TDOC ZTA245. 7 B873 8 3 l ] A8“ " "iv¢r$ i*Y B-1331 "' ‘ x4 Contents SUMMARY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 2 Export Demand of Study-Area Wheat by Port Area. . . .. 1 INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 3 RESULTS OF ANALYSIS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 12 Current System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . MARKETING SYSTEM AND POTENTIAL EFFICIENCIES ..... .. 3 2343i) ivstesmt ------------------------------------- -- OBJECTIVES AND PROCEDURES ........................ .. s " ' a’ Y5 em """""""""""""""""""" " Objectives . . . . . . . . . . . . . . . . . . . . . . . . . . i. . . . . .. 5 ggiiifiaiyagfcégsfiyssgsmgs ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' " 15 gtrsgggijrgéatgdisgsiifirrggtéczgss of the Analyuca MO e g for Each Distribution System . . . . . . . . . . . . . . . . . . . . .. 16 Tiyme Frame of gosts Included 8 EnergY .C°"S‘““P“°" for Each Dismpuuon System ' ' ' " 17 DATA 8 Sensitivity of Subterminal Organization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..‘........... t C ntsss_._ss"s____sss_17 Wheat Supply, Farm, and Country Elevator Storage . . .. 8 igipargtagénialncfeaszz Fafil/ieglgiage Cost s Farm Assembly Cost . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 9 on Subteiminai Organizations _ _ _ _ _ _ _ _ _ _ _ _ _ _ __ i7 Earm tHaigliling and ‘Stoirjage Costs . . . . . . . . . . . . . . . . . . . .. 9 im-pact of increased Trucioweighi Limit m 9"" TY 9V3 0" _" a" ' ermma I on Subterminal Organization . . . . . . . . . . . . . . . .. 18 and Port-Terminal COSIS . . . . . . . . . . . . . . . . . . . .' . . . . . . 9 impacts of Aitei-native Distribution Systems Cost of uPgradmg Comm)’ Ekwators to 5“bte'm'"a|5" 9 On Marketing-System Participants . . . . . . . . . . . . . . . .. 19 Commercial Truck Transportation Cost .» . . . . . . . . . . . . . .. 10 Raiiroad Costs _ _ . _ . _ . . . _ . _ _ . . . _ . _ . _ . . _ _ _ _ . _ _ _ _ _ _ _ . _ H i0 REFERENCES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 21 Barge Costs _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ __ 11 APPENDIX A . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 21 Grain Inspection and Grading Costs at Subterminals... 11 Summary The foci of this study were the economic feasi- bility and potential cost savings of operating wheat- carrying unit trains between a six-county area in the Texas-Oklahoma Panhandle and Texas Gulf ports. Costs of the current system are estimated and then contrasted with three alternative organizations that include operation of 20-, 50-, and 80-car trains from study-area origins. The current system involves sin- gle-car movements and transit privileges at existing inland-terminal locations. The following three alter- native distribution systems are studied: 1) a system involving the operation of 80-car unit trains between area inland-terminal lo- cations and Texas ports (referred to as the 80- car system); 2) a system of 80-car unit trains operating from area inland-terminal locations and of poten- tial subterminals served by 50-car unit trains for delivery of wheat to Texas ports (referred to as the 50-, 80-car system); and 3) a system of 80-car unit trains operating from area inland-terminal locations and of poten- tial subterminals served by either 20-, 50-, or 80-car unit trains for delivery of wheat to Texas ports (referred to as the 20-, 50-, 80-car system). This research indicates that subterminals served by 50-car trains (50-, 80-car system) would be feasible at five of the study area's 10 potential locations and would be responsible for handling 49 percent of the wheat destined for Texas ports. In addition, the analysis shows that either a 20-, 50-, or 80-car-train operation iwould be feasible at all 10 potential subterminal locations. Within the 20-, 50-, 80-car system, subterminals would capture 75 percent of the wheat moving to Texas port areas. The model analysis shows the 20-, 50-, 80-car Keywords: Unit train/transportation economics/grain transporta- tion/grain exports. 2 system to be the most efficient of the export-wheat distribution systems. For the six-county area, this would annually generate marketing-system savings of $2.49 million or 12.2 cents per bushel. The 50-, 80- car system was the second-most efficient system with expected annual savings of $2.08 million or 10.2 cents per bushel. Although the 80-car system ranked third in terms of potential efficiency gains, the analysis indicated this system would yield savings of 8.1 cents per bushel or $1.65 million. The principal source of marketing-system sav- ings was found to lie in the efficiency of the unit- train concept. The per-ton-mile cost savings to the railroads range from 23 percent for the 80-car system to 41 percent for the 20-, 50-, 80-car organization. Energy savings are also indicated within the alternative distribution systems. The current system consumes an estimated 399 billion BTU's (British Thermal Units of energy). The most energy efficient system is the 80-car-train organization which con- sumes approximately 297 billion BTU's — an energy savings of 26 percent. The 50-, 80-car and the 20-, 50-, 80-car systems have estimated energy savings of 25 and 24 percent, respectively. Several conclusions may arise from this re- search. First, the unit-train concept is a feasible means of improving the export-wheat marketing system efficiency in the Plains study area, and cost i savings are large enough that similar results may be concluded for other Plains areas. Second, a subter- minal organization served by unit trains is feasible and cost reducing as compared to an organization of unit trains operating from only inland-terminal loca- tions. Although the greatest opportunity for cost reduction and increased efficiency includes a sub- terminal organization (50-, 80-car and 20-, 50-, 80-ca i systems), this organization requires simultaneou ‘ development of several system components. Conse- quently, this would be the most difficult alternative to implement. A. n5 Alternative Export-Wheat Distribution Systems for the Texas-Oklahoma Panhandle Stephen W. Fuller and C. V. Shanmugham* Hard Red Winter wheat is a major source of income for U.S. and South Plains grain producers. Historically, wheat has ranked as one of the most valuable crops in Texas and Oklahoma, states that are major producers of the annual Hard Red Winter wheat national output and that are located in the southern portion of this grain’s production area (Figure 1). Exports are a significant outlet for the national annual wheat production and hold promise of be- coming even more important in the future. On a national basis, exports of Hard Red Winter wheat generally comprise from 43 to 75 percent of the annual production, except in the unusual market conditions of 1972 and 1973 when exports exceeded 75 percent of the current crop (Table 1). Exports of Hard Red Winter wheat increased from 336 million bushels in 1969 to 775 million bushels in 1973. Since 1973, Hard Red Winter wheat export volume has fluctuated between 418 and 625 million bushels? During the 1970's, Gulf ports were responsible for 50 to 62 percent of the Nation's total wheat exports and from 76 to 86 percent of the Hard Red Winter wheat exports (Table 2). Texas Gulf ports, particularly the North Texas Gulf ports (Beaumont, Port Arthur, Houston, and Galveston), are the prin- cipal Hard Red Winter wheat export locations. In 1977 and 1978, North Texas Gulf ports were respon- sible for 70 and 68 percent of the respective Hard Red Winter wheat exports in the United States. In the same time periods, South Texas Gulf ports A(Corpus Christi} and Brownsville) exported 8 and 12 I percent of the ‘respective U.S. foreign sales of this wheat class. WAssQciate professor (Department of Agricultural Economics) and assistant professor (Department of Industrial Engineering), The Texas Agricultural Experiment Station. 1 Hard Red Winter wheat typically constitutes about 50 percent of total wheat exports. i South Plains Export-Wheat Transportation/Marketing System and Potential Efficiencies The country elevator represents the first mar- keting agent in the South Plains export wheat marketing system. Wheat assembled to country elevators by producers will generally move to inland terminals which, in turn, distribute grain to Gulf ports as warranted by demand. The region's single- car rate structure allows for transit at the inland- terminal locations. The substantial capacity of the South Plains inland-terminal industry is, in part, a product of the railroad’s rail rate structure — the railroad’s transit privilege. This privilege permits wheat to be shipped from country elevators to Gulf ports on a single-car- through rate with intermediate stops for inland- terminal storage. In essence, the rate on a direct shipment from country elevator to Gulf port is equal to the sum of the rates from country elevator to inland terminal and from inland terminal to Gulf port. If follows that a grain shipper’s transportation charge on export-destined wheat is not unfavorably affected by transshipment at inland-terminal loca- tions. A second important aspect of the current rate structure involves equalized rail rates to Gulf port locations. This rate structure allows grain handlers to ship to most of the Gulf ports at the same rate. The export rate structure can be more easily understood through consideration of a specific ex- ample. Assume a country elevator located at Perry- ton, Texas has a Gulf export rate of 50.7 cents per bushel (Ex Parte 343). Because of the transit privilege and equalized Gulf rates, this grain may move at the 50.7-cents-per-bushel rate to any regional inland- terminal location for storage prior to its final move- ment to the Gulf port areas. Accordingly, up to 90 3 HARD RED WINTER WHEAT PRODUCTION AREA Figure 1. Location of the Hard Red Winter wheat production area in the United States (one dot equals 10,000 harvested acres). Source: U.S. Wheat Industry, Economics, Statistics, and Cooperative Service, U.S. Department of Agriculture, Agricultur- al Economic Report No. 432, August 1979. percent of the wheat produced in the South Plains moves through inland terminals. Recently, railroads have initiated truck- substitution and truck-allowance tariffs. The truck allowance or substitution allows for the transit privilege as if grain were being moved by rail to the inland terminals. With truck substitution, railroads pay the trucking cost and bill the elevator for the flat rail rate to the terminal. Truck allowance is handled by the elevator which pays for trucking to the inland terminal with the railroad deducting a predeter- mined amount from the rail rate. Transportation activities comprise about 75 per- cent of export-wheat marketing costs in the South Plains and thus represent one of the most critical and costlyiilinks in the system. On-going research at several Midwestern institutions has revealed the unit trainto be a more efficient means of long- distance grain transportation. ln general, Midwest- ern researchers havebeen investigating means of maintaining the efficiency of a grain transportation system that would abandon a major portion of the region's branchline segments. Researchers at Iowa State University investigated the practicality of re- 4 structuring the country elevator industry to include subterminals located on retained railroad mainlines [1]. In this reorganization scheme, the subterminal received grain from producers and country elevators and shipped on unit trains destined for port loca- tions. Their study revealed the