, the consumption of natural gas has shown particularly impressive gains in recent years and these gains have been, at least in part, at the expense of refined products. While increasing at an annual rate of more than 9 per cent from 1946 to 1956, natural gas consumption increased its share of the total energy market from 13 to 23 per ('<'Ill. Further gains of nearly 6 per cent in the consumption of this f ucl W('rc n•corclcd in both HJ.57 and 19,58, increasing the market penetration of natural gas to nearly 27 per cent. In part, this may have reflected the lesser sensitivity of natural gas to business cycles, but the large gains during recession years were quite impressive. In contrast, the rate of increase in domestic demand for refined petroleum products in the 3 years after 1956 averaged less than 2½ per cent. Components of Domestic Demand Understanding of changes in aggregate domestic demand for refined petroleum products is aided by an examination of trends in its various components. Over the entire postwar period the demand for gasoliue has increased roughly in line with total domestic demand for refined petroleum products. In 1946, gasoline consumption accounted for 41 per cent of domestic demand, while in 1959 it constituted 43 per cent of the total. The average rate of increase in consumption of this fuel has slowed considerably, however. A 6.5 per cent annual rate in the 1946-56 period slipped to a rate of 2.5 per cent in the past 3 years. During the entire period, on the other hand, there has been a continuing increase in octane ratings of gasoline for automobile use. 11 Recent Adjustments Demand For Refined Petroleum Products MILLION BARRELS 8000 SEMI-LOG SCALE 6000 4000 2000 1000 800 600L---~,,,,,,----::::::::.-----_:-_~;_;_::..=.;:::~ 400 200 100 80 60 40 20 * * EARLIER DATA NOT AVAILABLE NOTE : On a semi -logar i thmic chart, equal rates of change . SOURCE : U. S. Bureau of Mines . slopes indicate equal Similarly, distillate fuel experienced a reduction in the rate of growth in demand from an annual rate of 10 per cent in the postwar decade to a 2.3 per cent rate from 1956 to 1959. Consumption of this fuel, which is used in diesel engines of trucks, buses, and railway locomotives and for heating private homes, was accelerated by the displacement of steam locomotives and the sharp increase in the number of diesel trucks on U. S. highways after 1946. Since these transitions have been largely completed, future gains may be more moderate than the 1946-56 average. In 1946 no information was available on jet fuel consumption. Since that information became available in 1952, however, consumption has 12 increased by more than four times. While kerosene is an important component of jet fuel and has experienced increased demand as a result of gains in consumption of that fuel, other demands for kerosene have remained generally stable in the postwar period. In contrast to these relatively rapid rates of growth, demand for residual fuel oil increased at an average annual rate of only 1.3 per cent since 1946. A major reason for the relatively slow rate of increase in consumption of residual fuel oil was the intense competition this fu el faced from natural gas, since natural gas is probably a better substitute for residual fu el oil than for other refined p etroleum products. r11 broad perspective, the major change in the composition of demand since 1946 has b een the rapid rise in importance of the lighter end products. These high-value products which include gasoline, kerosene, distillate fuel oil, and jet fuel have assumed a greater relative importance during this period as their share of domestic demand for refined products has increased from 60 per cent in 1946 to 68 per cent in 1959. This shift in the composition of demand has had an important impact on existing refineries and on new plants in the industry. Capital Expenditures and Additions to Refining Capacity Investment in the Petroleum Industry In the light of changes in the structure and growth of demand for refined products, it is interesting to trace the capital expenditures which have given rise to the present surplus of refining capacity. After World War II, plant and equipment expenditures for manufacturers of petroleum and coal products increased, reaching a peak in 1948 and subsequently declining for 2 years. Shortly after the beginning of the Korean War in 1950, plant and equipment expenditures were stimulated by high demand and the program of in Petroleum Refining the D efense Production Administration which allowed accelerated amortization of new facilities. Expenditures for plant and equipment continued to increase after the end of the conflict and reached a peak in 1957. Expenditures in 1958 were down sharply from the year-earlier level and, although moderate increases occurred in 1959 and the first two quarters of 1960, anticipated expenditures for 1960 were still one fourth below the 1957 level. Nevertheless, recent levels of investment have been sufficiently large to support continuing additions to basic refining capacity. While increases in refining capacity are certainly relat<•d to plant and equipment expenditures in the petroleum industry, tlw n'lationship is by no means simple or fixed. A portion of investment outlays- the magnitude of which may vary from time to time and is not precisely known-goes for replacement. In addition, capital expenditures by petroleum companies are partly channeled into exploration, crude production, or distribution. In fact, as a result of the high level of excess capacity in the past few years, companies have indicated they are concentrating relatively more of total capital expenditures on distribution. Moreover, investment in new equipment motivated by efforts to reduce costs does not necessarily result