.IO'; ‘ .‘."‘J‘ 0'; .§‘|.: K. - ‘0‘ n'. £945 , , “0 o u 3,“: °' W | '. . fit! J 3’1. ! ll- .LI‘. . (II... I -.'.u .I" at: i ii. _. A. III. ‘1'! I I L 3 1293 10498 2552 .51“ " ‘ ' THESIS Pg": . i EFL“ "‘ '~‘. Mam-tram This is to certify that the dissertation entitled USING A REGIONAL INPUT-OUTPUT MODEL TO FORECAST RAIL FREIGHT TRAFFIC: WITH APPLICATIONS FOR THE SUBSIDY-ABANDONMENT DECISION presented by Jeffrey L. Jordan has been accepted towards fulfillment of the requirements for Ph.D. degree in Aqricul tural‘ Economics 3 5 aajor professor Date August 6, 1982 MS U i: an Affirmative Action/Equal Opportunity Institution 012771 ‘1'. Wmus‘ ‘ ‘ MSU LIBRARIES RETURNING MATERIALS: Place in book drop to remove this checkout from your record. EIN§§_wiTT be charged if book is returned after the date stamped below. _____,___....———---— M USING A REGIONAL INPUT-OUTPUT MODEL TO FORECAST RAIL FREIGHT TRAFFIC: WITH APPLICATIONS FOR THE SUBSIDY-ABANDONMENT DECISION BY Jeffrey L. Jordan A DISSERTATION Submitted to Michigan State University in partiai fulfiiiment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Agricuiturai Economics 1982 ABSTRACT USING A REGIONAL INPUT-OUTPUT MODEL TO FORECAST RAIL FREIGHT TRAFFIC: WITH APPLICATIONS FOR THE SUBSIDY-ABANDONMENT DECISION By Jeffrey L. Jordan The purpose of this study is to suggest and test a method that systematically links the level of statewide economic activity with the demand for rail freight tranSportation services and facilities. A 20-sector input-output model of the State of Michigan is combined with commodity flow data obtained from expanded, l percent waybills from all of Michigan's railroad lines. The test of the forecasting method is whether the 20-sector in- put-output model of Michigan can provide reliable estimates of rail traffic over a four-year period. The test is carried out not only on total rail traffic, but also on a sector-by-sector basis. The ability of the input-output model to disaggregate output projections is the key to its use in projecting rail traffic. This is so because the derived demand for rail services in a region, or on individual lines, depends in part, on the output of the particular commodities that are produced in the area. The results of the test of the rail forecasting method indicate that it is effective in producing estimates of rail traffic. The model projects rail traffic to within 1.2 percent of actual traffic over the Jeffrey L. Jordan total rail network. Additionally, the model is effective in providing sector-by-sector estimates in ll of the l3 sectors for which rail traf- fic is represented in l976 and l980. Following the testing of the forecasting procedure, the study pro- vides an illustration of the use of the model in regional rail planning decisions. The case of Michigan's Upper Peninsula is used to demon- strate the usefulness of having projections of rail traffic when making subsidy-abandonment, or rail rationalization decisions. The forecasting procedure is also used to examine a specific issue on one rail segment in the Upper Peninsula: the northern 300 Line. This illustration of the uses of the input-output model emphasizes the flexibility of the method in dealing with individual rail issues. Not only is the input-output model able to disaggregate the effects of changes in specific commodity production, but through the waybill sample, it is able to isolate particular rail segments. Dedicated to the memory of my father Gerald J. Jordan who instilled in me an appreciation for the written word 11' her 5; like t disserl I TAOADSt DVOiess tion t1 pAOies Harry DeDart Statio the di ACKNOWLEDGMENTS I would first like to acknowledge and thank Sherry A. Maddock for her support and good humor over the last eight years. I would also like to acknowledge her editorial assistance during the writing of this dissertation. I would like to thank my dissertation supervisor, Stanley R. Thompson, for his assistance and guidance, as well as my major professor, J. Roy Black. I would also like to give a word of apprecia- tion to my dissertation and program committee members including Professors Thompson and Black, James D. Shaffer, Kenneth D. Boyer, Harry Trebing, and A. Allan Schmid. Further, I would like to thank the Department of Agricultural Economics and the Agricultural Experiment Station, Michigan State University, for the financial support during the dissertation writing stage. Also very much appreciated were the able and diligent efforts of Nancy Creed, whose patience during the typing of this dissertation will always be remembered. A word of thanks is also due to James Pratt and Daniel Chappelle for their assistance and instruction in the use of input-output methods. I would like to pass on a very warm thank you to those at the Michigan Department of Transportation, without whose assistance and advice this dissertation would have been impossible: Richard Esch and Joyce Newell of the Transportation Planning Procedures Section, and Larry Britton and Jared Becker of the Railroad Planning Section. iii Hesa i. of voH this 2 Skills VAiUdI not not UDIVer Finally, a special word of thanks is appropriate to Mr. Jon M. Wesa of the Transportation Planning Procedures Section. The opportunity of working with Jon over the past few months has made the writing of this dissertation both rewarding and pleasurable. His programming skills and vast knowledge of the Michigan rail system have been in— valuable to the completion of this project. Without his help, I would not now be concluding my beneficial association with Michigan State University. iv DEDICA Acne.- LIST 0 LIST 0 CHAPTE I r—4 TABLE OF CONTENTS DEDICATION ............................ ACKNOWLEDGMENTS .......................... LIST OF TABLES .......................... LIST OF FIGURES .......................... CHAPTER I. INTRODUCTION ....................... Problem Setting ..................... Purpose and Objectives ................. The Input-Output Method ................. Applications to the Subsidy-Abandonment Decision . . . The Upper Peninsula Rail System: Applying the Method for the Subsidy-Abandonment Decision ...... Forecasting Procedure .................. Organization of Study .................. II. REVIEW OF LITERATURE ................... The Structural-Aggregate Approach ............ The Micro-Disaggregate Approach ............. Rail Traffic Forecasting Methods and Rail Planning in Michigan ................. Forecasting Rail Traffic and the Potential for Profitability of Currently Subsidized Lines ...... Conclusion ....................... III. THE MICHIGAN RAIL SYSTEM ................. The Upper Peninsula Rail System ............. IV. METHOD AND ECONOMIC FRAMEWORK ............... Framework of Study ................... The Structure of the Input-Output Model ......... Assumptions of Input-Output ............... Construction of the ZO-Sector Michigan Input-Output Model .................. Two-Region Model .................... Data, Updating, and Aggregation of the Michigan Input-Output Model .............. Output Multipliers and Projection of Total Output. . . . Linking the Input-Output Table to Commodity Flow Data. . The Use of Input-Output for Transportation Planning. . . V iii vii ix mag —-l—-l —"\IO\ -'O\O \lhw—‘d 25 VI. M "ZJZ‘IQ; CHAPTER V. TESTING THE USE OF INPUT-OUTPUT FOR RAIL TRAFFIC FORECASTING ....................... Estimating 1980 Final Demands .............. Personal Consumption Expenditures ........... Gross Private Capital Formation ............ Net Inventory Change ................. Net Exports ...................... Federal Government .................. State and Local Government .............. Testing the Rail Traffic Forecasting Method ....... Conclusions ....................... VI. APPLYING THE RAIL FORECASTING METHOD TO REGIONAL RAIL PLANNING ...................... The Upper Peninsula Rail System ............. Service Needs in the Upper Peninsula and the 500 Line ....................... Projecting Economic Activity, 1980-1983 and 1986 . . . . Projecting Upper Peninsula Rail Traffic, 1980-1983 and 1986 .................. The 500 Line and the North-South Route Issue ...... VII. CONCLUSIONS ........................ Implications for the Use of the Method in Rail Planning ....................... Implications for Further Research ............ APPENDIX 1. 1976, ZO-Sector Input-Output Model of Michigan with an Example of its Use .................. II. 1976, 44-Sector Input-Output Model of Michigan ...... III. 1983 and 1986 Projections of State Economic Activity and Total Rail Traffic ................. BIBLIOGRAPHY ........................... vi Page 93 100 103 106 111 122 125 127 130 142 164 167 TABLE 4.1 4.2 ' 4.3 4.4 4.5 4.6 4.7 5.1 5.2 5.3 5.1 5.2 6.3 6.4 LIST OF TABLES Page TABLE 3.1 Subsidy Phase-Out Schedule ............... 41 4.1 Sectors Defined by SIC, 1976 I/O Table for Michigan ..................... 63 4.2 1976, 20-Sector Input-Output Classification ....... 65 4.3 Output Multipliers Derived from the Michigan 20-Sector Input-Output Model ............. 67 4.4 Effect of a $1,697,000 Change in the Final Demand of Livestock and Products on the Entire Economy ..... 68 4.5 1976 Final Demands and Total Outputs Derived from the Michigan Input-Output Model (Thousands of 1976 Dollars) ....................... 70 4.6 STCC Codes in Each Input-Output Sector ......... 74 4.7 Expanded 1 Percent Waybills, Michigan, 1976, 1980, by Input-Output Sector ................ 75 5.1 4 1980 Total Output Projections: Michigan (Thousands of 1976 Dollars) ................... 84, 5.2 Expanded 1 Percent Waybills, Michigan, 1976, 1980, by Input-Output Sector ................ 85 5.3 Comparison Between 1980 Actual and Projected Rail Traffic Movements in Michigan (Tons)' ......... 87 6.1 Upper Peninsula Rail Line Abandonments Since April 1, 1976 ......................... 97 6.2 Upper Peninsula Rail Lines Currently Under Study for Possible Abandonment Applications ........... 98 6.3 Upper Peninsula Rail Lines Subject to a Pending Abandonment Application ................ 99 6.4 1983 Projection of State Economic Activity and Upper Peninsula Rail Traffic: High Projection ....... 105 vii TABLE 6.5 6.6 6.1 6.8 6.9 A1.1 41.2 TABLE >>>>>>>>> 6.5 6.6 _.a #0)“) O .—l 01wa 1983 Projection of State Economic Activity and Upper Peninsula Rail Traffic: Medium Projection ..... l986 Projection of State Economic Activity and Upper Peninsula Rail Traffic: High Projection ...... 1983 Projection of State Economic Activity and 500 Line Rail Traffic: Medium Projection ........ 1983 Projection of State Economic Activity and S00 Line Rail Traffic: High Projection ......... 1986 Projection of State Economic Activity and 500 Line Rail Traffic: High Projection ......... 20-Sector Transactions Matrix ............. 20-Sector Coefficient Matrix ............. 20-Sector Leontief Matrix ............... 20-Sector Inverse Matrix ............... 44-Sector Transactions Matrix ............. 44-Sector Coefficient Matrix ............. 44-Sector Leontief Matrix ............... 44-Sector Inverse Matrix ............... 1976 and 1980 Total Output Projections: Michigan 44 Sectors (Thousands of Dollars) .......... Comparison Between 1980 Actual and Projected Rail Traffic Movements in Michigan, 44 Sectors (Tons) . . . 1983 Projection of State Economic Activity and Total State Rail Traffic: Medium Projection ....... 1983 Projection of State Economic Activity and Total State Rail Traffic: High Projection ........ l986 Projection of State Economic Activity and Total State Rail Traffic: High Projection ........ viii Page 107 108 115 116 117 132 135 138 140 142 147 152 157 162 163 164 165 166 FIGURE 1 #hwwww (350303050101 01-wa .1 LIST OF FIGURES Flow Diagram of Rail Traffic Forecasting Method . . . . Rail Rationalization Plan, Rail Investment Decision Analysis Model, Tier II Flow Diagram ......... Michigan Abandonments Since 4/1/76 ........... State Subsidized Lines, 1981 .............. Pending Abandonments as of 11/24/81 .......... Lines Identified by Carriers as Under Study ...... Structure of the Intersectoral Transaction Table. . . . Structure of the Interregional Intersectoral Transaction Table of the Two-Region Input-Output Model ........................ Flow Diagram of Rail Traffic Forecasting Test ..... Railroads in the Upper Peninsula ............ Upper Peninsula Lines Possibly in Jeopardy ....... Existing Needs Map ................... North-South Alternatives--Soo Line ........... Frequency Distribution of Rail Density on All Michigan Railroads Classified in ICC Categories 1-4 .................... ix Page 12 28 37 38 39 40 49 59 78 96 101 102 112 119 caused issues Hichi; roads. facturm goods, ' the sea ACI of Referm a”31m ShOuld Ra the bani Rai1r0a( in Mich- Wat ,a. f1“Acig bank-IUD: federan CHAPTER I INTRODUCTION Problem Settigg In the last decade, the financial difficulties of railroads have caused federal and state governments to become increasingly involved in issues associated with preserving essential levels of rail service. In Michigan, 16,500 people are directly employed by the 32 operating rail- roads. Railroads are the dominant means of transportation for manu- factured products, carrying 67 percent of the state's manufactured goods, while Michigan's agricultural commodities also depend heavily on the service of the railroads. Since the Regional Rail Reorganization Act of 1973 (3R Act) and the Railroad Revitalization and Regulatory Reform Act of 1976 (4R Act), state governments have been charged with analyzing their rail systems and establishing priorities for lines that should be retained under subsidy. Rail freight planning began in Michigan in 1974 'hi response to the bankruptcies of the Penn Central Railroad (PC) and the Ann Arbor Railroad (AA), which together comprised 35 percent of the rail mileage in Michigan. The 3R Act provided for the reorganization of six bank- rupt railroads in the Northeast and Midwest region, and for federal financial assistance to continue rail freight service on lines of the bankrupt carriers (PC and AA in Michigan) which were excluded from the federally reorganized ConRail system. The Michigan Railroad Plan, 1 Phase in J; Michi tance tion . tiers so1ver miles aband ”Dre r abando land r 204 1a 16nd ‘ aband< SO1VEl 2 Phase II (revised and approved by the Federal Railroad Administration in June and August of 1976, respectively) was prepared to insure that Michigan would be eligible for federal rail service continuation assis- tance under Section 402 of the 3R Act. Implementation of the 1976 Michigan Railroad Plan provided reten- tion of service on 921 land miles of the 1,049 miles of bankrupt car- riers which were not absorbed into the ConRail system or acquired by solvent carriers. Upon passage of the 4R Act, another 419 land route miles operated by solvent carriers were identified as being subject to abandonment proceedings. From that time to the end of the decade, 318 more route miles were filed for abandonment and 185 miles were actually abandoned. As of September 1, 1930, the Michigan rail system consisted of 838 land route miles operated with state or federal rail service assistance, 204 land route miles which were subject to abandonment proceedings, 589 land route miles which were considered by carriers to be candidates for abandonment, and approximately 4,397 land route miles operated by solvent carriers and not potentially subject to or pending abandonment. In Michigan, the state's 1981 Appropriations Act specifies a phase-out schedule which reduces the 1981-82 state rail freight operat- ing assistance appropriation for currently subsidized lines by 25 per— cent, each year for four years. By 1985-86, the state's subsidies to these rail lines will be discontinued. This is in addition to a reduc- tion in federal support. While the Rail Service Act of 1978 provides subsidies through 1983 on some lines, little federal money has been appropriated. In the late 19705, federal rail subsidies were between $4-6 million per year. In 1982, that figure is $900,000, which funds 3 only a few capital projects. Michigan is currently in the process of submitting a request for some of this capital project funding. As federal and state funds disappear, rail planners in Michigan and else- where are faced with the need to define an essential core of rail service and to determine which rail lines will receive the declining financial support. Consequently, rail planners require methods to estimate rail traffic and to predict profitability of currently sub- sidized lines. Purpose and Objectives The purpose of this study is to suggest and test a method that systematically links the level of statewide economic activity with the demand for rail freight transportation services and facilities. A 20-sector input-output model of the State of Michigan will be combined with commodity flow data obtained from expanded 1 percent waybills from all of Michigan's railroad lines. The method that is discussed and tested in this study is designed to aid rail planners in forecasting rail traffic based on economic activity in the state. In Michigan, as elsewhere, rail traffic forecasting has been predominantly ad hoc in nature and not directly related to the state of the general economy. Not only is a systematic method of predicting rail traffic necessary for rail planning, but traffic projections are needed to analyze the issue of whether to subsidize a rail line or allow its abandonment. In accomplishing this purpose, the following objectives will be achieved: (1) The final demand sectors of the Michigan economy will be estimated for 1980. 4 (2) These final demands will be used to forecast total output, by sector, for 1980 from the 1976 Michigan input-output mode1.l/ (3) The changes in total output will be used to forecast the difference in rail traffic between 1976 and 1980. (4) The forecast results will be checked against the actual rail flows in 1980 to determine the forecasting capability of the input-output method. (5) The method will be used to forecast rail traffic, over various economic growth scenarios, to 1983 and 1986. Special attention will be paid to the rail system in Michigan's Upper Peninsula. (6) Finally, there will be an illustration of the method's usefulness in the rail planning process. The Input-Output Method The use of an input-output model to predict rail traffic is based on the relationship between the production and consumption of goods in an economy and the demand for transportation services and facilities. In order to produce a product, a given quantity of inputs is needed. Although economies of scale may be present in some industries, to pro- duce more of a product (or less) a larger (or smaller) supply of inputs is required. A production facility must acquire and store these inputs. A retail firm can also be viewed as a production facility, where l/Model obtained from Michigan Department of Commerce, Energy Administration, "A 44-Sector Input-Output Model of the Michigan EconomyJ' Energy Issues Analysis Series, Lansing, Michigan, March 1980. 5 production is the sale to final consumers. The process of acquiring inputs involves transportation. With the knowledge of inputs required for production of a good, and amounts required, it is then possible to determine what inputs (the output of other industries) and what volume must be transported. Freight transportation is a factor of production and is not desired in its own right, but only for the service it pro- vides in the production process. The demand for transportation is, in fact, a derived demand--derived from the level of production and econom- ic activity in a region or state. The input-output method uses a system of accounts that represent the transactions among the sectors of the economy. These determine the interrelationships that exist between sectors of the Michigan economy and the relationship between these and the economic activity outside of the state (Rest of Nation, RON). By using this accounting method, it is possible to show how outputs in dollars from each producing sector are distributed among other sectors in Michigan and RON. In terms of transportation requirements, the input-output table also shows how (and from whom) each sector procures its inputs (both in Michigan and RON). Input-output models are consistent in that the outputs of each sector must conform with the outputs of all other sectors from whom it buys inputs and to whom it sells outputs, as well as with the total output of the economy. It is therefbre possible to trace the interdependence in the economy and the required flow of goods needed 'hi the input-out- put process. Once the inputs and outputs that must move throughout the state are known (given the level of economic activity), it is necessary to link that information with data on rail transportation activity. The 6 characteristic of consistency in the input-output model must be matched by the transportation data. This requires that total product movements in the transportation sector correspond with the dollar flows between sectors in the input-output model. To accomplish this, an input-output model of Michigan will be linked with commodity flow data to determine movements of inputs on the state's rail network. Each of the 20 sectors in the input-output model will be matched to commodity classifications on the waybill tapes (the Standard Industrial Codes in the model are matched to the Standard Transportation Commodity Classification codes of the transportation data, SICs and STCCs). The input-output model represents the economy's structure and predicts the annual usage rate of commodities by industry. The model indicates who the suppliers of an industry are, who demands their product, and what their particular input-output relationship is to other sectors. The initial assumption is that economic activity in a particular sector or region generates the demand for the physical movement of the aggregate of products included in each of its input sectors in proportion to the change in that sector's output. Thus, corresponding to each of the dollar flows present in the input-output model, originating and receiving sectors can be identified on the com- modity waybill tapes. To forecast rail traffic for any year, it is necessary to estimate the final demands, by sector, for the forecast year. These final demands are used to "drive" the 1976 Michigan input-output table, pro- ducing projections of total output, by sector. This projection is ac- complished through the multiplication of the new final demands by the Leontief inverse matrix derived from the input-output table. After the tote the the the EC nu AH. 7 total outputs are generated, rail movements will be predicted based on the increase or decrease in output for each commodity. For example, if the final demand in the automobile industry is decreased by 10 percent, the total output in its input sectors will also decline. This decline represents a decrease in the flow of input goods on the rail lines that deliver goods to the industry. If the use of foundry output decreases by 3 percent due to the decline in the automobile industry, then on each rail segment that moves foundry output, projected rail movements will be decreased by 3 percent. This procedure will be used to adjust the movement of goods on the state's rail lines given the changes in total economic activity. Applications to the Subsidy-Abandonment Decision To determine whether a Michigan rail line should be subsidized or abandoned (the investment-disinvestment decision), the Michigan Depart- ment of Transportation's primary objective is to identify the minimum acceptable levels of rail service along endangered lines, demonstrating either profitability to the operator or social and/or economic benefits to communities. The approach mandated by the yearly "Michigan Railroad Plan" is the use of benefit-cost impact criteria. A decision to subsidize a branch line reflects an explicit judgment by decision makers that adverse consequences of a rail abandonment justify the cost of subsidization. While this benefit-cost approach is essential in individual abandonment proceedings, decision makers viewing the rail system from a statewide perspective require a more comprehensive framework which considers spatial and commodity aspects. The procedure suggested in this study seeks to provide the link between the total frei Mich in t1 a re the! and the caus that of i the t10r CFUC r611 ECOr 8 freight transportation system and the economic activity throughout Michigan. The abandonment of a rail branch line reflects not only a change in the supply of freight transportation, but also in its demand. When a rail branch line is abandoned, it demonstrates that the demand for the operator's service has declined, resulting in revenues below costs and thus, abandonment. Therefore, abandonment reflects the fact that the demand for goods in the entire economy or region has declined causing a decline in the derived demand for transportation. It follows that linking a model of the economy (representing the requisite amounts of input and output) with data on how these commodities move throughout the state, will enable planners to address the demand for transporta- tion services and facilities. This study is designed to aid in the answering of two questions crucial to the understanding of the relationship between the demand for rail service and the demand for intermediate and final goods in the economy: (1) Given the necessary volume of goods that must flow over a particular rail segment in order for it to produce revenues over costs, will the projected level of economic activity in the state or region generate enough demand for inter- mediate and final goods to meet this volume? If projected statewide economic activity will generate such a volume, then the rail segment would be a candidate for subsidiza- tion until the projected demand occurs. If statewide activity will not generate enough volume, the line may be subject to abandonment. 