cost savings of the unit train to more than offset other cost increases resulting from reorganization. Studies of analogous situations in the Corn Belt regions of Indiana and Ohio have reached similar conclusions [3,4]. Al- though previous studies contain some parallel as- pects with the South Plains wheat-producing region, there are important marketing-system differences that keep their results from being useful to this major wheat-producing area. In contrast to the Midwestern situation, rail abandonment is not a serious threat in the SoutlQT Plains, except on a few branchline segments. lt follows that the region's country elevators and subterminal's cost superiority and feasibility are le apparent here than in the Midwest where some country elevators are left without rail service. Another contrasting characteristic is the substantial subterminals alike would have rail service, thush P TABLE 1. HARD RED WINTER WHEAT: MARKETING YEAR, SUPPLY, AND DISAPPEARANCE, 1969-78 Year ; ~- eginning lune 1 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 (000,000 bu.) Beginning , stocks 475 574 492 471 287 169 224 377 603 631 Production 785 755 747 761 957 879 1,053 968 997 834 TOTAL 1,260 1,329 1,239 1,232 1 ,244 1 ,048 1,277 1,345 1,600 1,465 Domestic Use 350 387 432 326 300 314 314 324 443 426 Exports 336 450 337 704 775 510 581 418 565 625 TOTAL 686 837 769 1,030 1 ,075 824 895 742 1 ,008 1 ,051 Ending Stocks May 31 574 492 470 202 169 224 382 603 592 414 Source: Wheat Situation, Economic, Statistics and Cooperative Service, U.S. Department of Agriculture, issues WS-219 through WS-244. TABLE 2. PERCENT OF HARD RED WINTER WHEAT EXPORTS THROUGH VARIOUS COASTAL PORT AREAS Port Areas 1970 1971 1972 1973 1974 1975 1976 1977 1978 Percent Great Lakes 0.5 0.0 0.0 0.3 0.2 0.0 0.0 0.0 0.0 Atlantic 0.2 0.2 0.0 4.1 0.4 0.0 0.0 0.0 0.0 Gulf 77.9 86.4 87.1 85.4 81.6 79.9 76.4 84.7 i 983.1 Pacific 21.4 13.4 12.9 10.2 17.8 20.1 23.6 15.3 16.9 TOTAL 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 Source: Grain Market News, Agricultural Marketing Service, U.S. Department of Agriculture, various issues, 1970-78. inland-terminal industry located throughout the South Plains wheat-producing area. Because of ex- isting inland terminal’s substantial grain handling and storage capacity, no new plant investment would be necessary to accommodate unit trains; however, substantial investment would be neces- sary to upgrade country elevators into subterminals. In which case, a potential subterminal organization would be at a relative cost disadvantage when compared to the Midwest. An additional factor is the region's relatively low density of grain production, which is about one-fourth of that in the Midwest. Assembling large volumes to potential subterminal locations would involve larger market areas and increased assembly cost. Clearly, a subterminal organization in the South Plains appears to be at a relative disadvantage. To determine the economic feasibility and po- tential cost savings of unit-train operation in the South Plains, the Texas Agricultural Experiment Sta- tion in cooperation with the U.S. Department of Transportation, Texas Transportation Institute, Ok- lahoma Agricultural Experiment Station, and Kansas gricultural Experiment Station, has initiated a mul- tistate research project. The principal analysis cen- ters on a contiguous 27-county area located in outhcentral Kansas, Northcentral and Panhandle S fiareas of Oklahoma, and the northern-most counties in the Texas Panhandle.jThis report focuses on findings associated with a six-county area in the Texas and Oklahoma Panhandles (Figure 2). Because of the unique characteristics of this six-county area, particularly its very low density of wheat production and lack of an inland-terminal industry within the region, this report focuses on this area. Objectives and Procedures Objectives Focusing on the comparative efficiency of alter- native distribution systems for marketing Hard Red Winter wheat from the Texas-Oklahoma Panhandle region, this research had as its objectives to deter- mine 1) the economic feasibility of renovating selected country elevators into subterminals and operating unit trains between these facilities and Texas Gulf ports, 2) the economic feasibility of operating unit trains between inland terminals and Texas Gulf ports, and 3) the effect of these organiza- tions on the cost of handling export wheat. The alternative organizations involve various combina- tion of 20-, 50-, and 80-car shipments from potential subterminal locations and 80-car shipments from existing inland-terminal facilities. The following al- ternative distribution systems were studied: 1) a system involving the operation of 80-car unit trains between area inland-terminal lo- cations and Texas ports (referred to as the 80- car system); 2) a system of 80-car unit trains operating from area inland-terminal locations and of poten- 5 TEXAS BEAVER SHERMAN HANSFORD OCHILTREE LIPSCOMB Figure 2. Study-area counties in the Texas-Oklahoma Panhandle. tial subterminals served by 50-car unit trains for delivery of wheat to Texas ports (referred to as the 50-, 80-car system); and 3) a system of 80-car unit trains operating from area inland-terminal locations and of poten- tial subterminals served by either 20-, 50-, or 80-car unit trains for delivery of wheat to Texas ports (referred to as the 20-, 50-, 80-car system). Secondary objectives of this study were to 1) determine the differences in energy con- sumption for the current and alternative distribution systems, and 2) determine the sensitivity of a subterminal organization to unfavorable movements in system cost parameters. Figure 3 illustrates the grain handling and stor- age system elements and their involvement in the current and alternative distribution systems. The current system is characterized by flows from farms to country elevators with subsequent flows to inland terminals and ports. Most of the commercial trans- 6 portation in this system is via single-car rail ship- ments. The 80-car system would include the grain handling and storage elements of the current syste a and 80-car train shipments between inland and port l terminals. The 50-, 80-car and the 20-, 50-, 80-car systems would involve the introduction of a new marketing system element, the subterminal, which is served by either 20-, 50-, or 80-car trains. Both of these systems would include 80-car trains operating from inland terminals. Structure and Assumptions of theAnalytical Model Improving marketing efficiency is a goal that cannot be pursued in isolation. Because of the high degree of interdependence among the elements of this area's export-wheat marketing system, a cost- minimizing model of the entire system was con- structed. The following are the principal cost ele- ments of the model: 1) farm storage costs, 2) farm assembly costs, 3) truck, rail, and barge transporta- tion costs that link country elevators, potential subterminals, inland terminals, and port terminals, and 4) all facilities’ grain handling and storage costs. The system model represents a wheat crop year (June 1 - May 30) subdivided into three time periods to facilitate a temporal analysis. The first time period includes the first 21 days of the wheat crop year, when harvest is carried out and the annual wheat supply generated. The following 45 days constitute the second time period which represents post- harvest activity, while the final or third period consists of the remaining 299 days of the crop year. The six-county region was subdivided into 3 >< 3- mile areas (9 square miles), which resulted in 825 production origins. The harvest-time supply of wheat and available wheat storage at each produc- tion origin were predetermined. The predetermined wheat production reflected 1985 production, and the portion destined for export (84 percent) was based on historical data. Producers may store their annual wheat production at farms (production ori- gins) or ship directly by farm truck to country elevators or subterminals. As farmers appeared re- luctant to deliver grain in excess of 30 miles, only those country elevators or subterminals within 30 miles of a farm represented a potential delivery point. lf wheat is farm-stored, producers deliver to country elevators or subterminals in later time periods. Since most wheat enters the marketing system via the country elevator, the model was structured so that wheat must be assembled to country elevators or subterminals prior to furtherfi movement through the system. The model includes 58 country elevators located at 36 locations. Country elevators and potential \ subterminals have predetermined amounts of sto . age capacity available for area wheat production. All subterminals are renovated country elevators. Country elevators may ship to subterminals, inland P Inland Terminal Farm Production Country Elevator Port Terminal Figure 3. Wheat-flow patterns among marketing-system elements for the current and alternative export-wheat distribution systems. terminals (Enid, Fort Worth, Amarillo), Gulf port terminals (Houston, Galveston, Beaumont, Port Ar- thur, Corpus Christi, Mississippi River ports), and a river elevator on the Arkansas River (Catoosa, Ok- lahoma). The river elevator is linked to all Gulf ports via barge transportation. In the model all movement from country elevator to subterminal is restricted to 75 miles. Truck and rail modes are available for all country elevator shipments except those to subter- minals and the river elevator; in which case, only truck carriage is available. All country elevator rail shipments are represented as single-car movements in the cost-minimizing model. . The predetermined subterminal locations re- quire upgrading or investment costs in order to accommodate unit trains. Subterminals are reno- vated to accommodate 20-, 50-, or 80-car shipments. The level of investment is related to the size of unit train to be accommodated. Inland terminals, port terminals, and the river elevator have predeter- mined wheat storage capacities available for study- area wheat marketings. Inland terminals may ship by either truck or single-car movement to other inland terminals, by truck to the river elevator, or by 80-car shipments to port elevators. When estimating costs of the current system, rail movement from inland terminals was via single-car shipments. No invest- ment is required at inland terminals for loading of . the 80-car unit trains. Port terminals may receive "truck-, rail-, or barge-delivered grain which may be stored for short periods of time prior to loading aboard ship. Wheat demand per port and time ; y eriod was predetermined and based on historical s‘ flows to port areas. ' To estimate the grain-handling and transporta- tion costs associated with marketing the region's l I - - - - - - - ' - - - - - - - - - - _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _:l Subterminal represents principal wheat flows associated with current system represents potential wheat flows associated with two of the alternative distribution systems export-destined wheat supply under the current system, shippers faced with the current single-car, transit rate structure were assumed to continue to route wheat as historically practiced. Therefore, wheat flows in the cost-minimizing model followed the historic pattern to allow for the calculation of current-system costs. An estimate of current-system costs represented a benchmark against which alter- native distribution-system costs were compared. When calculating the latter, grain was not forced to follow an historic flow pattern except as