in capacity expansion. In addition, the recent trend toward increased octane ratings for gasoline and higher yields of light end products has required extensive investment aimed at increasing the quality and yield of certain products but not necessarily resulting in greater throughput capacity. The basic distillation process involved in refining al1ows no opportunity for varying the proportions in which the components of a given crude are divided, although different crudes yield different proportions of the various products. By means of cracking, however, the yield of the various components of crude oil can be modified to conMonthly Review • October 1960 form more closely to demand. Cracking is a means of increasing the yield and quality of light fuels by breaking large molecules into sma1ler ones. This accounts for the fact that catalytic cracking capacity increased nearly three times from 1946 to 1957 and reforming capacity increased more than 21/~ times from 1954 to 1959. In contrast, thermal cracking capacity has declined slowly since 1946 as certain obsolete thermal processes have been scrapped. The modification, replacement, and expansion of capacity has, in recent years, taken place at existing plants. As a result, most refin<'ri<'s consist of a conglomeration of old a11d new equipment. In this light, it is interesting to note that, while total capacity was expanded by almost 90 per ce11t during the postwar period, the number of refineries in operation declined by 20 per cent. Additions to Capacity It is apparent that inducements to invest in refining plant and equipment have arisen from the need to modernize plants, to develop capacity for increasing the percentage yield of light end products, and to replace aging equipment. Nevertheless, as mentioned earlier, part of the investment by petroleum companies in the last several years has taken the form of increases in basic crt.de oil refining capacity. Basic crude oil refining capacity is defined by the U. S. Bureau of Mines as the maximum average throughput attainable with allowance for normal downtime for repairs and routine maintenance. Total capacity is divided into both operating and shutdown components. Since operating capacity is more sensitive to relatively short-term developmen ts than total capacity, it is more suitable for the purposes of the article. Hence, subsequent referenc s to capacity will pertain to operating capacity. From January 1, 1946, to January 1, 1957, crude oil charging capacity increased at an 13 Recent Adjustments annual rate of 5.2 per cent. In the subsequent 3 years-despite low rates of utilization-capacity continued to increase at an annual rate of 2.8 per cent. Thus, additions to excess capacity have resulted from the sharply reduced rate of demand growth since 1956 combined with continuing additions to basic capacity. Excess Capacity in Petroleum Refining While refining capacity has clearly been large relative to output in the last several years, excess refinery capacity has existed virtually throughout the history of the industry . 1t has been common for expansion in capacity lo he unclcrlakc11 , allho11gh existing <'apacily was not being fully utilized . It is necessary to have a cushion of excess capacity in industries such as petroleum refining in order that short-run expansions in output may be undertaken. In this way, petroleum refineries are similar to steel mills. Plants in both industries are operated on a continuous 3-shift basis. As a result, it is not possible to expand capacity by working overtime or putting on another shift. The two types of plants are further similar in regard to the specialized equipment required. Both require more capital per worker than the average manufacturing industry. Capital invested in petroleumincluding extraction and pipeline transportation-was a large multiple of the average for manufacturing as a whole in 1957. In both cases, existing capacity represents a rather fixed ceiling to production which cannot be raised in the short run by the substitution of other productive factors. This is in contrast to the manufacture of products which require little specialized equipment, or where bottlenecks can be avoided by the su bstituti0n of other factors for capital equipment. In addition to a cushion being required to meet possible short-run demand surges, anticipation of continuing demand growth would likely lead to some degree of current excess capacity. 14 While the need for a cushion of extra capacity appears clear enough, it may well be asked what the normal or desirable magnitude of this cushion is likely to be. For the first 11 years after World War II, crude petroleum inputs amounted to about 90 per cent of operating capacity. During this period the per cent of capacity in operation fell below the 11-year average three times- in 1949, 1950, and 1954, all years of reduced business activity. If the average level of capacity utilization from 1946 to 1956 were considered to he the goal of petroleum refineries as an aggregate, utilization exceeding this level for a period of time would stimulate r<'finers to expand capacity. On the other hand , if utili zation fdl below this level for a significant p eriod, additions to capacity would be discouraged. If, then, the normal rate of capacity utilization is taken to be around 90 per cent, recent rates have been below normal. In 1957, capacity utilization was 89 per cent of the potential and indications were that this level was considerably higher than demand justified , as refined stocks piled up and at the end of the year were 9 per cent higher than the ycar-C'arli er level. In 1958, U. S. refin eries operated at 83 per cent of capacity. While this was a year of recession, this rate of capacity utilization was a low for the period since Per