9 (2) The corollary to the first question is, given projections of statewide economic activity, what will be the volume of the flow of goods generated for individual rail segments? Answering this question will help to identify which rail segments will eventually become profitable and which will continue to operate at a loss, providing for long-term planning. The immediate problem facing rail planners is the loss of subsidy money over the next few years. Consequently, they are faced with a yearly decision on which lines to continue to subsidize through 1986. Interim subsidies on lines which may never become profitable may not be an "efficient" use of subsidy funds. Thus, the motivation for using the method of this study is that rail planners must know which lines have a probability of becoming profitable and which do not. While service may be justifiable on a line in a more comprehensive benefit- cost basis, the method suggested here will help determine whether a line would require permanent operating assistance. Thus, a method that can predict rail traffic can be of value in the subsidy-abandonment process. The Upper Peninsula Rail System: Applying the Method for the SubsidyéAbandonment Decisién A demonstration of the model's usefulness in predicting rail move— ments and in aiding the subsidy-abandonment process will be applied to Michigan's Upper Peninsula. Michigan's Upper Peninsula was chosen as the case study for two reasons. First, the U.P. rail system is cur- rently (spring 1982) the subject of study by the Michigan Department of Transportation. This study is part of the rail rationalization process 10 of the transportation department, which includes decisions on sub- sidies and abandonments, requiring forecasts of U.P. rail traffic. Consequently, the U.P. case study is timely in that it can aid the decision-making process already underway. Further, the results of this study can be compared with the results of the rail rationalization report as a check on both methods. Additionally, the U.P. rail system is being used as a case study because of transportation data that has recently become available. The Marine-Ports Planning Section of the Michigan Department of Transporta- tion conducted a shipper survey of the U.P. in 1981. This data repre- sents nearly a 100 percent survey of resource-based and manufacturing traffic generated in the U.P. Thus, it is possible to use both the rail waybill samples provided by the Federal Railroad Administration and this primary data source to demonstrate the method of forecasting U.P. rail traffic. The primary data will be used as a check against the waybill sample. ForecastinggProcedure Once the 20-sector input-output model of Michigan is linked to rail data (expanded 1 percent waybills) by STCC and SIC codes, it is possible to use the model to project rail traffic and to aid in the subsidy-aban- donment decision process. First, the model will be tested to determine whether it can accurately forecast rail traffic, given known rail move— ments. The test will consist of projecting 1980 rail traffic on the basis of 1976, 1 percent waybill data and comparing that projection to the waybill sample observed for 1980. Second, the model will be used to forecast total state rail traffic and Upper Peninsula rail traffic beyond 1980 cedu 1976 sect 1976 put- outp put 11 1980, with application to the subsidy-abandonment process. The pro- cedure that will be used in both cases, with modifications, is outlined in Figure 1.1. To forecast rail traffic for any year, the starting point is the 1976, 20-sector input-output transactions table. Final demands, by sector, are estimated for the forecast year and are substituted for the 1976 final demands. Total output, by sector, is estimated by the in- put-out model through the multiplication of the new final demands by the output multipliers of the inverse matrix (I-A)'1. Projected total out- put is then taken as a percentage of 1976 output. The 1 percent waybills (expanded by 100) representing total rail traffic in 1976 are then multiplied by sector, by the real percentage changes forecast by the input-output model. This procedure yields a projection of rail traffic for the forecast year. Finally, the actual traffic of the forecast year is compared to the projected traffic to judge how accurately, both by totals and by commodity, the model pro- jects rail traffic. Organization of Study The literature review in Chapter II will focus on recent litera- ture dealing with rail traffic forecasting methods. The review will also examine studies on rail abandonments, particularly as they apply to Michigan. For the purpose of this review, the literature is divided into that which takes a structural-aggregate approach to rail freight forecasting and that which views the issue from a micro or firm level. Chapter II also includes a discussion of the estimation of rail freight traffic and the rail planning process in Michigan, with particular emphasis on a description of Michigan's two-tier rail rationalization 12 Figure 1.1 Flow Diagram of Rail Traffic Forecasting Method 20-Sector MI I/O Estimation of FinalDemands For Future Years Leontief In- verse Matrix Times New Final Demands Projections of Total Output: Percent of 1976 CD Base Year 1% Waybills Ex- panded: Mul - tiply by 100 ——-> 13 Base Year 1% Expanded Waybills Sector Percent Change Times Base Year Ex- panded Waybills W Projection of Rail Traffic Projected Rail Traffic Minus Actual Divided By Actual Jraffic WV Mode] for Rail Pro'ections ch till che mi l4 procedure. This latter part of Chapter II confirms the need for a more systematic method of estimating rail freight traffic. Chapter III is a description of the Michigan rail system, with an emphasis on those rail segments and regions where abandonments or sub- sidization have occurred or are likely to occur. The need for esti- mates of future rail traffic due to the reductions in subsidy appropria- tions mandated by Michigan's legislature is examined. Further, Chapter III provides a description of the Upper Peninsula rail system, includ- ing a discussion of the operations of each railroad and the traffic characteristics of the U.P. In Chapter IV, an examination is made of the economic framework of the study and the method suggested for estimating rail traffic. This chapter includes the rationale for using the input-output method for rail planning. The description of the 20-sector input—output model in- cludes an explanation of how the input-output method projects an economy's total output through the use of the Leontief inverse matrix. After a discussion of the assumptions behind the input-output table, the construction of the table itself is examined, based on work at the University of Minnesota by Hwang and Maki. Included in this part of Chapter IV are the data requirements, updating, and aggregation pro- cedures used to build the 1976 Michigan input-output model. Finally, Chapter IV includes a description of the procedure used to link the input-output table to commodity flow data and a description of the waybill sampling procedure. The forecasting ability of the input-output model is tested in Chapter V using the method outlined in Figure 1.1. Included is a description of the estimation of the 1980 final demands used to be in der fic OVe sub 198 com rev: "BY for line more IDQy l“Or f 15 forecast the total output percentages that predict rail traffic. The purpose of Chapter V is to show how accurately the link between the input-output model and commodity flow data predicts rail flows. Based on this knowledge, the model will be used to forecast rail traffic beyond the test period of 1976-1980. In Chapter VI an illustration of the use of the forecasting method for predicting rail traffic, as well as suggested applications to the subsidy-abandonment decision, is presented. Using estimates of final demands, a forecast of total U.P. rail traffic and 500 Line rail traf- fic for 1983 will be presented. These final demands will be estimated over various growth ranges. The procedure will be applied to the subsidy-abandonment decision by forecasting a high growth scenario to 1986. This forecast of rail traffic on selected segments will be compared to the rail traffic believed to be necessary to produce revenues over costs. It will be suggested that rail segments which may not be able to generate sufficient traffic, given high estimates for 1986, are more likely to be candidates for abandonments, whereas lines that may be able to generate enough traffic to be profitable are more likely to be candidates for continued subsidization in order that they be maintained until profitable, other things being equal. Finally, Chapter VII will conclude with a discussion of the use of the study's results for transportation policy analysis and implications for further research. CHAPTER II REVIEW OF LITERATURE In this chapter, a review of recent literature on rail traffic forecasting methods will be presented. Also included is a brief review of rail abandonment literature, particularly as it applies to Michigan. Finally, there will be a review of methods that transportation planners currently use to estimate rail freight traffic. It was not until the 19605 that a significant volume of work began to appear on the application of theory and empirical measurement of the demand for freight transportation services (Johnson, 1975). In general, the literature agrees on two points: (1) The demand for freight transportation is not simply a price—quantity issue; the determinants of demand involve factors such as travel time reliability, loss and damage, frequency and size of shipment, etc. A large body of the literature uses multiple regression analysis to distinguish the relative influence of each determinant. (2) Freight transportation is a factor of production and is not desired in its own right, but only for the service it provides in the production process. The demand for transportation is best described as a derived demand. The literature diverges, however, on the question of the proper level of aggregation from which to estimate the demand for freight 16 trm Stm prm ent abh to: Ira dis pr01 tha‘ TLe "Phe to c This econ The Disa tion rail mOVen CODSt ho“ 9 these raj] D l7 transportation. One approach considers the demand within the broader structure of an economy or region. This "structural-aggregate" ap- proach regards the derived demand for freight transportation as depend- ent on the state of the rest of the economy. Thus, structural vari- ables describing macroeconomic activity in different regions are used to explain demand. A second approach regards the demand fOr freight transportation as the sum of individual shipper decisions. This "micro- disaggregate" approach examines the specific inputs and outputs of each production process. In these studies, determining the volume of cargo that will flow in a given market is the starting point for any analysis. The Structural-Aggregate Approach The Canadian Transport Commission (CTC) reported in June 1976 on "Phase II" of their Program, employing an econometric forecasting model to determine the demand for freight transport in Canada (Hariton, 1976). This aggregate model attempts to link changes in broadly defined macro- economic variables to transport demand for specific commodity groups. The model concentrates on rail and marine modes and uses the Canadian Disaggregate Interdepartmental Econometric Model. In terms of aggrega- tion, it divides Canada into five regions, plus the United States (for rail and truck movements), adding nine regions of the world for marine movements. In terms of commodity aggregation, commodity classes were constructed to be as large as possible, while still retaining their homogeneity. In the report, 22 commodity groups were treated, most of these comprise bulk commodities totaling over 80 percent of Canadian rail movements and over 90 percent of total marine movements. Whl d93t trans 18 The CTC model considers the demand for freight transport as derived from the demand for interregional trade. The model is con- cerned with the movement of a given commodity from a given origin to a given destination by a given mode. The number of tons of commodity 9 moved from origin i to destination j by mode m during time period t is denoted by V The model introduces the notion of a "link“ as gimt' being characterized by a given origin, a given destination, and a given mode. Then, V denotes the movement of commodity 9 along link ijm gijmt at time t. Five main causal factors of the demand for freight transport for a given movement were identified: (1) excess production at the origin; (2) excess consumption at the destination; (3) transport costs; (4) com- plementary movements; (5) competitive (substitute) movements. In algebraic notation: vgijmt = f(EPgit’ Ecgjt’ Rgijmt’ vgi'j'm't) where Vgijmt is defined above. Epgit = excess production of commodity g in the origin region during time t, where Epgit is Productiongit - Consumptiongit - ASgit ‘(AS = stockpile of commodity g at start of period). git EC = excess consumption of commodity g in destination region git . . = - , - ' . + . . during time t, where ECth Consumptionth Productionth ASth R = the cost of transporting the commodity g from origin i to gijmt destination j by mode m, where freight rates are the only component of transport costs. a1 Su de' reg to: eclu COmi the IDnn 1:10”1 19 V = the links that are either complementary or competitive. gi'j'm't Certain movements of commodity g are complementary, resulting from transshipments either between modes or within a given mode (e.g., from lakers to ocean liners). These movements have either their origin the same as the destination or their destination the same as the origin. Certain other movements are substitutes, resulting from alternative op- portunities: (l) to ship to different destinations; (2) to receive from different origins; or (3) to use different modes. The CTC model uses macro level variables (i.e., activity in a region) with an accent on structural economic variables as the independ- ent variables. In terms of dependent variables, the model measured, in tons, the amount shipped between regions, using a linear approximation as the functional form. The regression analysis considered alternative origins, destinations, and modes, but not different commodities (see also Canadian Transport Commission Rgpggg, 1975). As part of the Northeast Corridor Transportation Project, the con- sulting firm Mathematica (Vol. II, 1969; also reported in Terziev, 1976) developed a four-stage aggregate model based on a system of linear regression equations. The Mathematica model began by projecting the total production of 16 commodity groups with 16 separate regression equations. The regressions included a time variable and projections of components of Gross National Product as independent variables. Next, the model projected the regional share of originating and terminating tonnage in each commodity group. In order to project regional demand, a regression model was used having as its independent variables popula- tion, retail sales, regional income, and per capita income. di re ti en je Cc je mi f1 SP CO 20 Two commodity groups were allocated to different regions and a distribution model was used to predict the flows of goods between regions. This involved two phases, beginning with a regression model which included as independent