dictated by past export demands at the various port areas. Rather, wheat was allowed to flow through least-cost channels in order to meet the predetermined export demands. Because of the need to include substantial details of the transportation and marketing system as well as a spatial and temporal dimension, the resulting model became very large. For this reason, a network-flow model was developed. Previous studies revealed network-flow models to be compu- tationally more efficient than linear-programming codes and capable of accommodating the system characteristics [2]. (See Appendix A for a full descrip- tion of the model.) System Data Requirements To construct the system model, substantial data were required. The following is a list of the model's data needs by marketing-system element. Production Origin 1. Quantity of wheat harvested on each produc- tion origin (3 x3-mile area) 2. Wheat storage capacity of each production origin 3. Cost of placing wheat into farm storage, storage, and removal from storage 4. Cost of transporting wheat from production origin to country elevators or subterminals within 30 miles Country Elevators and Subterminals 1. Facilities available for wheat handling and storage with associated capacities 2. Cost of receiving wheat from farm trucks, storing wheat at country elevators, and load- ing into commercial trucks and rail cars 3. Cost of upgrading a country elevator into a subterminal Inland Terminals and River Elevator (Port of Catoosa) 1. Facilities available for wheat handling and storage with associated capacities 2. Cost of receiving wheat from trucks and rail cars, of storing, and of loading into commer- cial trucks, barges, and rail cars 3. Cost of loading barges at the Port of Catoosa (terminal on the navigable portion of the Arkansas River) Transportation 1. Cost of transporting wheat by commercial truck from country elevators to subterminals, inland terminals, port terminals, and the Port of Catoosa 2. Cost of single-car rail movement from coun- try elevators to inland terminals and port terminals 3. Cost of barging wheat from Port of Catoosa to Gulf ports 4. Cost of transporting wheat by commercial truck from inland terminals to other inland terminals and port terminals 5. Cost of single-car rail movement from inland terminals to other inland terminals and port terminals 6. Cost of unit-train operation from subtermi- nals and inland terminals to Texas Gulf ports Port Terminals 1. Facilities available for wheat handling and storage with associated capacities 2. Cost of receiving wheat from trucks, rail cars, and barges and cost of loading grain into ocean-going vessels 3. Demand for study-area wheat at each Gulf port Time Frame of Costs Included in Model As the focus of this study was the comparable efficiency of the alternative distribution systems, emphasis was directed at estimating comparable transportation cost parameters for each mode. 8 Therefore, all transportation cost parameters were calculated to include total costs. For existing grain handling and storage facilities, a only variable or short-run costs were included in the model. It can be argued that these facilities will continue to operate as long as variable costs are covered. The life of most commercial grain storage facilities ranges from 25 to 30 years, which indicates that their long-run time frame is considerable. When new capital is invested in a country elevator for purposes of renovating into a subterminal, total cost is included in the analysis. New capital will only be invested by an entrepreneur if the capital can be recovered and a comparable return on the invested capital generated. Data The following section relates parameter values entered into the study-region model. For a detailed description of the data and methodology employed to estimate model parameters, Rail Based Transpor- tation of Export Destined Wheat: An Efficiency Study (Department of Transportation, Federal Railroad Ad- ministration, 1980) should be consulted. Wheat Supply, Farm, and Country Elevator Storage Based on historical production trends, the es- timated 1985 wheat output for the six-county area was 24.2 million bushels. Based on historical grain flows 84 percent of total production (20.4 million bushels) was estimated to be destined for Gulf ports, the remaining wheat would move into domes- tic markets. A county's estimated production was distributed among its production origins (3><3-mile areas) to agree with the portion of the county's cultivated land area in each production origin. To estimate existing on-farm storage in the study area, a mail questionnaire was distributed to a 10-percent random sample of farmers. Based on this survey, the Texas and Oklahoma counties were estimated to include 12.187 and 8.183 million bushels of on-farm storage, respectively. Approxi- mately two-thirds of this capacity was available for wheat storage. On-farm storage estimates were allocated among farms (3><3-mile areas) to agree with each farm's expectedgrain production. Storage capacity for each of the region's 58 country elevators was obtained from an on-site visit, secondary sources, or a telephone interview. Stor-J age capacity available for export-destined wheat was calculated by subtracting from each elevator’s stor- age capacity that storage necessary for 1) working space, 2) domestically consumed wheat, and 3' carryover of wheat and other grains. Storage capaci- ty for export-destined grain was estimated at 21.3 million bushels. F»: q, Farm Assembly Cost Distance from each farm (3><3-mi|e areas) to each country elevator within a 30-mile radius was calculated. Farm truck-delivery cost to each elevator was determined by a cost function which used distance to predict per-bushel assembly cost. Farm truck costs were determined for a 2.5-ton tandem, tag-axle straight truck; a 2-ton straight truck; and a 1.5-ton straight truck. A survey of elevator receipts indicated that these truck sizes were most commonly employed in farm-to-country elevator delivery. The largest truck (2.5 tons) was assumed to carry 500 bushels and to assemble 35 percent of the country elevator’s receipts. The 2-ton truck was assumed to assemble 50 percent of elevator receipts and have a load size of 300 bushels. The 1.5-ton truck was assumed to assemble 15 percent of receipts and have a load size of 250 bushels. Based on these assumptions, a weighted average assembly cost was estimated for alternative distances (Table 3). Farm Handling and Storage Costs Farm storage cost includes three cost items: 1) cost of placing wheat into storage, 2) cost of wheat storage, and 3) cost of removing wheat from storage. A survey of wheat producers provided information on sizes and characteristics of existing farm storage. With this information cost parameters were cal- culated using the economic-engineering estimation technique. The analysis revealed the per-bushel variable cost of placing wheat into storage to be 2.19 cents, while the per-bushel removal cost was estimated at 1.5 cents. Per-bushel variable cost of storing wheat for 12 months was calculated at 8.3 cents. These costs are for steel bins of 10,000-bushel storage capacity. Country Elevator, Inland-Terminal, and Port-Terminal Costs The Economic Research Service (U.S. Depart- ment of Agriculture) has conducted a series of studies on cost of grain handling and storage in TABLE 3. ESTIMATED FARM TO COUNTRY ELEVATOR ASSEMBLY COST IN CENTS PER BUSHEL, 1977-78 Distance of Assembly haul cost 1F‘ (miles) (¢/bu) 5 6.86 1O 7.73 Q; 15 8.50 _ i 20 y 9.46 25 10.33 30 11.19 country elevators, inland terminals, and port termi- nals. With use of regression analysis, these costs were updated to 1977-78 (Table 4). The parameters reveal the per-bushel costs of receiving and loading grain by truck, rail, and barge at each elevator type and per-bushel costs of storage. Cost of Upgrading Country Elevators to Subterminals To determine the feasibility of unit-train ship- ments from country origins, an assessment was made of the investment needed to modify elevators for this purpose. Analysis revealed that, in general, the level of investment was closely related to the storage capacity of the elevator to be upgraded, the TABLE 4. ESTIMATED COST IN CENTS PER BUSHEL OF RECEIV- ING, STORING, AND LOADING GRAIN BY ELEVATOR TYPE, 1977-78 Country Inland Port Function elevators terminals‘ terminals ------------------- -- (¢/bU-) Receiving Grain Truck Fixed Cost .373 1.013 1.958 Variable Cost 1.934 1.650 1.309 Total Cost 2.307 2.663 3.267 Rail Fixed Cost ----- 1.396 1.265 Variable Cost ----- 2.002 1.317 Total Cost 3.398 2.582 Barge Fixed Cost ----- 1.182 .532 Variable Cost ----- 3.938 1.685 Total Cost ----- 5.120 2.217 Loading Grain Truck Fixed Cost .565 1.395 5.251 Variable Cost 2.065 1.058 2.089 Total Cost 2.630 2.453 7.340 Rail . Fixed Cost .579 1.171 1.640 Variable Cost 2.011 1.514 1.497 Total Cost 2.590 2.685 3.137 Ship/Barge Fixed Cost .096 .348 .498 Variable Cost .974 .758 .772 Total Cost 1.070 1.106 1.270 Storage (annual cost) Fixed Cost 16.212 14.635 26.986 Variable Cost 5.545 4.144 5.131 Total Cost 21.757 18.779 32.117 ‘The river elevator was assumed to have the same cost structure as the inland terminal. Source: Costs of Storing and Handling Grain in Commercial Elevators, 1970-71, and Projections for 1972-74, Economic Research Service, U.S. Department of Agriculture, ERS-501, March 1972. (The tabulated, updated parameters were based on costs taken from the referenced study.) 9 20-car trains) to $872,700 (required of a small elevator to accommodate 80-car trains). Annual fixed _ costs for these respective elevator types are $5,004 l size of multicar shipment to be handled by the renovated elevator, and expected volume of grain to be handled by the subterminal. The investment associated with the smaller elevators (300,000- 750,000 bushels of storage capacity) was greater than that of the larger elevators (over 750,000 bushels of storage capacity). Accordingly, two levels of upgrad- ing costs were estimated (one for each elevator-size category). In addition, the level of investment was dependent on the size of unit train serving the elevator. Since the analysis included three train sizes (20-car, 50-car, and 80-car), three upgrading costs were estimated for each elevator-size category. Ex- isting unit-train rates from Corn Belt origins allow 24 hours for loading; thus, all unit trains were assumed to load within a 24-hour period. It follows that greater elevator upgrading investment was required to accommodate the larger train sizes. In addition, elevator-renovation cost was found to be affected by the annual volume to be handled by the subtermi- nal: the larger the subterminal’s annual volume, the greater was the grain-handling capacity and corre- sponding investment. Investment cost was es- timated for annual volumes less than 1.5 million bushels, from 1.5 to 5.0 million bushels, and greater than 5.0 million bushels. Estimated investment and annual costs of up- grading each elevator-size category are shown in Tables 5 and 6. Investment costs include elevator- equipment costs and rail-equipment costs. Elevator- equipment costs include reclaiming belts, altering spouts, increasing leg capacity, and installing auto- matic samplers and scales. Rail-equipment costs include additional rail siding, a switch, and a track- mobile. Investment costs range from $48,350 (neces- sary for upgrading a large elevator to accommodate and $124,212. Commercial Truck Transportation Cost Truck movement of wheat among and within all elevator categories is assumed to be by commercial truckers. The types of