Cent of Refining Capacity in Operation Crude Oil Runs to Stills as a Per Cent of Operating Capacity* PE.R CENT CAPACITY IN OPERATION 100 90 80 1946 '48 '50 '52 ' 54 '56 ' 58 '60 * Capacity is calculated as of July 1 by averaging capacity at the beginning and end of each year . SOURCE: U. S. Bureau of Mines . in Petroleum Refining World War II-below even the rate of previous postwar recession years. The following year the rate of capacity utilization showed little improvement. In spite of the significantly increased level of excess capacity, further increases in capacity were made during this 3-year period. In fact, while crude runs to stills increased less than l per cent from 1956 to 1959, crude oil refining capacity increased more than 8 per cent from January 1, 1957, to January 1, 1960. These developments are, at least at first glance, contrary to what might be expected. However, a variety of fa ctors may help to account for the seemingly incongruous developments of the past S<'V('ral y('ars. An influence which woul<l contribute to increases in the industry's operating capacity is the decision of an individual refinery to expand capacity in an attempt to capture a larger share of the market. Although it can be concluded that the increase in total demand has not kept pace with recent additions in operating capacity for the industry as a whole, individual companies may have realized increases in their market share in line with the capacity expansion undertaken. In addition, the geographical distribution of new crude discoveries combined with population shifts have resulted in the construction of some facilities to take advantage of proximity to developing population centers and convenient access to crude supplies. Also, modifications are often necessary in refineries before crude from a new source can be refined. This is a result of the unique properties of various crudes and would of course tend to raise investment if not capacity. Decisions to expand operating capacity during 1957, 1958, and 1959 also may have in Monthly Review • October 1960 part resulted from the difficulties in gauging the extent to which increases in demand since 1950 were a consequence of the Korean War and the Suez crisis. These two influences served to obscure changes in long-run trends in demand for refined petroleum products. The process of separating and identifying short- and longer-term factors was further complicated by the recession which folJowed upon the stimulation of Korea and Suez. Another factor which may be significant in explaining the continued increases in capacity is the fact that actual expansion in capacity is usually not realized for some time after the decision to expand is made. This lag may be of substantial length in th e petroleum industry where e<1uipme11t is elaborate and produceJ. to custom order. It is conceivable that at least some capacity increases planned in late 1956 and early 1957 might not have been completed until 1958 or 1959. The latter two explanations for the continued increases in capacity in 1957, 1958, and 1959 lead to the conclusion that a certain period of time may be required for the industry to adjust to changing patterns in demand. The most recent survey of refinery capacity taken by the Bureau of Mines on January 1, 1960, indicates adjustments to excessive capacity may be under way. This is suggested by the fact that refinery capacity under construction as of January 1, 1960, was at the lowest level since 1945. Although the magnitude of capacity building on January l has not been a particularly reliable indication of the annual increase in operating capacity in the past few years, the low level of capacity construction supports the hypothesis that the 1960 increment to refining capacity will be small. 15 New Construction Plant and Equipment Outlays 45 BILLIONS OF DOLLARS 5 ·5 SEASONALLY 40 40 5.0 5.0 35 35 4.5 4 .5 30 30 4 .0 4.0 25 25 3.5 BILL IONS OF DOLLARS BILLIONS OF DOLLARS 45 SEASONALLY ADJUSTED ANNUAL RATES BILLIONS OF DOLLARS 5 5 · ADJUSTED 3.5 RESIDENTIAL ~ 2.0 20 20 3.0 15 15 1.5 ~ 1.5 10 1.0 * 1.0 PUBLIC 10 --- .. _...,_ PUBLIC UTILITIES ,/ ~ ..... ._ • .....- 5 ~ .................. TRANSPORTATION ..-v- BUSINESS ALL OTHER .5 5 PRIVATE .5 o .............................................~~_._..__._....__.........._.__.........J....._._...............................................__.__. 1953 '54 '56 '55 '57 '58 '59 1953 '60 '54 '55 '56 '58 '57 * CHANGE *ANTICIPATED District and States Reserve City Member Banks Deposits Country Member Banks Reserve City Member Banks 0 '60 IN SERIES PRICE INDEXES, UNITED STATES BANKING IN THE TENTH DISTRICT Loans '59 Index Country Member Banks Aug. 1960 July 1960 Aug. 1959 124.8 Consumer Price Index (1947-49 = 100) 126.6 126.6 Wholesale Price Inde x (1947 -49 = 100) 119.2 119.7 119.1 Prices Rec' d by Farmers (1910-14 = 100) 234 238 239 Prices Paid by Farmers (1910-14 = 100) 298 298 297 August 1960 Percentage Change From TENTH DISTRICT BUSINESS INDICATORS July Aug. July Aug. July Aug. July Aug. 1960 1959 1960 1959 1960 1959 1960 1959 Tenth F. R. Dist. Colorado Kansas Missouri* Nebraska New Mexico* Oklahoma* Wyoming +1 +4 +2 +12 -1 -2 t +2 t +7 +4 +13 +1 - 1 +4 +3 +3 t +3 +17 -2 t +1 +s - 3 t +1 - 1 +1 +3 +1 - 3 :j::j: t ** • ; ••·· +7 ** *Ten th District portion only. t Less than 0.5 per cent. 16 - 1 +3 +6 +9 ** District and Principal Metropolitan +12 +3 +3 0 +0 +4 +2 - 13 -1 Denver +2 t Wichita - 4 Omaha +10 :j:* +3 +16 - 2 - 5 - 1 +3 Oklahoma City ,;,* ** t + 2 Tulsa + 0 Year to date - 1 Kansas City - 7 +3 ** No reserve cities in this state. Year to date Aug . +10 +2 - 1 Percentage change-1960 from 1959 Areas Tenth F. R. District Value of Department Store Sales Value of Check Payments Aug. - 5 - 11 +11 +2 +7 +2 +18 +2 +0 +3 +13 +3 0 0 +5 0 - 1 - 4