variables: consumption at the destina- tion points, production at the origin, distance, and population and employment at the destination. It was found that regional flows pro- jected in this manner did not match the total flows projected originally. Consequently, a flow adjustment algorithm was used to make the two pro- jections consistent. The adjustment used Lagrange multipliers to minimize the flow adjustments, subject to the constraints on the total flow. The fourth stage of the Mathematica model dealt with the modal split. A market share regression was used for rail, private truck, common carriers, air, water, and "other." The independent variables included the fraction of shipments falling into each of five weight groups, the fraction of shipments in each distance group, and the value and average gross revenue per ton of each commodity. Terziev (1976) points out that there are two major drawbacks to this sort of sequential modeling system which attempts to address a full range of freight shipment decisions. First, some choices are made jointly rather than sequentially; for example,the choice of mode and shipment size. Secondly, when two decisions are not made jointly, there is still some feedback from short—run decisions to long-run decisions. These drawbacks brought forth joint demand models that attempt to model several choices in the same equation. As part of his study in 1964, Perle estimated a joint demand model using rail and truck flows in five commodity groups, in nine regions, during each of 21 five years. Perle estimated a truck and rail model from the form: 9 109(Vm1) = 80 + 81 109(rm1) + 82 [109(rmzll + .2 c1-Ri 1=1 + g d.Y. + g f C i=1 ~13 k=1 k" where le = volume of traffic carried by mode ml; rm] = average revenue/ton on mode m1; rm2 = average revenue/ton on mode m2; R1 = l for region i, 0 otherwise; Yj = l for year j, 0 otherwise; Ck = l for commodity k, 0 otherwise. The results showed poor R2 and t-statistics, partly due to the fact that the dependent variable included both the choice of production level and the choice of mode. However, as Terziev (1976) illustrates, the biggest problem with this and other joint demand models is that they do not reflect the fact that the demand for transportation is a derived demand based on the demand for goods in an economy. While the dependent variables include the volume to be shipped, the independent variables do not explain the demand for the commodities being trans- ported. The Micro-Disaggregate Approach One type of disaggregate model concerns the abandonment of rail branch lines. These types of studies are included here because the abandonment of a rail branch line reflects not only a change in the supply of freight transportation, but also in its demand. When a rail branch line is abandoned, it reflects the fact that the demand for the 22 operator's service has declined, resulting in revenues below cost, and thus, abandonment. These studies are considered as disaggregate models because they are generally commodity or region specific, focusing on a single branch line, a single commodity, or a specific geographic region. Examining two regions in Saskatchewan, Fost (1972) found that changes in local economic conditions resulted in changes in the grain handling and transportation system. This implies that a community's economic condition first affects the transportation system, rather than abandonment causing the economic decline of an area. This conclusion is similar to the aggregate models that suggest that changes in a region's economic activity affect the derived demand for transportation. In a study that compared nine counties that lost rail service to nine counties that retained service, Sloss, Humphrey, and Krutter (1975) found that the growth rate of income in the two areas was not signifi- cantly different. The short—run impacts of rail abandonment in 10 com- munities was studied by Allen (1975) who discovered that in only two of the 10 communities did rail abandonment cause significant declines in employment. Allen concluded that rail abandonment caused the largest shippers (feed and fertilizer) to shift to other modes of transport, increasing the costs to farmers for shipment. That the impact of rail line abandonment can best be seen in the costs imposed was demonstrated by Vollmers and Thompson (1980) who examined the impact of abandonment of four grain subsector branch lines in the Thumb of Michigan. Vollmers and Thompson suggested that rail abandonment forces shippers to find alternate shipping methods--probably truck--as well as a modification of destinations. It was found that abandoning the four branch lines serving the Thumb region of Michigan 23 would increase the 1976 grain marketing costs (before farmer exit) by $473,921, or 2.5 percent. This change in distribution costs modifies the competitive environment and encourages farmers to alter marketing patterns, causing smaller elevators to exit the market. While rail abandonment may not cause the economic decline in the first place, it will impose higher distribution costs and in the long-run force some shippers from the market; thus, further affecting the employment condi- tions in a region. Studies at the Regional Science Research Institute (Stevens, et a1” 1979 and 1980) have used input-output models to estimate the impact of new transportation facilities. When a new highway is proposed, the input-output method is suggested as a way to examine the direct, in- direct, and induced effects. The RSRI work concentrates on the use of input-output multipliers, particularly the income and employment multi- pliers. This differs from the approach taken in this study where only the output multipliers are used. Output multipliers represent the most straightforward use of input-output and are not subject to difficulties associated with the use of income and employment multipliers. The output multiplier simply indicates the degree of structural interdependence between individual sectors and the rest of the economy. Income and employment multipliers require more data, increasing the already sizable data problems that exist, and in the case of income multipliers, the assumptions of input-output make their use problematic. The assumption of an aggre- gate linear homogeneous consumption function is restrictive, particu- larly at the theoretical level (Richardson, 1972). Further, under- lying the income estimate is a questionable assumption that changes V0 fh All the Dr: out PUr tha ind 1011 hi duct R059! 24 in consumer spending are proportional to changes in income, both in terms of quantity of income spent and expenditure patterns (Jones, 1978). Using this simplified consumption function tends to overstate the in- come effects of changes in final demand. Thus, in using input-output, one must use care in the application of multipliers. The approach to transportation demand employed by researchers at the Massachusetts Institute of Technologyl/ begins by stating that the flow of cargo in a given market is simply the sum of individual ship- pers' decisions. These decisions are, in turn, conditioned by the specific inputs and outputs of each of the production processes in- volved. This disaggregate approach estimated the output levels of 'firms and using input-output analysis, estimated the inputs required to Ioroduce the given output. These input-output coefficients represent tflie dollar value of inputs purchased from a particular industry i to produce one dollar of output in industry j. When multiplied by the curtput of the industry, this coefficient gives the dollar value of purchases of inputs from industry i--and hence, the amount of goods ‘thaiz require transportation services. The objective is to explain the iruiividual shippers' decision on commodity k as the probability of .joiiitly selecting the frequency (f), mode (m), size (q), and location (1) (of purchases, given user location (j) and usage rate (u) required by the final demand for a good and the input requirements of the pro- ducti on process: Pk(f.m,q.i 1 uj) l/See: Chiang and Roberts, 1975; Chung and Roberts, 1976; Roberts, 1977; and Terziev, 1975. 25 Thus, determining the location of economic activity and the rate at which commodities are used in the production process is crucial to determining freight demand. Rail Traffic ForecastinggMethods and Rail Planning in Michigang/ With the passage of the 1981 Michigan Appropriations Act, a reduc- tion of 25 percent per year in financial support to currently subsi- dized rail lines was mandated, eliminating rail subsidies by 1986. This anticipated reduction in state support has produced a need to identify an essential core of transportation services. Consequently, a "rail rationalization" process was begun by the Michigan Department of Trans- portation which involves a two-tier procedure to define an essential core system of rail service and to rank all rail lines in terms of the requirements of that system. The Tier I analysis (adopted January 27, 1982 by the State Trans- portation Commission) screened all lines considered as candidates for financial assistance. The intent was to use readily available data, without exhaustive investigation, to produce a rank-ordering to indicate those line segments which warrant state assistance, and thoSe which do not. Tier I analysis included the review, evaluation, and placement of each segment into one of the following categories: 2/ - The following discussion is taken from the following Michigan Department of Transportation documents: Annual Update, 1978; Tier I Report, 1981; Tier II, Phase I Report, 1982; Tier II, Phase II Report, 1982; Northwest Rationalization, 1980. For full citations, see Bibliography. 26 (1) Lines which are viable or that should be included in the essential core system because of their contribution to program objectives. (2) Lines with questionable viability, but with the potential to be included in the essential core system. (3) Nonviable lines, not to be included in the essential core system. To do this, three sets of rail service indicators were used. The first set of indicators were called "service characteristics“ and in- cluded the number of shippers on a line, carloads, carloads per mile, carloads team tracked, and a rail dependency factor. The second set of indicators, "service effectiveness and cost characteristics," in- cluded revenue to operating cost trends, revenue to cost ratio, revenue to operating cost ratio, reliability costs, operating subsidies, and righttyfway costs. The third set of indicators were "potential and existing demand" and included the growth potential and production in agriculture,forestry,and extractive products, as well as an employment dependency factor and consideration of production centers served. The procedure was designed to objectively rank rail segments with respect to service criteria and economic potentials, minimize statistical bias, and illustrate the relative contribution to the state and regional rail systems of specific rail segments. Tier II is designed to assess benefits relative to the costs of state financial support for individual lines. A guiding concept is program opportunity costs, which addresses the question: What are the state's best investment opportunities to achieve rail service objec- tives, given limits on financial resources? th Se 27 The Tier II procedure employs a decision theory approach, as sum- marized in Figure 2.1. The approach is applied to those rail lines that were assigned to Category II in the Tier I process. After collect- ing traffic, cost, and revenue data for each rail segment under ques- tion, the Tier II process attempts to forecast potential rail demand and future requirements fur commodity transport services. A rail abandonment impact/cost method was adopted for use in the third step of the Tier II procedure. The impact/cost analysis compares the social and economic costs of discontinuing local rail service with the net capital and operating costs (subsidy) of continuation. For lines cur- rently being analyzed in Michigan, both of these sets of costs can be calculated quarterly over 10 years, or 40 calendar quarters. The method also estimates the personal income loss, state and federal tax losses, and unemployment and welfare compensation payments resulting from job losses reported by shipper surveys. The impacts are then discounted over time, which allows for the gradual reemploy- ment of labor. Reemployment rates are determined by statistics obtained from the Michigan Employment Security Commission (MESC). The limits on reemployment potential are determined by local area unemployment rates that are also obtained from the MESC. Finally, the Tier II process in- cludes the development of an adaptive decision process for minimizing the potential risks, or opportunity costs, associated with alternate service strategies. While this impact/cost approach is traditionally employed in individual abandonment hearings, decision makers viewing the rail system from a statewide perspective require a more comprehensive frame- work which considers spatial and commodity aspects. The impact/cost Economic Base Data Input- Output Data Cost Analysis Strategy Analysis 28 Figure 2.1 Rail Rationalization Plan Rail Investment Decision Analysis Model Tier II Flow Diagram —————> \/ \/ \/ Collect Primary and Secondary Source Data l System Demand Forecasting @— [K 1 Segment Cost/Benefit Analysis <+_______. l Rail System Investment Optimization 4g________ Publish Modal Plan Operational Data Waybill Data Sensitivity Data Risk Analysis Ci Re ti ti. Sh' rat Ella and 29 method is by design a procedure that examines each case on a line-by- line basis and offers little insight into how significant any given line is in the total rail system. A few shippers on a line with little total