vehicles operated by grain truckers vary; the most common types among inter- viewed firms were diesel-powered, cab-over, twinscrew, tractor-trailer rigs. Analysis revealed that a truck's per-mile cost was influenced by distance of N‘) trip. For this reason, two cost functions were cal- culated — one function for trip distances less than 350 miles, another for distances equal to or in excess of 350 miles. Hauls of less than 350 miles were assumed to have no backhauls, while the longer distances (specifically from the study area to Gulf ports) were assumed to have backhauls one out of every five trips. All loads were assumed to be 860 bushels, except when sensitivity of subterminal organization was tested; in which case, loads were assumed to be 1,100 bushels. Tables 7 and 8 show the calculated costs for the short- and long-distance hauls, respectively. Railroad Costs For purposes of this study, it was necessary to estimate costs of single-car movement from country elevators to inland terminals and port terminals, and from inland terminals to Gulf ports. In addition, the study required cost estimates of 80-car unit trains operating between inland terminals and Gulf ports, as well as 20-car-, 50-car-, and 80-car-train costs of TABLE 5. ESTIMATED INVESTMENT AND ANNUAL COST OF UPGRADING A 300,000 — 750,000-BUSHEL ELEVATOR TO ACCOMMODATE 20-CAR, 50-CAR, and 80-CAR UNIT TRAINS, 1977-78 --- 20-Car-Train Annual Volume --- --- 50-Car-Train Annual Volume --- --- 80-Car-Train Annual Volume --- Over 5.0 1.5 to 5.0 Up to 1.5 Over 5.0 1.5 to 5.0 Up to 1.5 Over 5.0 1.5 to 5.0 Up to 1.5 million million million million million million million million million bushels bushels bushels bushels bushels bushels bushels bushels bushels Investment cost $247,250 $162,150 $121,250 $560,500 $418,500 $518,500 $872,700 $702,250 $560,250 Annual i cost $ 36,400 $ 21,088 $ 14,054 $ 82,150 $ 51,287 $ 51 ,287 $124,212 $ 95,655 $ 64,792 TABLE 6. ESTIMATED INVESTMENT AND ANNUAL COST OF UPGRADING A GREATER THAN 750,000-BUSHEL ELEVATOR TO ACCOM- MODATE ZO-CAR, 50-CAR, AND 80-CAR UNIT TRAINS, 1977-78 --- 20-Car-Train Annual Volume --- --- 50-Car-Train Annual Volume --- --- 80-Car-Train Annual Volume --- Over 5.0 1.5 to 5.0 Up to 1.5 Over 5.0 1.5 to 5.0 Up to 1.5 Over 5.0 1.5 to 5.0 Up to 1.5 million million million million million million million million million bushels bushels bushels bushels bushels bushels bushels bushels bushels Investment cost I $55,850 $48,350 $ 48,350 $513,250 $375,750 $375,750 $825,850 $655,000 $517,500 Annual cost $ 6,028 $ 5,004 $ 5,004 $ 77,768 $ 50,721 $ 50,721 $ 91,163 $ 59,724 $119,770 10 TABLE 7. ESTIMATED COST IN CENTS PER BUSHEL OF COMMER- CIAL TRUCK HAULS FOR DISTANCES LESS THAN 350 MlLESl, I 1977-78 Miles of Per-Bushel haul . cost (miles) (¢/bu) 50 11.10 75 13.3 100 15.5 125 17.7 150 . 19.9 175 22.1 200 24.3 225 26.5 250 28.7 275 30.9 300 33.1 lAssumes no backhaul. TABLE 8. ESTIMATED COST IN CENTS PER BUSHEL OF COMMER- CIAL TRUCK HAULS FOR DISTANCES EQUAL TO OR IN EXCESS OF 350 MILESI, 1977-78 Miies of Per-Bushel haul cost (miles) (¢/bu) 350 37.3 400 42.2 450 47.1 500 52.0 550 56.8 600 61.7 650 66.6 700 71.5 ‘Assumes a backhaul on 20 percent of the trips. operating between potential subterminals and Gulf ports. Railroad costs were estimated by reconstructing the formulae of the ICC cost scales (according to instructions for adjusting cost estimates in Rail Carload Cost Scales 1974, Interstate Commerce Commission, 1976). The railroad freight-rate index (Bureau of Labor Statistics, U.S. Department of Labor) was used to convert 1974 cost estimates to 1977-78 estimates. For ease in estimating the cost of v point-to-point movements, a computerized rail-cost program was developed. With use of the cost algorithm, the per-bushel variable cost associated with single-car and multicar movements was cal- culated. To estimate the variable cost for each rail movement, the values for 21 variables were Wspecified, including: number of cars in shipment, origin, destination, routing, way-train and through- train mileage, number of intra- and inter-company switches, gross tons in way and through train, value _ f grain loss and damage, car days in movement, and switch-engine minutes per car. To make rail costs comparable to the total-cost parameters of the other transportation modes, the variable-cost parameter was multiplied by 1.35. This total-cost parameter was entered into the model for purposes of determining least-cost routings. However, be- cause of the study’s focus on the potential operating efficiency of unit trains, the ”Results” section re- ports only railroad’s variable cost. The analysis revealed savings for multicar ship- ments of 8 to 13 cents per bushel relative to single- car movements. This represented per-bushel sav- ings of 23 to 37 percent relative to the current system. Barge Costs Barge transportation of study-area wheat to Gulf port destinations may occur by way of the Port of Catoosa on the Arkansas River. Estimated barge rates are used in this study as a proxy of barge costs. Barge transportation is a highly competitive industry since the rate on bulk shipment of grain is unre- gulated. Under these circumstances, rates over a period of time should approach long-run costs. Waterway-transportation rates for bulk grainare closely tied to the Waterways Freight Bureau, Freight Tariff No. 7. Rates for this study were estimated by using the Guide to Published Barge Rates on Bulk Grain, Schedule No. 8. and checking these values against the results of a previous cost study [1]. Table 9 indicates values entered into the model to repre- sent costs for barging grain from Catoosa to alterna- tive Gulf ports. Grain Inspection and Grading Costs at Subterminals The necessary grain inspection of unit trains at subterminals was found to be more expensive than the current system where official grades are deter- mined at inland-terminal locations. The additional cost at subterminals was associated with a courier service which traveled between subterminals and inland-terminal locations, or the site of official graders. Estimated cost of this service was .1 cent per bushel for 50-and 80-car trains and .2 cent per bushel for 20-car trains. I TABLE 9. ESTIMATED COST IN CENTS PER BUSHEL OF BARGING WHEAT FROM CATOOSA, OKLAHOMA, TO ALTERNATIVE GULF PORTS, 1977-78 From Catoosa, Oklahoma Cents Per to bushel (¢/bu.) Mississippi River Ports‘ 16.92 Houston, Galveston, Beaumont, Port Arthur 26.82 Corpus Christi 37.26 ‘Includes Ama, Baton Rouge, Destrehan, Myrtle Grove, New Orleans, Reserve and Westwego, Louisiana. 11 Export Demand of Study-Area Wheat by Port Area Export demand for the study region's export- able wheat production was estimated for each port area and by time period. These estimates were based on the study area's historical grain-flow pat- terns. Table 10 indicates the results of these predic- tions. Results of Analysis This section reports the results associated with the current system; the 80-car system; the 50-, 80-car system; and the 20-, 50-, 80-car system. The current- system solution represents a benchmark to which the costs of alternative organizations may be com- pared. Current System To estimate the grain handling and transporta- tion costs of the current system, duplication of the existing system's wheat-flow patterns was necessary. This was accomplished with flow data gathered for the 1976-77 crop year. Table 11 shows the six-county study area's estimated truck and rail flows through the various inland-terminal locations. The structure of the cost-minimizing model forced grain to be moved through these locations in the indicated quantities (Table 11). TABLE 10. ESTIMATED 1985 EXPORT DEMAND FOR STUDY-AREA WHEAT PRODUCTION BY TIME PERIOD Port Areas Time Beaumont — Corpus New period‘ Houston Galveston Port Arthur Christi Orleans (000,000 bu.) 1 .62 .17 .13 .08 .03 2 1.67 .52 .27 .30 .04 3 10.96 1.62 1.12 2.53 .34 TOTAL 13.25 2.31 1.52 2.91 .41 ‘As indicated in an earlier section, the model includes a crop year which has been divided into three time periods. The first time period includes the 21 days associated with harvest, while the second is a 45-day period following harvest. The final period represents the remainder of the crop year, a 299-day period. TABLE 11. THE STU DY AREA'S ESTIMATED 1985 WHEAT RECEIPTS AT INLAND TERMINALS UNDER- THE CURRENT SYSTEM BY MODE OF TRANSPORTI Inland _ Mode terminal location Truck Rail ------------- -- (000 bu.) --------------- Enid 1898 11361 Fort Worth 0 2534 Amarillo 1183 2842 ‘Flow pattern based on 1976-77 crop-year data. ‘l2 Based on predetermined export demands (Table 10) and 1976-77 flow patterns through the inland-terminal organization (Table 11), costs were calculated. Table 12 shows the estimated grain handling and storage costs and associated truck and rail costs for the six-county area. This solution's grain handling and storage costs include those incurred at country elevators, inland terminals, and port terminals. Truck-shipping costs include farm- to-elevator costs, and country elevator-to-inland terminal costs. Rail-shipping costs include the vari- able costs associated with assembling wheat from country elevator to inland terminal, shipping wheat from country elevator to Gulf port, and rail transpor- tation from inland terminal to Gulf port. Tabled rail costs represent single-car movements (Table 12). Based on the 1976-77 flow patterns, the estimated grain handling, storage, and transportation cost for the study region is 56.4 cents per bushel. 80-Car System This alternative involves the operation of 80-car unit trains between inland and port terminals, the only modification relative to the current system. Here, country elevators have the option to ship by the least-cost mode, either by commercial truck or by rail (single-car costs), to inland terminals for subsequent movement on 80-car trains to Gulf ports. In addition, the country elevator may ship directly by truck or rail (single-car costs) to port terminals or ship by truck to the Port of Catoosa for purposes of barging to port terminals. In contrast to the current-system solution, wheat is not forced through the inland-terminal organization (i.e., a historic flow pattern is not being duplicated), but grain is allowed to flow through those channels that are least costly. Information in Table 13 shows estimated 80-car- train costs associated with transporting wheat be- tween the inland-terminal locations and Texas Gulf ports. For these rail movements, the 80-car-train average variable costs were approximately 8 cents per bushel less than single-car costs. All other costs within the model were unchanged from thecurrent system. Existing inland-terminal facilities were as- TABLE 12. THE STUDY AREA'S ESTIMATED COSTS FOR MARKET- ING 1985 EXPORT WHEAT UNDER THE CURRENT SYSTEM BY TYPE OF COST, 1977-781 \ I ‘ x N Type of Cost Cost ($) Total Variable Grain-Handling and -Storage Cost 2,381,334 Total Truck-Shipping Cost 2,345,652 Total Variable Rail-Shipping Cost’ 6,771,248 Total Cost 11,498,234 ~- Per-Bushel Cost .564 l‘ " ‘Costs are associated with marketing 20.4 million bushels. zTotal Rail-Shipping cost may be estimated by multiplying the total variable cost by 1.35. a "TOTAL 2346 ‘Grain flows through inland terminals exceed study-area production TABLE 13. ESTIMATED TOTAL COST FOR 80-CAR TRAIN OPERA- TING BETWEEN INLAND TERMINALS AND TEXAS GULF PORTS IN I@ENTS PER BUSHEL, 1977-78‘ Texas Gulf Portsz Inland Beaumont — Corpus terminal Houston Galveston Port Arthur Christi (¢/bu.) Enid 19.58 20.51 21.55 24.85 Fort Worth 12.04 12.97 14.00 17.04 Amarillo 21.01 21.92 21.97 