traffic could bias the results of the method. It appears that the procedure is biased toward the retention of rail service. Further, the interdependency between rail service and economic activity on a line or in the entire state is not addressed. The procedure suggested in the following chapters seeks to provide the link between rail demand on a particular line or region, the total freight transportation system, and economic activity throughout Michigan. Forecasting Rail Traffic and the Potential for Profitability of Currently subsidizedILines In both Tier I and Tier II of the Michigan Rail Rationalization Plan, it is necessary to forecast both potential commodity movements and traffic flows. In the three reports where forecasting has been used, the method has relied primarily on shipper-carrier interviews. Discussions with Michigan Northern management were used to develop fore- casts of potential traffic in the Grand Rapids to Mackinac Corridor Report. In both the Northwest and Ann Arbor Reports, traffic projec- tions fur 1985 were developed for the Michigan Department of Transporta- tion by A.T. Kearney, Inc. The projections were based on interviews of shippers in major commodity categories. A.T. Kearney used the 1979 rate structure, service quality, and traffic levels, as a base and evaluated the reported sensitivity of traffic to further rate changes and changes in the quality of service. In the Ann Arbor Report, these projections were adjusted by the Michigan Department of Transportation for changing commodity mixes. Cc ma es St: MEI est! dOmj 30 In all, the traffic projections used thus far in Michigan rail planning have been ad hoc in nature and not related to changes in economic activity in general. In fact, the transportation department recognizes this deficiency and is presently investigating alternate methods for estimating potential traffic changes. The foremost of the alternatives is a variation of the input-output method suggested in this study. Traffic projections in other states follow the same pat- tern as in Michigan. Shipper surveys are the predominant means of fore- casting rail traffic in Arizona (Transportation Research Circular, pa 2), while in Tennessee most work appears to be of the impact/cost method variety, with little forecasting ability. In Washington and California, however, preliminary work has been carried out in using an input-output model to aid in rail traffic fore- casting. In Washington, an input-output table has been used to pro- duce baseline data for use in forecasting. Thus far, the work in Washington appears limited to Standard Metropolitan Statistical Areas. Current use of the input-output model in Washington is focusing on modal choice and is based on shippers' surveys. In their West Coast Corridor Study in California, Cambridge Systematics, Inc. (1978), has made limited use of the 1967 input-output table of the United States to estimate commodity production in Pacific Region SMSAs. The California study relied entirely on the national model, rather than the two-region method,which will be discussed in Chapter IV. Further, Cambridge Systematics, Inc., relied on national census data for all of their estimates. Thus, not only are traffic projections needed to replace the pre- dominantly ad hoc procedures used, but traffic projections are crucial sys prC ing 599 Stu: miss lite that the tran duce: Chang OCCUr What ECOnol mOVemE SectOr 31 in analyzing the potential for profitability of currently subsidized lines. This study addresses the issue of whether future rail services on currently subsidized lines can be premised on the eventual profit- ability of private carriers. Existing systems with low traffic density can be justified on financial grounds only if future traffic increases are capable of overcoming existing losses. Conclusion The method suggested here addresses both the need to improve the systematic forecasting of rail traffic and the need to judge future profitability. It is possible to view the issue of transport forecast- ing as the interaction of aggregate economic variables with the dis- aggregate decisions by individual firms: The method suggested in this study combines both the level of analysis of the Canadian Transport Com- mission and Massachusetts Institute of Technology models. Abandonment literature suggests that it is the economic activity of an entire region that affects the demand for freight transportation services. As with the CTC model, changes in the economic climate of a region affects transportation demand. The use of a statewide input-output table pro- duces projections of economic activity in a region through estimated changes in the final demands of its commodities. Once these changes occur, individual shippers then face decisions on how, whether, and at what rate to ship their commodities. The initial assumption is that economic activity in a particular sector or region generates physical movements of the aggregate of products included in each of its input sectors in proportion to the change in that sector's output. an ti See The l‘ai the: 32 Once the aggregate change in economic activity has been estimated, the disaggregate structure of input—output models is amenable to elabor- ation of the transportation sector. Corresponding to each of the dollar flows present in the input-output model, originating and receiv- ing sectors can be identified and an indication of the type of product which is physically moving can be derived. Further, the use of com- modity flow data with the input-output model indicates over which rail lines products actually move. Consequently, the use of input-output as a forecasting method is particularly suited to the needs of transporta- tion planners. Input-output does not simply project total economic activity using macro variables such as GNP, income, or employment. Input-output is disaggregated by commodity groups--the same commodity groups whose future production will determine the demand for individual rail service. With this information, it would be possible to both estimate future traffic and evaluate the potential for profitability on currently subsidized lines--given estimates of future rail demand. The review of both rail forecasting literature and rail planning leads to the conclusion that there is a gap between the theoretical understanding of rail freight demand and an operational use of this understanding. A large body of work exists that deals with the factors associated with rail demand. However, this work is theoretical in nature and has not been employed in any meaningful sense in the actual estima- tion of rail traffic by planners. The method examined in this study seeks to bridge the gap between the theoretical and the operational. The goal is to make the link between the level of economic activity and rail demand suggested in the theoretical literature useful in the sense that rail planners have a method to address that link. In essence, the 33 purpose of this study is to operationalize the theoretical understand- ing of rail freight demand. pres regi thei syst mean‘ and l route and f repre: Opera: tains CHAPTER III THE MICHIGAN RAIL SYSTEM In this chapter, a description of the Michigan rail system is presented.l/ Particular attention is given to those rail segments and regions where rail abandonment and/or subsidization have occurred or their occurrence is likely. Additionally, a description of the rail system in Michigan's Upper Peninsula is presented. This chapter is meant to provide a perspective on the variety and extent of the issues and problems to be addressed by the Michigan rail planning procedure. The Michigan railroad system consists of approximately 6,028 route miles serviced by nine Class I carriers, 16 Class II carriers, and five switching and terminal railroads. Of this total, 5,654 miles represent actual first track mileage, while the remaining 374 miles are operated under trackage rights agreements. The Lower Peninsula con- tains 4,807 route miles and the Upper Peninsula contains 1,061 route miles. These figures are supplemented by 373 miles of active car ferry routes maintained by Michigan railroads. In any analysis of Michigan's rail network, the state's peninsular geography must be remarked upon. Due to this geographic configuration, three distinctive characteristics of Michigan railroading have emerged. lllhe material in this chapter is taken from the following Michigan Department of Transportation documents: Railroad Plan, 1980; Tier I Report, 1981; Rationalization Plan, 1981. Material on the Upper Peninsula is taken from: U.P. Rail Planning Commission, 1982. 34 Fi ar Pe bu b0 b0: POI int of rai fer eig ODe Pen maj. Teri n01c of p thro comm, Mlch- 0f fr ”Orth 35 First, Michigan, in reality, has two different railroad systems which are not physically linked by a direct rail connection. The Lower Peninsula system is oriented towards Chicago, Toledo, and Cincinnati gateways. Its traffic consists primarily of transportation equipment, bulk agricultural commodities, and chemical products which move out- bound; and coal, primary metal goods, and food products which come in- bound. The Upper Peninsula system serves mainly as an outbound rail route for metallic ores and pulpwood. Most of its traffic runs southerly into Wisconsin or to lake ports. Second, since Michigan's two peninsulas have a total of 3,126 miles of Great Lake shoreline, which drastically restricts direct overland rail connections with other states and Canada, an extensive rail car ferry service has evolved in Michigan. Currently, five carriers operate eight different car ferry routes from Michigan ports. The bulk of these operations involve Lake Michigan service between Wisconsin and the Lower Peninsula. While the Great Lakes rail car ferry routes have played a major role in the development of Michigan's railroad network, the ferries' extensive capital needs and the emergence of new railroad tech- nology and equipment now threaten the existence of this unique element of Michigan railroading. Third, the state's peninsular geography inhibits the growth of through traffic and the resultant heavy density freight routes which are common in most Midwestern states. Therefore, a significant portion of Michigan's railroad activities involve the origination or termination of freight carloadings, i.e., a terminal or switching function. As the Northeastern rail crisis has clearly indicated, a railroad's financial pe ha an ca aci tot cat ren yea- mil, Whlc aCco aCcO term ture ran 36 viability is frequently endangered when this particular function becomes dominant. As mentioned earlier, as of September 1, 1980, the rail system consisted of 838 land route miles operated with state or federal rail service assistance, 204 land route miles which were subject to abandon- ment proceedings, 589 land route miles which were considered by carriers to be candidates for abandonment, and approximately 4,397 land route miles operated by solvent carriers and not potentially subject to or pending abandonment. Figure 3.1 illustrates those rail lines which have been abandoned since 1976; Figure 3.2, the lines which were sub- sidized in 1981; Figure 3.3,the pending abandonments as of late 1981; and Figure 3u4, the lines which have been identified by the solvent carriers as under study for future abandonment. Service provided, which is indicative of the level of economic activity dependent on the availability of rail, can be summarized by total carloads originating or terminating on lines in each respective category. The 838 miles to be operated with assistance under the cur- rent program originate or terminate 28,400 carloads of rail freight per year. Approximately 39,000 carloads originate or terminate on the 204 miles currently subject to abandonment proceedings, and the 589 miles which are candidates for a future attempt by the carrier to abandon account for about 11,300 carloads each year. The balance of the system accounts for approximately 1.5 million carloads of originating and terminating traffic. In an effort to stabilize the policy and financial support struc- ture for the continuation of essential core services on the state's rail system, a "rail rationalization" process has been implemented by 37 2. 3. 4. 6. 7. 8. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23m 24. 25. 26. 27. e) .1 1 PF“ 3 1 . d / Ida-Lenavee Jct. (PC) Leaf-State Line (DTSI) Ackerson Lake-State Line (PC) Osseo-Hillsdale (PC) Haires-Jonesville (PC) Nottawa-Wasepi (PC) Mendon-Vicksburg (PC) Richland-Doster "T" (PC) Otsego-Byron Center (PC Coopersville-Grand Haven (GTW) North Horn Spur (C50) GTW Car Ferries (GTW) Ludington-Milwaukee Carferry (0&0) Lansing-Owosso (PC) Oxford-Millington (PC) Coleman-Mt. Pleasant (C60) Edmore-Remus (C&0) Williamsburg-Elk Rapids (C60) Race-Race Jct. (Soc) Rapid River-Eben (Soo) Hunising-Marquette Lawson-Little Lake (L851) Champion—Republic (Milw) Ishpeming-Martina Landing (CNW) Scott Lake-State Line (CNW) Nestoria-Bergland (Soo) Dollar Bay-Lake Linden Hancock-Calumet: (500) State Line-Marenisco (CNW) Figure 3.1 Michigan Abandonments Since 4/1/76 38 1. 2. 3. 4. 6. 7. Constock Park-Mackinaw City Toledo-Frankfort Prankfort-Kewaunee and'Manitowoc Straits Carferry Grand Rapids-Vermontville Noctawa-Sturgis and wasepi-Mendon Hillsdale Area Lenawee Area rascals-Saginaw Bay Figure 3.2 State Subsidized Lines 1981 39 Manistee-Bay View (C60) an. a... mu- Ludington-Manitowoc (CEO) Montague-Hart (C&O) Fuller—Kinney GRGI Branch (Conrail) Hartfort-South Haven (C80) Niles-Benton Harbor (Conrail) Vickesburg-Austin Lake Battle Creek (Conrail) Eaton Rapids-rives Jct. (Conrail) Jackson-Three Rivers (Conrail) Ottawa Lake-Clinton (Conrail) Monroe Branch (Conrail) Detroit Area (Conrail - four segments) Scott Lake-State Line (CNW) Watersmeet—State Line (CNW) Baraga-Dollar Bay (Soo) Figure 3.3 Pending Abandonments as of 11/24/81 40 auteur.“— I O O O O O Portland-Ionia (C80) Ashley-Greenville (GTW) Inlay City-Caseville (GTW) St. Ignace-Shingleton (Soo) Hunising (L881) Bessemer-State Line (Soo) Figure 3.4 Lines Identified by Carriers as Under Study 41 the Michigan Department of Transportation (MDOT) to identify those lines warranting a long-term state commitment. Rail rationalization was placed in the 1981 Michigan Appropriations Act (MPA 32) as a require- ment prior to the expenditure of funds after October 1, 1981. Another section of the Appropriations Act outlined a financial reduction of 25 percent per year on the currently subsidized lines so that at the end of a five-year period, they would no longer be funded. These two legisla- tive requirements are interpreted as requiring the Michigan Department of Transportation to reconsider the elements of the current rail preser- vation program and reduce its cost, while maintaining essential services. The total of state funds available under the legislated phase-out for rail and water operating assistance, by year, is shown in Table 3.1. This phase-out of the state appropriation for operating assistance ap- plies to state-assisted rail systems operating on the date the rail rationalization plan for the entire state is approved. Table 3.1 Subsidy Phase-Out Schedule Fiscal Year Operating Assistance Reduction from 81/82 1981-82 $11,156,100 1982—83 8.367.075 25% (5 2,789,025) 1983-84 5,578,050 50% ( 5,578,050) 1984-85 2,789,025 75% ( 8,367,075) 1985-86 0 100% ( 11,156,100) , 42 The systems indicatedg/ are currently the subject of competitive bid proposals solicited by the MDOT to ensure the continuation of es- sential services over the next five years at minimal cost to the state. While the state appropriations available for continuation assistance on these systems is to decrease by 25 percent per year, there are potential economies and sources of funds yet to be exploited. Beyond economies in operation which may be possible, cost savings can be achieved through system restructure, if necessary. One potential source of funds is local financial participation by shippers and local units of government. The Upper Peninsula Rail System Rail service in Michigan's Upper Peninsula is used principally by the mining and forestry industries, including related pulp and paper in- dustries (this discussion from the U.P. Rail Planning Commission, 1982). Service is characterized as light density with dispersed shipping points and low volume of shipments. Furthermore, most U.P. traffic moves north to south, rather than in the east-west direction of the railroads. Consequently, transportation costs are high because freight is often shipped over circuitous routes. This problem particularly affects freight traveling over the 500 Line from the Munising, Marquette, and Keweenaw Peninsula areas. In order to get freight from Marquette, south, the 500 Line ships east to Trout Lake, then west through Gladstone before heading south to Wisconsin. The following is a brief discussion of the operations of each railroad in the Upper Peninsula and their traffic characteristics. g/Michigan Northern, Kent-Barry-Eaton Connecting Railway, Tuscola and Saginaw Bay Railway, Hillsdale County Railroad, Lenawee County Rail- road, Ann Arbor Railroad System, Straits of Mackinaw Car Ferry Service, and two ConRail segments. 43 500 Line: The 500 Line operates 514 miles of right-of-way, with a main line running from Sault Ste. Marie to the Wisconsin line via Rapid River and Gladstone. Other lines provide service to points along the Lake Superior shoreline, including the Keweenaw Peninsula; the Chief Wawatam Ferry at St. Ignace; the White Pine Mine in Ontonagon County; and the Ironwood/Bessemer area. These lines provide the only link with the International Bridge at Sault Ste. Marie, and the only direct link with the Lower Peninsula via the Chief Wawatam Ferry at St. Ignace. Chicago and Northwestern: The C&NW operates primarily north and south in the U.P. and currently maintains 282 miles of right-of-way. This railroad operates the only direct north-south route in the U.P. between Palmer and Escanaba as its main line. It also provides service along the Wisconsin border between Powers and Marenisco and owns and operates the principal ore dock in the U.P. Escanaba and Lake Superior: The E&LS, in addition to its original line from Wells to Channing, has acquired the bulk of the property abandoned by the Milwaukee Road in the U.P. This property includes lines from Iron Mountain to Channing, Channing to Ontonagon, and Channing to Republic. Additionally, the railroad operates on the old Milwaukee Road Line from Iron Mountain to Green Bay, Wisconsin, where it links with the Milwaukee Road. Lake Superior and Ishpeming: The LS&I primarily transports iron ore pellets from mines in western Marquette County to the Presque Isle dock in Marquette. This dock serves as a secondary alternative ore handling facility to the C&NW dock at Escanaba. Currently, 67 miles of track are maintained. This includes an island operation between 44 Munising and Munising Junction, which links the Kimberly Clark paper mill in Munising with the 500 Line at Munising Junction. Milwaukee Road: While abandoning nearly all of its operations in the U.P., the Milwaukee Road still retains approximately a two-mile spur serving the City of Menominee. Due to the peninsular geography of the U.P., most rail traffic either originates or terminates in the area. Through traffic (freight transported through the U.P. with origin and termination points outside the U.P.) is primarily transported via the Soo Line from Sault Ste. Marie south through Wisconsin. This traffic consists primarily of lumber and wood products being transported from Canada to various points in the United States. Highest densities of rail traffic occur on the C&NW Line transporting iron ore from the Marquette range to the Escanaba ore docks, the LS&I Lines transporting are between the Marquette range and the Presque Isle dock, and the 500 Line main line between Sault Ste. Marie and Wisconsin. Principal outbound commodities shipped by rail are: lumber and wood products; metallic concentrates; pulp, paper, and allied products. Inbound commodities consist of bentonite clay; various chemicals used by the mining, pulp, and paper industries; coal; and propane gas. Some other nonmetallic minerals (such as limestone) are shipped as well. For the most part, these commodities are high bulk and low cost per unit volume; commodities of this type can usually be shipped more profitably by rail than by truck. As in the case in many areas, the U.P. rail system was built for a greater demand than currently exists. Thus, abandonments have occurred with most of the excess lines being eliminated. In discussing the 45 abandonment process, the U.P. Rail Planning Commission (1982) is con- cerned with the piecemeal approach where only the line in question is taken into account. What is lacking, according to the Commission, is an examination of how a particular line might fit into a more efficient system. As noted in their draft report, the Commission states: Conceivably, some lines which when viewed in isolation appear unprofitable could be utilized in a restructured system that would reduce excessive travel distances and increase traffic densities over the entire system, both now and in the future. Continuation of piecemeal abandonments may well mean elimination of lines that could play a key role in the future growth and development of the Upper Peninsula. It is the concern with fragmented abandonment and subsidy decisions that this study addresses. As in the case in Michigan's Upper Peninsula, a rail system should be seen from a statewide, or region-wide, perspec- tive in order to arrive at decisions that provide the whole region, and not just pieces of it, with a viable transportation system. Not only must the system be viewed from this broader scope, but economic activity and the future demand for freight transportation must be considered in a comprehensive context as well. In the next chapter, a method is examined that should provide transportation planners with the wider perspective necessary to make system-wide decisions. CHAPTER IV METHOD AND ECONOMIC FRAMEWORK Framework of Study The conceptual framework of this study focuses on the structural relationships in the transportation system; the location of freight transportation facilities and services, as well as the location and level of economic activity in the state. To be measured are the results of economic changes and public policy decisions which alter the struc- ture. Transportation facilities and economic activity are highly inter- dependent; a structural change in one sector of the economy or trans- portation system will induce changes in other sectors. Plant location decisions are based, in part, on existing and potential transportation structures, while transportation location decisions must take into consideration existing or potential demand. Investment encourages both the development of transportation systems and economic activity, while disinvestment in either element will hinder the viability of the other (Pratt and Thompson, 1979). Given an interrelationship between the location of economic activity and the transportation system, the con- sequences of investment-disinvestment decisions in freight transporta- tion policy depend on the industrial structure of the state. Especially important is a knowledge of the specific inputs consumed by each in- dustry. Therefore, determining the volume of commodities that will 46 47 flow in a given market is the starting point for any quantitative analysis of transportation policy (Chiang and Roberts, 1976). The method used here is also a structural one. On the one hand, an input-output model represents the economy's structure and predicts the annual usage rate of commodities by industry. Commodity flow data illustrates how these commodities move across the state by rail. In an input-output sense, the model identifies an industry's suppliers and its clients and the input-output relationship to other sectors. On the other hand, since the demand for transportation is a derived demand, it is dependent on the state of the rest of the economy. Thus, structural variables describing economic activity are used to explain the demand for freight transportation. Such a model allows for the investigation of changes in both economic activity and transportation facilities; it is possible to simulate the impact on rail freight demand or variations in the level of the economy (changes in final demand). The Structure of the Input-Output Model The model employed in this study is made up of five sectors: (1) producers, who combine intermediate goods and inputs to produce final goods (or inputs to other industries); (2) consumers, who demand final goods (as well as other industries that demand intermediate goods); (3) governments, which also demand final goods; (4) the trading sector, which is a demander of final goods for export and supplies of inputs for import between the state's economy and outside economies; and (5) carriers, who transport both final and intermediate goods spatially between supply and demand points. In using the input-output method developed by Leontief and ex- panded by others (Dorfman, Samuelson, and Solow, 1958), the concept is 48 to use a system of accounts that represent the transactions among the sectors of the economy. This determines the interrelationships that exist between sectors of the Michigan economy and the relationship between them and the economic activity outside of the state (Rest of Nation, RON). By using this accounting method, it is possible to demon- strate how outputs in dollars from each producing sector are distributed among other sectors in Michigan and RON. In terms of transportation requirements, the input-output table also shows how (and from whom) each sector procures its inputs (both in Michigan and RON). Input-output models are consistent in that the outputs of each sector must conform with the outputs of all other sectors from whom it buys inputs and sells outputs, and with the total output of the economy. It is then possible to trace the interdependence in the economy and the required flow of goods needed in the input-output process. Following is a description of the input-output matricies and their analytical use.l/ The 20-sector, input-output model of Michigan used in this study is reproduced in Appendix 1. Also in Appendix 1 is an example of the use of the model. In Appendix 2, a 44-sector version of the input-output model is included. To use an input-output table, the first step is the construction of the transactions or interindustry matrix (Appendix 1, Table l), the structure of which is outlined in Figure 4.1. The three major sectors of this table are the producing sector (X), primary inputs (W), and final demands (Y). The numerical entries are represented as variables, vectors, and matricies, in constant dollar terms: l/This discussion follows the layout example from: Pratt, Smith, and Connor, 1976. The mathematical formulation is from Jones, 1978; and Adiarte and Venegas, l980. 49 mpnmh covauemcmeh Pocopoomcmpcm esp mo wczuuzepm Asks. .axez nee aces: "mucsomv _.e ec=m_e ¥.> H.x e E 3 . 