25.83 ‘Costs are calculated by multiplying variable costs with the 1.35 ratio. N‘ Ffhe analysis did not include unit-train movement to New Orleans. Given the current rail system's regional configuration, the study region is most efficiently served by Texas Gulf ports. Historically, only a small quantity of the study region's production has moved through the Port of New Orleans. sumed capable of loading unit trains, and the existing grain inspection and grading costs as- sociated with loading these trains were unchanged relative to the current system. Based on the model's least-cost solution all wheat (except that demanded at New Orleans) is estimated to flow through the inland-terminal sys- tem prior to shipment to Texas Gulf ports (Table 14). Grain would move through the inland-terminal sys- tem in order to capture the substantial cost savings associated with the 80-car-train shipments. In the model, unit-train costs were not included to New Orleans; consequently, the least-cost means of filling this demand would be direct single-car ship- ments from country elevators. Based on the analysis, the volume flowing to the various inland-terminal locations would be substan- tially altered relative to that observed with the current system. ln particular, flows from the six- county area to Amarillo would be significantly in- creased, while the quantity transported to Enid would be markedly reduced. Also, no grain would flow to Fort Worth inland-terminal locations. TABLE 14. ESTIMATED 1985 WHEAT FLOWS FROM STUDY- REGION COUNTIES TO INLAND TERMINALS BY MODE OF TRANSPORT UNDER THE 80-CAR-TRAIN SOLUTION IN THOUSANDS OF BUSHELSI _ Enid Fort Worth Amarillo County Truck Rail Truck Rail Truck Rail (000 bu.) Beaver, OK 1122 0 0 0 0 0 Texas, OK 0 0 0 0 1680 4659 g Sherman, TX 0 0 0 0 292 2981 Hansford, TX 0 0 0 0 151 7000 Ochiltree, TX 675 0 0 0 982 846 Lipscomb, TX 549 0 0 0 0 1280 0 0 0 3105 16766 because country elevators along borders receive grain from adjacent counties, in particular, Ellis and Harper Counties, Oklahoma, and Seward and Meade Counties, Kansas. The analysis indicates that about 75 percent of the movement from country elevator to inland terminal would be by rail (Table 14). All counties would ship grain by rail to Amarillo except Beaver, Oklahoma (the only county not serviced by a rail- road). Country elevators in all counties would truck some wheat either to Amarillo or Enid. The system costs of this solution are shown in Table 15. Estimated per-bushel cost is 48.3 cents per bushel, compared to an estimated current per- bushel system cost of 56.4 cents, which results in a cost savings of 8.1 cents per bushel. 50-, 80-Car System This organization is analogous to the 80-car system with the exception that selected country elevators are upgraded to subterminals capable of loading a 50-car train. Ten country elevator sites were identified as potential subterminal locations for wheat produced in the six counties. Five of these locations were within the six-county area; remaining sites were located in adjacent counties (Figure 4). The largest elevator at each location was assumed to be upgraded, and the subterminal was allowed to receive wheat directly from producers located with- in 30 miles (the same potential market area as a country elevator) and from country elevators located within 75 miles. Fifty-car, unit-train costs to Texas Gulf ports are substantially less than the single-car cost (10.8 cents less when only variable rail costs are included) (Table 16). However, the opportunity to move grain to inland terminals for shipment on the 80-car train creates a very cost-competitive situation between subterminal and inland-terminal organizations. In order for a subterminal to exist, its associated upgrading, handling, storage, and transportation costs must be less than those similar costs incurred when assembling wheat to inland terminals for shipment on 80-car trains. a The analysis revealed that five of the potential 10 subterminals would be economically feasible. Table 17 identifies these locations, the expected annual volume for each subterminal, and their estimated annual upgrading costs. The model's solution indi- TABLE 15. THE STUDY AREA'S ESTIMATED COSTS FOR MARKET- ING 1985 EXPORT WHEAT UNDER THE 80-CAR SYSTEM BY TYPE OF COST, 1977-781 Type of Cost Cost ($) Total Variable Grain-Handling and -Storage Cost 2,621,856 Total Truck-Shipping Cost 2,918,558 Total Variable Rail-Shipping Cost’ 5,386,071 Total Cost 10,926,485 Per-Bushel Cost .483 ‘Costs are associated with marketing 22.611 million bushels. zTotal Rail-Shipping Cost may be estimated by multiplying the total variable cost by 1.35. 13 w; TABLE 16. ESTIMATED TOTAL COST FOR THE 50-CAR TRAINS OPERATING BETWEEN SUBTERMINALS AND TEXAS GULF PORTS IN CENTS PER BUSHELI, 1977-78 Texas Gulf Portsz Subterminal Beaumont — Corpus location Houston Galveston Port Arthur Christi (¢/bu.) Buffalo, 0K3 28.67 28.50 29.82 36.94 Ashland, KS 30.81 30.58 32.40 35.18 Meade, KS3 32.85 33.84 35.07 40.12 Liberal, KS3 31.77 32.99 34.70 38.11 Shattuck, 0K3 20.30 28.57 29.79 33.60 Guymon, OK 32.82 34.09 35.76 39.18 Perryton, TX 30.40 30.12 31.37 35.17 Gruver, TX 30.05 31.31 32.99 35.94 Spearman, TX 31.08 30.85 32.08 35.88 Stratford, TX 27.46 27.28 30.08 31.94 ‘Costs are calculated by multiplying variable costs with the 1.35 ratio. zThe analysis did not include unit-train movement to New Orleans. Given the current rail system's regional configuration, the study region is most efficiently served by Texas Gulf ports. Historically, only a small quantity of the study region's production has moved through the Port of New Orleans. 3Subterminals located outside the six-county study area but who may receive wheat from the study area. cates that subterminal annual volumes would range from 1.625 million bushels at Meade, Kansas, to 5.562 million bushels at Perryton, Texas. All or nearly all of Perryton, Stratford, and Liberal subterminal receipts would originate from the six-county area, whereas only a fraction of the Meade and Buffalo subterminal receipts would originate from the study area. TABLE 17. SUBTERMINAL LOCATIONS THAT WOULD BE SERVED BY 50-CAR TRAINS WITH ESTIMATED UPGRADING COSTS AND BUSHELS OF WHEAT SHIPPED UNDER THE 50-, 80-CAR SYSTEM 1977-70‘ Necessary Volume to Feasible annualized Total subterminal subterminal upgrading - volume/ from study locations cost subterminal area ($) ----------- -- (000 bu.) ----------- -- Meade, KS 51,287 1625 586 Liberal, KS 50,721 2714 2464 Buffalo, OK 51 ,287 2564 398 Perryton, TX 77,768 5562 5562 __g Stratford, TX 50,721 2756 2756 ti‘ ‘In this subterminal analysis, Seward County, Kansas, was included. Liberal, Kansas, is located in the southern portion of Seward County, which is adjacent to Beaver and Texas Counties, Oklahoma, and received substantial quantities from the six-county area. The solution revealed that 49 percent of the wheat destined for Texas Gulf ports would move through the subterminal organizations. Conversely, only 51 percent of the export-destined wheat would move through inland terminals —- a substantial decrease relative to the current and 80-car systems. As with the 80-car solution, the distribution of wheat among inland-terminal locations would be altered. In particular, Fort Worth would receive no study- area wheat, while the volume moving to Amarillo would be substantially increased and the flow to Enid substantially reduced. Seventy-two percent of inland-terminal receipts from country elevators would be rail transported (Table 18). SEIIIID‘ IBEIIAL CLIIIK ASHLAND ENGLEWOOD PTIMA 0v‘ ADAMS BUFFALO GUYMON AROESTY "up" *Y S . x J’ HITCHLANIT rnnrono m". 2 Slmill" “us § Figure 4. Potential subter- .. Gruver " minal locations for receipt SIEIMI 1,. of study-area wheat pro- .,._ , LIPSCOII d t. a’ (I? nus “c '°"- 1" mm "o" _ gt ocmm: | 14 ITO I TABLE 18. ESTIMATED 1985 WHEAT FLOWS FROM STUDY- REGION COUNTIES TO INLAND TERMINALS BY MODE OF l’ RANSPORT UNDER THE 50-, 80-CAR SOLUTION IN THOU- JANDS OF BUSHELSI Enid Fort Worth Amarillo County Truck Rail Truck Rail Truck Rail (000 bu.) Beaver, OK 434 0 0 0 0 0 Texas, OK 0 0 0 0 1546 2104 Sherman, TX 0 0 0 0 212 572 Hansford, TX 0 . 0 0 0 0 4828 Ochiltree, TX 355 0 0 0 144 18 KLipscomb, TX 448 0 0 0 0 640 TOTAL 1237 0 0 0 1902 8162 ‘Grain flows through subterminals and inland terminals exceed study-area production because country elevators along borders receive grain from adjacent counties, in particular, Ellis and Harper Counties, Oklahoma, and Seward and Meade Counties, Kansas. Results from the model" indicate that it would not be a least-cost alternative for area country elevators to ship to subterminals as subterminals would receive all grain from farmers rather than from country elevators because of the keen cost competition between subterminal and inland- terminal organizations. That is, when all costs are considered, it would be more efficient for a country elevator to ship grain to an inland terminal for movement on an 80-car train rather than to ship to a nearby subterminal for movement on a 50-car train. On-farm storage would be an integral part of the 50-, 80-car organization. The analysis indicates that subterminals would fill their storage at harvest time, and these receipts would represent about 20- 33 percent of the subterminal's annual volume. After harvest, farmers would deliver farm-stored wheat to subterminals where storage would be available. In the model, approximately 35 percent of the available on-farm storage was filled for delivery to subtermi- nals at later time periods. Costs of the 50-, 80-car solution are shown in Table 19. The solution indicates total system costs to average 46.2 cents per bushel, 2.1 cents per bushel less than the 80-car system and 10.2 cents per bushel less than the current system. 20-, 50-, 80-Car System This organization involves 80-car unit trains operating from inland terminals and either 20-, 50-, or 80-car trains operating from selected subterminal locations. The 50-, 80-car solution served as a guide ‘lfor deciding which subterminal locations would be served by the alternative-size unit trains. From this solution, the low volume and impractical locations . - were determined to be partially the results of subter- . _ inal-upgrading costs necessary to accommodate a 50-car train. If a subterminal in the 50-car organiza- tion had an annual volume less than 1.67 million bushels (the volume carried by 10 trains of 50 cars), it TABLE 19. THE STUDY AREA’S ESTIMATED COSTS FOR MARKET- ING 1985 EXPORT WHEAT UNDER THE 50-, 80-CAR SYSTEM BY TYPE OF COST, 1977-781 Type of Cost Cost ($) Total Variable Grain-Handling and -Storage Cost 2,571,305 Total Truck-Shipping Cost 2,717,617 Annualized Subterminal-Upgrading Cost 203,492 Additional Grain-Grading Cost at Subterminals 11,766 Total Variable Rail-Shipping Cost’ 5,354,113 Total Cost 10,858,293 Per-Bushel Cost .462 ‘Costs include six-county area and Seward County, Kansas. Costs are for marketing 23.497 million bushels. ' zTotal Rail-Shipping Cost may be estimated by multiplying the total variable cost by 1.35. was designated a 20-car-train cost. Those subtermi- nal locations receiving in excess of 2.67 million bushels (the volume carried by 10 trains of 80 cars) appeared to have cost attributes that would support larger unit trains; these locations were then desig- nated 80-car-train locations and given the corre- sponding rail rates (Table 20). The 80-car trains operating between subtermi- nals and Texas Gulf port locations had, on the average, variable-cost savings of 12.8 cents per bushel relative to single-car movements. The vari- able