3 .ax x_s fl_x O O o O O O N —- z 0 o o 0 o o O 0 O O o O o O O O o N F panaao mcouumm _eece accede Page; mcouumm mewszmcou OZ r- mpaacH _maoe mcouumm usacH Agnew»; mcouumm mementos; where: 3x2 Xij ik mi 50 the number of producing sectors; the number of final demand sectors; the number of primary input sectors; h the purchase of the output of the it producing sector by the jth consuming sector for the purpose of produc- ing the output of it“ sector, i = l, 2, N; j =1, 2,... N; the final demand of the sector k for the output from the ith producing sector or primary input class, i = 1, 2,... N+M; l:==1,.2,... K; th the purchase of the m sector of primary input by the jth consuming sector for the purpose of producing the h output of the jt sector, m = 1,2, ...M; j = 1,2,... M; th total output of the i sector, i = l, 2,... N; h total input of the jt sector, j = 1,.2,... N; total final demand for the output of the it“ sector, i = l,.2,... N+M; total final demand of the sector k, k = 1,2,... K; the purchase of total primary inputs by the jth consum- ing sector for the purpose of producing the output of the corresponding jth producing sector, j =‘l,2,... N. The interrelationships in the economy expressed in the transactions matrix are as follows: (1) x1»; x..+ v. J1.” 1 51 The sectors of the economy are arrayed horizontally and vertically, in identical order. The purchasing sector is listed horizontally and the selling sector vertically. Thus, the distribution of a sector's product sales is read across its "row" in the matrix and the distribu- tion of a sector's input purchases is read down its “column." The intersection of any column and row represents gross purchases from, or gross sales to the corresponding sector. The transactions matrix quanti- fies the intersectoral flows in thousands of dollars. The Michigan table is divided into the processing sectors, 1-20, and the payments and final demand sectors, 21-23. Row/Column 24 represents the total sales and total purchases in the economy. Second, it is necessary to obtain the direct requirements or tech- nical coefficients matrix (Appendix 1, Table 2). Derivation of these technical coefficients assumes a linear relationship between purchases of an endogenous sector and the level of output of that sector: (2) aij = xij/Xj = amount of industry i's output neces- sary to produce one unit of industry j's output; where: aij = technical coefficient; Xij = value of sales from industry i to industry j; Xj = total output of industry j: Xj is the same as Xi in equation (1), where i=j, corresponding row and column totals in the processing sector are equal. Technical coefficients are computed for each industry in the processing sector, yielding the following matrix: a” 612 ............ a]: 321 322 ............ 62" A = an] anz ............ ann Where there are n sectors in the processing sectors, (A) represents the complete matrix of technical coefficients. The technical coef- ficients are found by dividing each column entry for the processing sectors by the corresponding column total for the sector. Each element in the coefficient matrix indicates the dollars of inputs required from each selling sector (horizontal) in order to produce one additional dollar of output in the purchasing sector (vertical). Third, the Leontief matrix (Appendix 1, Table 3) is found by sub- tracting the coefficients matrix for the processing sectors (A) from an identity matrix of the order (I-A). Each element of the coefficients matrix at the intersection of an identical horizontal and vertical sector is subtracted from one. The other elements are subtracted from zero. The coefficients matrix identifies the direct purchases that a sector must make to deliver one dollar of output to final demand. Yet, the sales of one dollar of output to final demand by one sector, require the other sectors that supply it with inputs to increase their output by more than the direct amount. Inverting the Leontief matrix produces the inverse matrix (Appendix 1, Table 4), commonly called the interdepend- ency coefficient matrix. These coefficients yield a table of direct, indirect, and induced requirements per dollar of final demand. The coefficients are obtained from the matrix: A“ A], ............ Ami A21 A22 ............ A2n (I-A)" = An] An2 ............ Ann Determining the interindustry transactions or production over all sec- tors required to sustain a given level of final demand represents the input-output problem, solved as: (3) x = (I-A)" Y where: x = total output vector of economy; y = final demand vector facing economy; (I-A)’1 = matrix of interdependency coefficients. The sum of each sector in the inverse matrix is an output multi- plier. These column sums, or output multipliers, show the direct, in- direct, and when the household sector is endogenous, the induced economic activity that will be generated by the economic system as a whole for each sector to meet an increase of one dollar in final demand. Assumptions of Input-Output (l) Constant Production Coefficients for each sector: This as- sumption implies that the input structure of a sector remains fixed; the physical inputs for etonomic sectors change in the same proportion as output. From this assumption, four other relationships follow: (a) There is no technological change which alters factor- factor or factor-product relationships. Where a relatively stable 54 economy exists, the constant technology assumption is not a large problem. When an economy undergoes large economic changes, the input-output model can be modified, the direct requirements may be adjusted converting the model from "static" to "compara- tive static“ (Diamond and Chappelle, 1981). (b) Sectors cannot substitute inputs from one sector for inputs from another. This fixed rate of substitution assumption implies that relative use levels of resources do not change over time. (c) There are no scale or size economies of production. (d) The relative prices of all goods and services remain constant. (2) Homogeneity of Sectors: Each sector produces only one pro- duct, with one process. It is obvious that firms in sectors have diverse input and output structures. Yet, it is not feasible to at- tempt a model that would incorporate all of these differences. This assumption is essentially an aggregation problem, endemic in economic analysis. There are no set rules for aggregating firms into individual sectors. The aggregation of any input-output model depends mostly on the region to be studied and the uses of the input-output model itself. The homogeneity assumption is most harmful when attempting to extend the input-output model to uses outside of its accounting framework. In building an input-output table, most firms are grouped in sectors based on a similarity of product type and production process. This normally fullows the classification of firms by Standard Industrial Classification codes. The best basis for aggregation is similarity of input structure (Chenery and Clark, 1959). 55 (3) Additivity and Divisibility: The outputs of each sector are assumed to be additive in that the total effect of several production activities is the sum of their separate effects (Pratt, Smith, and Conner, 1976). The output of a combination of sectors is the sum of their inputs, regardless of the combination of outputs. The direct requirements coefficients represent the only interaction between pro- ductive activities. It is also assumed that inputs are available to produce the outputs required to meet final demands. As is the case in many forms of model building, any sector is as- sumed to be divisible. Sectors are assumed able, within the model, to produce a fraction of a unit of production. However, the aggregation of total output in most input-output models is large enough to minimize any departure from reality imposed by this assumption (Pratt, Smith, and Conner, 1976). These three assumptions do not negate the usefulness of input- output analysis. However, the researcher must be aware of the limita- tions implied when interpreting the results. Construction of the 20-Sector Michigan Input-Output Model The structure of the Michigan input-output table is derived from the work of Hwang and Maki (1979). The procedure that Hwang and Maki applied to Minnesota was used to build the Michigan model by Adiarte and Venegas (1980) for the Michigan Energy Administration. (The fol- lowing discussion on the building of the two-region, input-output table for Michigan is taken from these two reports.) The procedure begins by estimating gross output for each industry sector to serve as controls for the rows and columns of the trans- actions table. Data on value added, personal income, taxes, etc., are ma se va pi 111 tre acc rel cie wit lev ind por obt tha com fic the fOr 56 used to fill in several cells of the table. Next, the gross outputs are distributed among all sectors using a secondary data approach. The table is then rebalanced with respect to row and column totals and the validity of the model is checked against other estimates of gross state product, personal income, employment, etc. After the row and column totals of the transactions table are esti- mated, the cells representing intersectoral flows are estimated using a secondary data approach. The secondary data on industrial outputs, value added, demand components of gross state product, etc., were com- piled from the U.S. Department of Commerce's Ceg§g§_reports and statis- tical publications of various state agencies. This approach builds a transactions table based on national coefficients and makes adjustments according to the existing industry mix and other differences of Michigan relative to the national average. In order to alter national coeffi- cients for Michigan conditions, a commodity balance approach is used (Adiarte and Venegas). This approach compares industry gross output with estimated purchases using technical coefficients at the national level, and adjusts these proportionally whenever total purchases from an industry exceed the gross output of that industry. Schaffer and Chu re- port that comparisons with survey results show that the coefficients obtained in this manner are larger, but not statistically different, than those obtained using survey data. Finally, Adiarte and Venegas compared control totals in Michigan based on primary data, with the coef- ficients derived for Michigan from the national model and concluded that the derived coefficients lend confidence to the adequacy of the model for impact analysis. 57 Two-Region Model In order to produce a two-region model, the national input-output table was split into two regions, Michigan and the Rest-of—Nation (RON). The industries, final demands, and primary inputs outside of Michigan were aggregated across states, but treated as distinct sectors trading among themselves with similar sectors in Michigan. Two transactions tables were produced, Michigan and RON, and developed simultaneously from estimates of four commodity flows: (1) sales within Michigan; (2) exports from Michigan; (3) imports into Michigan; and (4) sales of outside industries among themselves. Two balanced input-output tables are created when imports and exports at the industry level are combined into the import and export sectors. In doing this, three additional assumptions were necessary: (1) the input structure at the national level prevails at the state level; (2) surplus output in Michigan becomes an export to RON, while deficit output in Michigan becomes an import from RON; and (3) cross-hauling of a similar product between the regions is indicated only as a net flow (Adiarte and Venegas). The two-region procedure begins with the national transactions table, in the form of Figure 4.1. The technical and interdependence coefficients are found as described earlier. Thus, as Hwang and Maki note, the following relationships are true by the previous definitions: k vi = E Yik where 1 = 1, 2,... N+M; k-l m W. = E W . where j = 1,2,.”. N. ti fi to SUI 58 And since it is assumed that the total output of a sector is equal to the total input of the corresponding sector, the following relationships exist: "M: X + _< X. = X. when i=j when Xi = J When the national table is disaggregated into two regions, the result is four quadrants for each sector, primary inputs, and final demand components, as shown in Figure 4.2. The quadrants are product flows defined for: (1) first quadrant of intraregional intersector of transactions for Michigan; (2) second quadrant of interregional inter- sector of transactions from Michigan to RON; (3) third quadrant of interregional intersector of transactions for RON; and (4) fourth quadrant of the intraregional intersectoriyftransactions for RON. The interregional flows assume the surplus of the region to be an export to the other, while a deficit is an import from the other region. National coefficients are used initially for both Michigan and RON, in which linear adjustments are made for estimated surpluses or deficits. The method for doing this is similar to the commodity balance approach, except for the simultaneous adjustments in the coefficients of both tables. In Figure 4.2, a single superscript indicates a particular region, and a double superscript indicates the source of flow by the first letter and the destination of flow by the second letter. The first step in compiling the two-region transactions tables is to estimate the total consumption of a sector's output by all the con- suming sectors and final demand sectors in a region. This is derived by: .3...» mucoEoo —o:_u — cc.E:mcou sex zo~umm .IR sq: zo~umz 59 Adam. ._xcz eee wees: "muczomv _muoz u:au=¢-u=n=~ :oFmomuozh mg» we opamh cowaummcmgh pagouuomcmacm pocowmwcamucm new no wesuoaeum .e. zoned: .a. zone”: ~.e oc=m_n am» nmx am» npx u=a=_ pouch Aueaeuesc suev Aueugvozc tiny : «Louuom he he H unac— caz no: — xsaewga .a. z . z . c Augustus: zuev Augustus: guev Augustus: ugnv Aucmgcmau ucmv . ~ 0 I u _ x.» a. a? a, . meouuom u ex we aux as» sex . ae_o=eoca a N p Announce: ucmv Augustus: umpv 1 «nouumm we we . «sac— aa: an: F unsaved 3Q.