rail-cost savings of the 20-car train relative to the single-car movement was estimated at 9.9 cents per bushel. As indicated in the analysis, all of the potential subterminals would be feasible and would receive study-area wheat production. Five of these subter- minals are located within the six-county region, while the remaining are in adjacent counties. Table 21 shows the activated subterminal locations, an- nualized upgrading costs, total volume received per subterminal, and volume received at each subter- minal from study-area origins. The model analysis indicates volume per subterminal would range from 483,000 bushels at Spearman, Texas (20-car train), to 5.8 million bushels at Perryton, Texas (80-car train). _ Subterminals would receive 75 percent of the study area's volume destined for Texas Gulf ports. Con- versely, only 25 percent of the export-destined wheat would move through inland terminals. Table 22 identifies the expected wheat flows from study- region counties to inland terminals by truck and rail. As with the previous alternative distribution-system solutions, the distribution of receipts at inland- terminal locations were substantially altered relative to the current system. Approximately 69 percent of the inland terminal’s receipts would be rail- delivered. As with the 50-, 80-car system, subterminals in the 20-, 50-, 80-car system would depend on produc- ers to store wheat for later delivery. Approximately 46 percent of the available on-farm storage would be 15 1 ‘s4 TABLE 20. ESTIMATED TOTAL COSTS FOR EITHER 20-, 50-, or 80-CAR TRAINS OPERATING BETWEEN SUBTERMINALS AND TEXAS GULF PORTS IN CENTS PER BUSHEL, 1977-781 Texas Gulf Ports3 Size Subterminal of Beaumont — Corpus location train Houston Galveston Port Arthur Christi (cars) (¢/bu.) Buffalo, OK3 50 27.67 23.50 29.82 36.94 Ashland, KS3 20 32.25 32.72 33.86 36.54 Meade, KS3 50 32.85 33.84 35.07 40.12 Liberal, KS3 80 28.57 29.67 31.17 34.30 Shattuck, 0K3 a 20 30.28 30.04 31.27 35.03 Guymon, OK 20 33.62 34.90 36.61 40.07 Perryton, TX 80 27.38 27.15 28.26 31.67 I Gruver, TX 20 30.79 32.08 33.82 36.76 Spearman, TX 20 32.62 32.36 33.59 37.37 Stratford, TX 80 24.80 24.66 27.11 28.84 lCosts are calculated by multiplying variable costs with the 1.35 ratio. 3The analysis did not include unit-train movement to New Orleans. Given the current rail system's regional configuration, the study region is most efficiently served by Texas Gulf ports. Historically, only a small quanitity of the study region's production has moved through the Port of New Orleans. 3Subterminals located outside the six-county study area but who may receive wheat from the study area. TABLE 21. SUBTERMINAL LOCATIONS THAT WOULD BE SERVED BY EITHER 20-, 50-, or 80-CAR TRAINS WITH ESTIMATED UP- GRADING COSTS AND BUSHELS OF WHEAT SHIPPED UNDER THE 20-1, 50-, 80-CAR SOLUTION IN THOUSANDS OF BUSHELS, 1977-78 Necessary Volume to Feasible annualized Total subterminal subterminal upgrading volume/ from study locations - cost subterminal area ($) ----------- -- (000 bU-l ----------- -- Buffalo, OK 51,287 2194 369 Ashland, KS 14,054 1223 29 Meade, KS 51,287 1090 504 Liberal, KS 91,163 3142 2892 Shattuck, OK 14,054 1149 57 Guymon, OK 5,004 1680 1680 Perryton, TX 119,770 5787 5787 Gruver, TX 5,004 2155 2155 Spearman, TX 14,054 483 483 Stratford, TX 91 ,163 3053 3053 ‘In this subterminal analysis, Seward County, Kansas, was included. Liberal, Kansas, is located in the southern portion of Seward County, which is adjacent to Beaver and Texas Counties, Oklahoma, and received substantial quantities from the six-county area. used for this purpose. The analysis indicated no wheat flow from country elevators to subterminals. The estimated cost of this organization would be 44.2 cents per bushel (Table 23). When compared to the current, 80-car, and 50-, 80-car systems, the respective} per-bushel savings of the 20-, 50-, 80-car system are 12.2 cents, 4.1 cents, and 2.0 cents. Summary of Cost Savings for Each Distribution System Each of the distribution systems exhibited sub- stantial savings relative to the current system (Table 16 24). The smallest savings was associated with the 80- car system, the system most analogous to the current one. Based on model results, the estimated ‘ per-bushel cost of this organization would be 48.3 cents, a cost savings of 8.1 cents per bushel relative to the current system or an annual cost savings of $1.642 million. With the second alternative distribu- tion system, existing country elevators are upgraded to accommodate 50-car trains. This system had I estimated costs of 46.2 cents per bushel, represent- ing a 10.2-cents-per-bushel savings relative to the current system. Based on study-region production, this system would yield an annual savings of $2.081 million. Based on model results the cost of the final alternative system (20-, 50-, 80-car system) would be 44.2 cents per bushel, a savings of 12.2 cents per bushel relative to the current system. This is the most efficient of the analyzed systems and would TABLE 22. ESTIMATED 1985 WHEAT FLOWS FROM STUDY- REGION COUNTIES TO INLAND TERMINALS BY MODE OF TRANSPORT UNDER THE 20-, 50-, 80-CAR SOLUTION IN THOUSANDS OF BUSHELSl ~ Enid Fort Worth Amarillo County Truck Rail Truck Rail Truck Rail (000 bu.) Beaver, Ok 241 0 0 0 0 0 Texas, OK 0 0 0 0 891 1389 Sherman, TX 0 0 0 0 120 275 Hansford, TX 0 0 0 0 0 1897 Ochiltree, TX 139 0 0 0 0 18 Lipscomb, TX 303 0 0 0 0 261 TOTAL 683 0 0 0 ‘Grain flows through subterminals and inland terminals exceed study-area production because country elevators along borders receive grain from adjacent counties, in particular, Ellis and Harper Counties, Oklahoma, and Seward and Meade Counties, Kansas. 1011 3840 5.; TABLE 23. THE STUDY AREA'S ESTIMATED COSTS FOR MARKET- t, ING 1985 EXPORT WHEAT UNDER THE 20-, 50-, 80-CAR SYSTEM Y TYPE OF COST, 1977-781 ‘l Type of Cost I Cost ($) Total Variable Grain-Handling and -Storage Cost 2,428,949 Total Truck-Shipping Cost 2,360,088 Annualized Subterminal-Upgrading Cost 352,219 Additional Grain-Grading Cost at Subterminals 21,012 Total Variable Rail-Shipping Costz 4,996,861 Total Barge-Shipping Cost 30,456 Total Cost ' 10,189,586 Per-Bushel Cost .442 ‘Costs include the six-county area and Seward County, Kansas. Costs are for marketing 23.063 million bushels. 2Total Rail-Shipping Cost may be estimated by multiplying the total variable cost by 1.35. TABLE 24. THE STUDY AREA'S ESTIMATED COST SAVINGS AS- SOCIATED WITH THE ALTERNATIVE DISTRIBUTION SYSTEMS, 1985 EXPORTS AND 1977-78 COSTS Per-Bushel cost savings Per-Bushel relative to Annual System cost current system saving‘ -------------- -- (¢/bu.) ($) Current 56.4 80-Car 48.3 8.1 1,652,400 50-, 80-Car 46.2 10.2 2,080,800 20-, 50-, 80-Car 44.2 12.2 2,488,800 ‘Calculated by multiplying per-bushel savings (column 3) by the study area's expected 1985 volume entering export channels, 20.4 million bushels. yield annual marketing-system savings of $2.489 million. The principal source of savings for each of the three alternative distribution systems is reduced railroad-shipping costs. Variable rail cost for the current system is estimated at 33.2 cents per bushel, whereas the 80-car-train system had an estimated variable rail cost of 23.9 cents per bushel, which results in a cost savings to the railroad of 9.3 cents per bushel. The estimated rail costs of the 50-, 80-car y and the 20-, 50-, 80-car systems were 22.8 cents and 21.9 cents per bushel, respectively. This represents respective variable-cost savings to the railroads of 10.4 cents and 11.2 cents per bushel. Rail cost per- ton-mile was calculated for each system to gain additional insight into potential railroad efficiency. )With the current system, variable rail cost per-ton- mile was estimated at 1.398 cents. The estimated variable rail costs per-ton-mile for the 80-car, 50- 80- car, and 20-, 50-, 80-car solutions were 1.084 cents, ._ 897 cent, and .818 cent, respectively. The railroad’s “ton-mile cost savings ranged from 23 percent for the 80-car system to 41 percent for the 20-, 50-, 80-car solution. Energy Consumption for Each Distribution System Due to the decrease in available energy supplies and the associated increase in energy’s value, re- search was carried out to estimate energy consump- tion by the current system and the three alternative systems. This was accomplished by aggregating ton- miles generated by a particular mode in a specific movement and multiplying this value by an appro- priate parameter reflecting BTU (British Thermal Units) consumption per ton-mile. The BTU con- sumption parameters were taken from secondary sources (Table 25). The three, alternative systems show energy savings relative to the current system, with the most efficient system being the 80-car solution which displayed energy savings of 26 per- cent. The least efficient of the alternatives was the 20-, 50-, 80-car solution which consumed 24 percent less energy than the current system. Sensitivity of Subterminal Organization . to Unfavorable Cost Movements Impact of Increased Farm Storage Cost on Subterminal Organizations The subterminal system requires substantial farm storage in order for it to be a feasible organiza- tion. Some producers may lack facilities or manage- ment skill to maintain grain quality relative to other storage facilities; thus, the impact of increased farm storage costs on the viability of the subterminal organization and the impact of these increased costs on the 50-, 80-car system were determined. The annual variable costs of storing wheat on farms is currently estimated at 8.32 cents per bushel. In the model, cost is assigned to the three time periods by duration of each period. The most critical farm storage period was determined to be the first 66 days, which includes a 21-day harvest and a 45-day period following harvest. Based on earlier analysis of the 50-, 80-car solution, grain was found to be farm- stored during these time periods and then shipped to the subterminals as storage space became availa- ble. Increased farm storage costs were therefore as- signed only to the 21-day harvest period and the following 45 days. Because of possible additional increases in grain shrinkage, farm storage cost was increased by .25 cent per bushel in the first period and .50 cent per bushel in the second period; thus, annual variable farm storage cost increased to 9.07 cents per bushel. In general, the results of this analysis were as anticipated. Increasing farm storage cost reduces the volume handled by subterminals or, conversely, increases the volume moving to inland terminals (Table 26). With the original 50-, 80-car solution, about 49 percent of the wheat destined for the Texas Gulf ports was handled by subterminals; with in- 17 * w; TABLE 25. ESTIMATED ENERGY CONSUMPTION OF THE FOUR SOLUTIONS BY MODE OF TRANSPORT IN MILLIONS OF BTU'S1 1 Car to 1 Car 20-, 50-, and 80- Total inland to car unit trains Commercial Farm energy System terminalsz Gulf ports3 to Gulf ports‘ trucks Barge“ assembly’ consumption 000,000 BTU'S Current 95,991 242,888 0 27,539 0 32,140 398,558 80-Car 43,019 7,207 168,414 45,331 0 32,826 296,797 50-, 80-Car 15,904 6,824 207,730 25,944 0 42,233 298,635 20-, 50-, 80-Car 8,458 3,891 223,287 16,242 2,565 47,052 301,495 ‘Energy consumption estimated for marketing 21.59 million bushels. zBased on estimate of 890 BTU's/ton-mile. 3Based on estimate of 600 BTU's/ton-mile. ‘Based on estimate of 400 BTU's/ton-mile. SBased on estimate of 2323.26 BTU's/ton-mile. “Based on estimate of 500 BTU's/ton-mile. 7Based on estimate of 4195.62 BTU's/ton-mile. O‘ Sources of Rail and Barge BTU consumption were: The Replacement of Alton Lock and Dam 26, September 1979, An Advisory Report of the DOT to the Senate Commerce Committee, and a paper authored by John B. Hopkins, A. T. Newfall, and Martin Hazel, Fuel Consumption in Rail Freight Service: Theory and Practice, DOT, Transportation Systems Center, Cambridge, Massachusetts. The paper was presented at the 56th Annual Meeting of the Transportation Research Board, January 1977. TABLE 26. ESTIMATED QUANTITIES OF 1985 EXPORT WHEAT HANDLED BY SUBTERMINALS BEFORE AND AFTER INCREASED FARM STORAGE COST UNDER THE 50-, 80-CAR SOLUTIONI Total Total Necessary volume/subterminal volume/subterminal Volume to Feasible annualized prior to increasing after increasing subterminals subterminal upgrading farm-storage farm-storage from locations costs costs costs study area ($) (000 bu.) Meade, KS 51,287 1625 1556 586 Liberal, KS 50,721 2714 2596 2346 Buffalo, OK 51,287 2564 2541 398 Perryton, TX 77,768 5562 5244 5244 Stratford, TX 50,721 2756 2160 2160 ‘In this subterminal analysis, Seward County, Kansas, was included. Liberal, Kansas, is located in the southern portion of Seward County, which is adjacent to, Beaver and Texas Counties, Oklahoma, and received substantial quantities from the six-county area. creased farm storage cost, this portion is reduced to approximately 44 percent. Aggregate volume han- dled by subterminals decreased 9 percent. Volume per subterminal did not decrease uniformly. For example, the Stratford subterminal location had a volume decrease of 22 percent, while the Buffalo subterminal location had volume reduced less than 1 percent. Subterminals which are relatively access- ible to inland terminals appeared to lose greatest volume. Those having relatively small costs in mov- ing grain to inland terminals are particularly vulner- able to unfavorable cost changes. As expected, the costs of this solution (46.3 cents per bushel) are increased slightly relative to the original 50-, 80-car solution which had an expect- ed total per-bushel cost of 46.2 cents (Table 27). Impact of Increased Truck-Weight Limit on Subterminal Organization Increasing the truck-weight limit reduces per- bushel transportation cost. For example, increasing 18 the load limit from 860 bushels to 1,100 bushels on a 100-mile haul will reduce per-bushel cost from 15.5 cents to 12.3 cents, or by 20 percent. The purpose of this section of the study was to determine the impact TABLE 27. THE STUDY AREA'S ESTIMATED COSTS FOR MARKET- ING 1985 EXPORT WHEAT UNDER THE 50-, 80-CAR SOLUTION WITH I1NCREASED FARM STORAGE COSTS BY TYPE OF COST, 1977-78 Type of Cost Cost ($) Total Variable Grain-Handling and -Storage Cost 2,584,866 Total Truck-Shipping Cost 2,603,281 Annualized Subterminal-Upgrading Cost 206,741 Additional Grain-Grading Cost at Subterminals 10,734 Total Variable Rail-Shipping Costz 5,499,484 Total Cost 10,905,106 Per-Bushel Cost .463 F“ \ ‘Costs include the six-county area and Seward County, Kansas. Costs are ” for marketing 23.577 million bushels. zTotal Rail-Shipping Cost may be estimated by multiplying the total variable cost by 1.35. \ \ ’\ of increased truck-load limits on the 50-, 80-car Nubterminal organization. a. K L, I‘ A priori, several effects may be hypothesized to result from lowered trucking cost: 1) Reduced truck- ing cost may increase the use of trucks in the assembly of grain to inland terminals for movement on 80-car trains. In which case, subterminals may receive reduced volumes. 2) Lowered trucking cost may make it practical to truck-transport additional wheat directly to Gulf ports and bypass the subter- minal and inland-terminal organization. 3) Transpor- tation of wheat from country elevator to subterminal rmay become more economical; in which case, the volume flowing through the subterminal system would be increased. For purposes of this analysis, truck-weight limit was assumed to increase from 860 bushels to 1,100 bushels, and backhauls were assumed available for all truck movements. Increasing the truck-weight limit from 860 bushels to 1,100 bushels adversely affects the sub- terminal organization. Because of the reduced truck cost, wheat was increasingly shipped by truck to inland terminals. Study-area wheat production des- tined for subterminals decreased 13 percent. Simi- larly, the portion of study-area wheat handled by subterminals decreased from 49 percent (original 50-, 80-car solution) to 43 percent. As in the previous solution involving increased farm storage cost, the subterminals are affected unevenly (Table 28). Volume at all facilities decreased relative to the original solution, except for the Meade, Kansas, subterminal location where expected volume in- creased from 1.625 to 1.994 million bushels. With increased truck-weight limits, a subterminal adja- cent to Meade lost its comparative advantage to that facility, and grain was diverted away from the neigh- boring plant. Truck was the least-cost assembly mode for 71 percent of the wheat flowing to inland terminals. With the original 50-, 80-car solution, this mode assembled approximately 28 percent of the inland- terminal receipts. The analysis revealed that no wheat would move by truck from country elevator to subterminal or to Texas Gulf port terminal. Howev- er, the analysis did indicate that grain would be trucked to the Port of Catoosa for barging to Mississippi River ports. This flow was to meet the predetermined demand at this location which, in the original 50-, 80-car solution, was met by single-car haulage. Table 29 identifies the costs associated with this organization. As expected, the solution's total per- bushel cost (46.1 cents) is less than that expected for the original 50-, 80-car solution (46.2 cents) with 860- bushel truck-weight limits. Impacts of Alternative Distribution Systems on Marketing-System Participants If implemented, each of the three studied grain transportation organizations would affect the mar- keting-system participants differently. With the 80- car solution, farmers and country elevator managers would not be forced to act in a manner that differs substantially from the current system. Country TABLE 29. THE STUDY AREA’S ESTIMATED COSTS FOR MARKET- ING 1985 EXPORT WHEAT UNDER THE 50-, 80-CAR SOLUTION WITH INCREASED TRUCK-WEIGHT LIMITS BY TYPE OF COSTS, 1977-781 Type of Cost Cost ($) Total Variable Grain-Handling and -Storage Cost 2,536,669 Total Truck-Shipping Cost 3,248,912 Annualized Subterminal-Upgrading Cost 180,840 Additional Grain-Grading Cost at Subterminal 10,254 Total Variable Rail-Shipping Costz 4,541,122 Total Barge-Shipping Cost 72,755 Total Cost 10,590,553 Per-Bushel Cost .461 ‘Costs include the six-county area and Seward County, Kansas. Costs are for marketing 22.973 million bushels. rfotal Rail-Shipping Cost may be estimated by multiplying the total variable cost by 1.35. TABLE 28. ESTIMATED QUANTITIES OF 1985 EXPORT WHEAT HANDLED BY SUBTERMINALS BEFORE AND AFTER INCREASED TRUCK- WEIGHT LIMITS UNDER THE 50-, 80-CAR SOLUTlONl Necessary Total Total Volume to Feasible annualized volume/subterminal volume/subterminal subterminals subterminal upgrading prior to increasing after increasing from study location costs truck-load limit truck-load limit area‘ ($) (000 bu.) . Meade, KS 51,287 1625 1994 586 Liberal, KS 50,721 2714 2596 2346 Buffalo, OK 51,287 2564 1568 293 Perryton, TX 50,7212 5562 5130 5130 Stratford, TX 50,721 2756 1899 1899 ‘ -- * ‘In this subterminal analysis, Seward County, Kansas, was included. Liberal, Kansas, is located in the southern portion of Seward County, which is adjacent to Beaver andTexas Counties, Oklahoma, and received substantial quantities from the six-county area. zlt was not feasible for Perryton to upgrade into a facility capable of handling an annual volume in excess of 5.0 million bushels. The associated upgrading cost would have been $77,768. 19 elevators would ship a slightly greater percent of their export-destined wheat volume to inland termi- nals. All wheat destined for the Texas Gulf ports would move through inland terminals for purposes of capturing the more efficient 80-car trains. Al- though the quantity of wheat moving to inland terminals would tend to increase with the 80-car system, the portion handled by the various inland- terminal locations would be substantially altered. In particular, the six-county study area would ship no wheat to Fort Worth, whereas shipments to Amarillo would increase and those to Enid would decrease. Within the current system, about 15 percent of the flow from country elevators to inland terminals is estimated to be truck-transported. If the cost-based, 80-car solution were implemented, trucks would have to haul 25 percent of the country elevator-to- inland terminal flow for optimum efficiency. Percent total ton-miles generated by railroads would then be reduced from 96.5 in the current system to 96 in the 80-car system. Both the 50-, 80-car and the 20-, 50-, 80-car solutions would, in general, affect farmers, country elevators, and inland terminals more so than the 80- car solution. Since these distribution systems in- volve subterminals which receive much of their supply from farm storage, farmers will be required to store their own grain. In addition, farmers would need to assemble grain over greater distances to participate in the subterminal's cost-saving advan- tage made possible by the operation of unit trains. In general, the average distance of farmer assembly increases 2 to 4 miles relative to the current and 80- car-train solutions. Country elevators located near subterminals may be forced to exit the industry, since farmers would likely bypass them in favor of subterminal facilities. Within the current and 80-car systems, 56 country elevators are involved in receiv- ing producer's export-destined wheat. The 50-, 80- car-train solution includes 30 country elevators and five subterminals while the 20-, 50-, 80-car solution includes 21 country elevators and 10 subterminals. Therefore, the number of initial assemblers is re- duced in both of these alternatives. The analysis shows that the volume of wheat received by the inland-terminal industry would be reduced within the subterminal organization, and the portion mov- ing to various inland-terminal locations would be altered relative to the current system. Within the 50-, 80-car and the 20-, 50-, 80-car systems, only 51 and 25 percent of the respective export-destined wheat is estimated to move through inland terminals. Within the subterminal organization, the amount of grain hauled from country elevators to inland terminals by truck would be greater than within the current system but less than within the 80-car system. The analysis indicates that the current grain- handling industry (country elevator and inland ter- minal) could be seriously affected with the adoption of a subterminal organization. Reduced volumes of export-destined grain at country elevators and in- 2O land terminals could jeopardize the value of these facilities unless alternative income sources were found. In this study, analysis focused on export f destined grain flows or about 84 percent of produc- tion and does not indicate least-cost flow patterns for domestically consumed wheat. Perhaps those country elevators not necessary for the handling of export-destined wheat would specialize in the hand- ling and blending of domestically consumed grain. However, it seems likely that some country elevators and most inland terminals would lose in an organiza- tion that includes subterminals. To some extent, the grain-merchandizing prac-fii tices of subterminal management may need to be altered relative to those necessary for country elevator management. The current single-car rate allows management to merchandize wheat as it is purchased from producers and, therefore, pur- chased inventory is not great. In contrast, subter- minal management would be required to accumu- late the purchased inventory to meet the multicar shipment requirements of the unit train. In which case, purchased inventory levels would become substantial, and the risk associated with change in value of inventory would increase relative to the current situation. This risk would likely need to be reduced through acquired futures-hedging skills. For a subterminal system to evolve, the country elevator industry will need to make substantial capital investment to upgrade facilities. The analysis indicated an incentive for this investment; however, it further revealed that the subterminal system would be in strong competition with the inland- terminal industry. Therefore, an unfavorable move- ment in a cost parameter (increased farm storage cost, change in trucking costs) can neutralize the subterminal's cost advantage and create an addition- al investment risk. It would seem that some form of rate assurance by railroads would need to be pro- vided to potential subterminal investors for this system to evolve. Railroads are the principal source of potential system efficiency and therefore are the most critical marketing agents for the remodelling of the current system. Their potential actions can have significant impacts on other system participants, in particular, the grain-handling industry. It would appear that system changes must be initiated by railroads with the understanding that the grain-handling industry must be provided a financial incentive to modify activity and to invest new capital; that is, available” savings to the railroads must, in part, be passed on to grain handlers and, in turn, to farmers in order to obtain system changes. The extent that savings wi ll be passed on to other system participants dependm‘ on the level of competition between railroads and" other transportation modes and requires further study, which is beyond the scope of this analysis. References Baumel, C. P., John l. Miller, and Thomas P. Drinka. A Summary of an Economic Analysis of Upgrading Branch Rail Lines: A Study of 71 Lines in Iowa. FRA-OPPD-76-3. DOT-FR-55045. Iowa State University. March 1976. 2. Fuller, S. W., and C. V. Shanmugham. ”Network Flow Models: Use in Rural Freight Transportation Analysis and a Comparison with Linear Pro- gramming." Southern journal of Agricultural Eco- nomics. Vol. 10. No. 2. December 1978. "3. Hilger, D. A., Bruce A. McCarl, and l. W. Uhring. "Facilities Location: The Case of Grain Subter- minals." American lournal of Agricultural Eco- nomics. Vol. 59. No. 4. November 1977. 4. Larson, Donald W. and Michael D. Kane. ”Effects of Rail Abandonment on Grain Marketing and Transportation Costs in Central and Southwest- ern Ohio." North Central journal of Agricultural Economics. Vol. 1. No. 2. July 1979. Appendix A To provide additional insight into the structure of the network-flow model, a prototype is present- ed. A network model is constructed of nodes repre- senting elements comprising the system — includ- ing production origins, country elevators, subter- minals, inland terminals, the river elevator, and port terminals — and arcs, which connect nodes and include information regarding lower and upper bounds on the arc flow and, in addition, the unit cost of this flow. The prototype model includes two production origins (P=2), one country elevator (C=1), one subterminal (S=1), one inland terminal (l=1), one river elevator (R=1), and one port terminal (E=1). It also includes two time periods (T=2) (Figure A1) which are represented through three points in time. Grain stored from point 1 in time to point 2 will have been stored through the first time period. Grain stored from point 2 in time to point 3 will have been carried through the second period. The nodes of the network are defined as follows: P1,‘: represents production location iat point k in time i = 1, 2, ..., P k =1,2, ...,T+1 C111: represents country elevator i at point k in time i = 1, 2, .. ., C k = 1,2, ...,T+1 1k: represents subterminal i at point k in time i=1, 2, ..., S k= 1,2, ...,T+1 l,-k: represents inland terminal iat point k in time i = 1, 2, ...,l k =1,2, ...,T+1 R11‘: represents river elevator iat point k in time i = 1, 2, ..., R k = 1,2, ...,T+1 E,-k: represents port elevator iat point k in time i=1,2, ...,P k=1,2, ...,T+1 The quantity of grain which may be shipped from a country elevator in the first time period is constrained to available transportation. Accordingly, a set of artificial nodes are created. These are represented in Figure A1 as A,‘ and A1’, where: A}: represents the truck node associated with country elevator iduring harvest, and A1’: represents the rail node associated with country elevator iduring harvest, and i=1,2, ...,C. An additional set of dummy nodes are required to include a subterminal's renovation costs. These are represented in Figure A1 as S, nodes. Each arc connecting the various nodes includes three parameters. These parameters include a lower bound or a required flow, an upper bound or maximum flow, and a cost of unit flow through the arc. The lower bound on all arcs has been set equal to zero except for those linking the source node with the production origin nodes and the port terminal nodes with the sink node (Figure A1). The lower and upper bounds on the arcs terminating at the production origin node (P11 and P12) are set equal to the quantity of harvested wheat at each produc- tion origin which is export-destined. The arcs con- necting the port elevator with the sink node have their lower and upper bounds set equal to the exogenously estimated foreign wheat demand at the port elevator. Arcs connecting the C,-k nodes with the A,‘ and A,’ nodes have their upper bounds set equal to the respective quantity of truck and rail service available at harvest. All remaining arcs have their upper bounds set equal to infinity, except for those arcs representing wheat storage activities. Many arcs include costs of either grain receiv- ing, storage, loading, and/or transportation. The multiperiod or storage characteristics of the model are introduced through the use of storage arcs which link a storage facility through points in time. For example, the vertical arc connecting P11 (produc- tion origin 1, time frame 1) and P12 (production origin 1, time frame 2) represents the first time period and includes the farm storage cost associated with production origin 1. The upper bound on this vertical arc reflects the wheat storage available at this production origin. The vertical storage arcs linking the subterminal, inland terminal, river elevator, and port terminal through alternative time 21 TEXAS AltM UNIVERSITY llllllllllllllll ALHGBB E933 Figure A1. Prototype of network-flow model. _. Q ‘Q PZI \ 5n 5' Q2 S12 9' P23 5:3 5' In fi I12 / / 1:3 / / // PM Rn T R12 R13 EIZ Sink frames include similar types of cost and upper- bound information. If wheat is not farm-stored at harvest-time, it must move to a country elevator or subterminal. For production origin 1, this will involve flow over the arc linking P11 and C11 or over the arc linking P11 with S12. These arcs include farm assembly cost and unloading costs at these respective facilities. Arcs linking the country elevator with the subterminal, inland terminal, and river elevator include the unit loading cost of the appropriate mode, transporta- tion cost, and the unloading cost at the respective facilities. Similar types of costs are included on those arcs which connect the inland terminal, river elevator, and port terminal. To estimate the costs of the currentsystem, rail arcs included only single-car movementi When the feasibility of multicar ship- ments are examined, the rail arcs connecting the inland terminal with the port terminal include 80-car costs, whereas the arcs linking the subterminal with the port elevator include either 20-, 50-, or 80-car cost. Arcs connecting the S”, node (subterminal) and the S, nodes include estimated subterminal unit fixed renovation cost. 22 To include the annual fixed cost associated with renovating a country elevator into a subterminal, a procedure involving a series of iterative computer solutions of the model was carried out. The initial solution of the model included no subterminal renovation costs (i.e., costs on the arcs connecting the S”, with the S, level nodes were set equal to zero). The initial solution provided necessary input to the subsequent computer solution, in particular, the subterminal's annual fixed renovation cost was divided by the subterminal's annual volume as determined by the initial solution. The resulting unit cost parameter was entered on the arc connecting the 51k with the S, nodes, and a solution was again obtained. Again, the annualized cost was divided by the resulting subterminal volume, yielding a new and larger cost parameter, etc. This procedure wa carried out until the volume of the potential subter- minal stabilized or was changed from solution to solution. The iterative procedure was internalized and typically five or six solutions were required pri ; to stabilization. The feasible subterminal's stabilizeu /' volume was one where the unit fixed renovation costs multiplied by the associated volume yielded the total fixed cost. For the unfeasible subterminals, yflhe stabilized volumes were zero. i a The network solution technique requires that a return arc be created, originating at the sink node and terminating at the source node (Figure A1). The lower and upper bounds on this arc are set equal to the total wheat supply which is identical to wheat demand. After construction of the network-flow model, a network algorithm is applied to resolve the least-cost solution. Any network-flow model can be formulated as a linear-programming model. In a linear- ‘programming model, each node is represented by a row (constraint) and each arc by a column (activity). The direction of flow on an arc linking node ito node j is indicated in the linear-programming table as a +1 coefficient in the row for node iand a -1 coefficient in the row for node j. Linear- programming and network-flow models yield identi- cal least-cost solutions. Acknowledgments Several individuals made substantial contribu- tions to this research effort. These include Orlo Sorenson (Kansas State University), who estimated grain handling and storage cost parameters, and Jack Lamkin (Texas Transportation Institute), who estimated rail cost parameters. Others making sig- nificant contributions were Sadler Bridges (Texas Transportation; Institute), Robert Oehrtman (Ok- Uahoma State University), C. Phillip Baumel (Iowa State University), Lowell Hill (University of Illinois), Frank Hardesty (Federal Railroad Administration), and lack Dezik and Darrell Fannin (Texas A&M fnilniversity). Appreciation is expressed to Donald Farris, John Nichols, and Roland Smith for review of the manu- script. Mention of a trademark or proprietary product does not constitute a guarantee or warranty of the product by the Texas Agricultural Experiment Station and does not imply its approval to the exclusion of other products that may also be suitable. All programs and information of the Texas Agricultural Experiment Station are available to everyone without regard to race, ethnic origin, religion, sex, or age. The Texas Agricultural Experiment Station, Neville P. Clarke, Director, College Station, Texas. 2M — 11-80