- z . z . a aueucuoac ucwv Augustus: oeuv “accrues: amp. Aucacvoac ampv . ~ . u ._ x. a. x. a. . neoaoem m ex ea» sax an» sax . me_o=ooca a N p x . . . . . N P z . . . . . . N x . . . . . N _ z . . . . . . ~ _ unnuao meouumm acouuom . pouch mucosa: —~=_m me.e=mcoo mucosa: .a:_d ocpeamcou sect the same then corre total regio techn ages. the o 4.2.? 1mDOr SECOn While EStin Of tr exDOr qUadr *p n p k p X. = Z a..X. + Z a. Y 1 i=1 13 j k=l ik k * n k 9 - 9 X. — E a. X. -+ Z a ‘Yq 1 3‘1 13 J k=1 k k when i= 1,2, N+M. Secondly, the quantity of the surplus or deficit of a producing sector of the region is found. The surplus or deficit is the amount of the difference between the total output and the total consumption of the same producing sector. If a surplus exists for a sector in region p, then this amount is allocated proportionally to the subsectors of the corresponding consuming sector of region q. If the difference between total output and total consumption is zero, then there is no inter- regional flow of the output for that sector. This procedure produces technical coefficients of each region different from the national aver- ages,and more appropriate to a particular region. The allocation of the surpluses or deficits among industries in the two regions fills out each cell of the expanded matrix of Figure 4.2. Balanced regional tables are thus produced by constructing the import and export sectors. For the Michigan table, the sum of the second quadrant from each producing sector row represents total exports, while the sum of the third quadrant for each producing sector column estimates imports. The same is true for RON, except that the cohnmisum of the second quadrant represents imports, while the row sums are exports. A balanced table for Michigan and RON is organized with the first quadrant representing the interindustry transactions in Michigan and the fourth representing the interindustry transactions in the RON. 61 The consolidation of the second and third quadrants generates imports and exports by industry in each region. Finally, Adiarte and Venegas ran the two-region model to generate the single Michigan input-output tables. The 44-sector input-output table for 1976 in Appendix 2 represents the table produced at Minnesota for Michigan. The 20-sector table is an aggregation of the 44-sector model. Data, Updating, and Aggregation of the Michigan Input-Output Model The procedure described above began with the 367 sector, 1967 input-output model of the national economy, produced by the Bureau of Economic Analysis (BEA), U.S. Department of Commerce. The national model was updated by the BEA to 1970. This 367-sector model was first aggregated into 53 sectors and data for Michigan was assembled from U.S. census sources as well as Michigan specific data from the Michigan Statistical Abstracts. Michigan data was that of gross outputs for each of the 53 sectors and total expenditures of each final demand sector. The Minnesota Energy Agency repeated the estimation procedures for a 1972 model. The updating was done by recalculating gross outputs and final demands and balancing the transactions table. Estimates of value added, gross state product, and other aggregate variables were verified during this updating process. The model was reaggregated to 47 sectors and combined with 1972 estimates of final demands and the 1972 update of the national input-output table. The estimated gross outputs were used to rerun the two-region model. The same procedure was used to update the input-output table to 1976. The sectors were further reduced to 44 industry groups, by combining pipelines with gas 62 utilities, and commercial service sectors into selected services. Table 4.1 shows 'the 1976, 44-sector aggregation; this aggregation is reflected in the 44-sector model of Appendix 2. When linked to transportation data, the 1976 model was aggregated to 20 sectors, as shown in Table 4.2. This was done to better match the mix of commodity groups moving over Michigan's rail system to the input- output sectors. Additionally, this aggregation was performed to deal with those commodities in the 44-sector model that have small absolute amounts moving over the rail system. Similar commodity groups with low tonnage were aggregated, producing the 20-sector model. Output Multipliers and Projection of Total Output The use of an input-output model allows for the calculation of output multipliers; the main benefit being that they are disaggregated. Aggregate multipliers, used in macroeconomic analysis, do not distinguish between the sectors in which expenditure changes begin. Input-output multipliers, on the other hand, are valuable in that they account for the fact that the total impact of changes in the economy on output will vary according to which sectors experience the initial expenditure change (Richardson, 1972). Output multipliers represent total requirements per unit of final output. These multipliers serve as an indication of the degree of structural interdependence between each sector and the rest of the economy. The higher the multiplier, the greater the interdependence of the sector with the rest of the economy. Output multipliers measure the total value of industry requirements per unit of final demand for each industry (Richardson, 1972). As Leontief (1941) points out, the 63 Table 4.1 Sectors Defined by SIC, 1976 I/O Table for Michigan Industry Number 1972 Sector Industry 1967 I/O SIC Code 1 Livestock and Products 1 02 2 Other Agriculture, Forestry 2, 3, 4 01, O7, 08 Services 3 Metals and Minerals 5, 6, 9, 10 10, 11, 12, 14 4 Crude Petroleum, Natural Gas 8 1311, 1321 5 Construction 11, 12 part 138, 15, 16, 17 6 Meat Products 14.01 201 7 Dairy Products 14.02 - 14.06 202 8 Preserved Food 14.07 - 14.13 203 9 Grains, Bakery, Sugar, Fats 14.14-14.20, 204-207, 209 14.24-14.32 lO Beverages 14.21-14.23 208 ll Textile, Apparel 16-19 22, 23 12 Lumber, Wood 20, 21 24 13 Furniture 22, 23 25 14 Paper and Allied Products 24, 25 26 15 Printing and Publishing 26 27 16 Chemicals and Allied 27, 29.02, 29.03, 28 excluding 30.00 282, 283 17 Plastics and Synthetic 28 282 18 Drugs 29.01 283 19 Petroleum Refinery 31 29 20 Rubber and Leather Products 32-34 30, 31 21 Stone, Clay, Glass ' 35, 36 32 22 Foundries and Metal Products 37.02-37.04, 38 332-336, 339 23 Blast Furnace, Basic Steel 37.01 331 24 Metal Containers, Heating, 39.01, 40 3411, 343, 344. Plumbing, Fabricated Metals 25 Screw Machine Products 41.01 345 26 Metal Stampings 41.02 346 27 Other Fabricated Metal 39.02, 42 342, 347-349 28 Engines and Turbines 43 3511, 3519 29 Farm and Construction 44-46 3523-353 Machinery 30 Metal Working Machinery 47 354 31 Industrial and Service 48-50, 52 355, 356, 358, Machinery 359 . 32 Office and Computing 51 357 Machines 33 Electrical Equipment 53-58 36 34 Trucks,Trailers, Motor 59 371 Vehicles 64 Table 4.1 (cont.) Industry Number 1972 Sector Industry 1967 I/O SIC Code 35 Aircraft and Parts, Other 60, 61 372-379 Trans. 36 Miscellaneous Manufacturing 62-64 38, 39 37 Transportation, Communica- 65-67 40-48 tion 38 Electric Utilities 68.01 491, part 493 39 Gas Utilities 68.02 492, part 493 40 Water and Sanitary Services 68.03 494-497, part 493 41 Wholesale and Retail Trade 69 50, 52-59 42 Finance, Insurance, Real 70-71 60-67 Estate 43 Services 72-77 70-89 44 Government Enterprises 78-79 - 65 Table 4.2 1976, 20-Sector Input-Output Classification Sector Number from 44- Sector Industry Sector I/O 1 Livestock and Products 1 2 Other Agricultural Products 2 3 Mining 3, 4 4 Construction 5 5 Food and Kindred Products 6-10 6 Lumber, Furniture, Paper, Printing 12-15 7 Chemicals, Drugs, Plastics 16-18 8 Petroleum Refinery 19 9 Rubber, Leather, Stone, Glass, Clay 20, 21 10 Primary and Fabricated Metals 22-27 11 Machinery, Except Electrical 28-32 12 Electrical Equipment 33 13 Motor Vehicles and Parts 34 14 Aircraft and Other Transportation 35 Equipment 15 Transportation and Communication 37 16 Utilities 38-40 17 Wholesale, Retail, Miscellaneous Manu- ll, 36, 41 facturing, Including Textiles 18 Financial, Insurance, Real Estate 42 19 Selected Services 43 20 Government Enterprises 44 66 concept of final demand has as its counterpoint the concept of inter- mediate demand or output, while total output embraces both. The final demand for a commodity represents that part of total output which is treated as an independent variable. The derived or intermediate demand is considered to be a dependent variable. The purpose of using an input-output model for forecasting rail traffic rests on the concept of the output multiplier. To forecast rail traffic, it is necessary to forecast changes in the economy, given changes in the final demand sector. The disaggregated nature of output multipliers is particularly well-suited to transportation planning because the demand for specific rail lines or segments is often com- modity dependent. A method is needed to disaggregate the effects of economic changes (final demands) on a commodity-by-commodity basis in order to estimate the future demand for rail services. The effect of the change on individual sectors of the economy, given a change in the final demand in one sector, can be evaluated. The total output multipliers produced by the inverse matrix and shown in Table 4.3 provide estimates of the total economic effects of a change in each sector. It is also possible to examine how the effect of a change in the final demand in one sector affects individual sectors in the economy. This is done by multiplying the sectors column of the in- verse matrix by the change in final demand, as is done for the livestock and products sector in Table 4.4. Table 4.4 shows the impact of a $1,687,000 increase in final demand in the livestock and products sector (equivalent to the 1976-1980 change that will be discussed in Chapter V). Each product of the multiplication of $1,697,000 with the output multi- pliers for livestock and products indicates the effect of that change 67 Table 4.3 Output Multipliers Derived from the Michigan 20-Sector Input-Output Model Sector Industry Output Multiplier 1 Livestock and Products 1.69 2 Other Agricultural Products 1.48 3 Mining 1.51 4 Construction 1.66 5 Food and Kindred Products 1.61 6 Lumber, Furniture, Paper, Printing 1.55 7 Chemicals, Drugs, Plastics 1.56 8 Petroleum Refinery 1.49 9 Rubber, Leather, Stone, Glass, Clay 1.52 10 Primary and Fabricated Metals 1.66 11 Machinery, Except Electrical 1.56 12 Electrical Equipment 1.48 13 Motor Vehicles and Parts 2.02 14 Aircraft and Other Transportation 1.55 Equipment 15 Transportation and Communication 1.29 16 Utilities 1.52 17 Wholesale, Retail, Miscellaneous 1.28 Manufacturing, Including Textiles 18 Financial, Insurance, Real Estate 1.32 19 Selected Services 1.39 20 Government Enterprises 1.47 68 Table 4.4 Effect of a $1,697,000 Change in the Final Demand of Livestock and Products on the Entire Economy (I-A)" Ele- ments in Livestock and Products Change in Sector Industry Sector Output 1 Livestock and Products 1.13273844 $1,922,257 2 Other Agricultural Products 0.19944936 338,466 3 Mining 0.00229381 3.893 4 Construction 0.01317304 22,355 5 Food and Kindred Products 0.12047448 204,445 6 Lumber, Furniture, Paper, Printing 0.01175515 19,948 7 Chemicals, Drugs, Plastics 0.02081888 35,330 8 Petroleum Refinery 0.00510590 8,665 9 Rubber, Leather, Stone, Glass, Clay 0.00705966 11,980 10 Primary and Fabricated Metals 0.01260634 21,393 11 Machinery, Except Electrical 0.00519639 8,818 12 Electrical Equipment 0.00118786 2,016 13 Motor Vehicles and Parts 0.00148287 2,516 14 Aircraft and Other Transportation 0.00017908 304 Equipment 15 Transportation and Communication 0.02254404 38,257 16 Utilities 0.00973669 16,523 17 Wholesale, Retail, Miscellaneous 0.05078442 86,181 Manufacturing, Including Textiles 18 Financial, Insurance, Real Estate 0.03286972 55,780 19 Selected Services 0.03713335 63,015 20 Government Enterprises 0.00198564 3,370 Total 1.68857513 $2,865,512 69 in final demand on the rest of the sectors in the economy. The sum of these individual sector output changes will equal the total change found originally by multiplying the total sector output multiplier (here, 1.68857513) by the change in final demand. However, the disag- gregation that is possible allows for an accounting of the sector-by- sector changes necessary to forecast rail traffic demand (Jones, 1978). The results of such a process across all sectors is shown in Table 4.5. The figures under "final demands" are those estimated as part of the original process of constructing the 1976 input-output table for Michigan. When each sector's final demands are multiplied by the individual sector Leontief inverse elements, and then summed, the input-output model projects total output as shown in the column labeled "output." Table 4.5 represents the base from which the rest of this analysis proceeds. The forecasting procedure begins by estimat- ing new final demands by sector for a year subsequent to 1976, and repeating the process outlined above to achieve total output projec- tions. These projections will then be compared with the 1976 total output figures to forecast the change in total output. Linking the Input-Output Table to Commodity Flow Data In order to forecast rail traffic using the input-output model, it must be linked to data on commodity flows over the Michigan rail net- work. This is done by linking the 20 input-output sectors to data from rail waybill samples. Transportation data has been compiled from a sample of audited revenue waybills (the rate of freight charges have been verified in the railroad accounting offices) submitted to the Federal Railroad 70 Table 4.5 1976 Final Demands and Total Outputs Derived from the Michigan Input-Output Model (Thousands of 1976 Dollars) Final Sector Industry Demands Output 1 Livestock and Products 13,766 691,909 2 Other Agricultural Products 210,555 709,533 3 Mining 926,488 1,742,344 4 Construction 4,302,101 5,837,164 5 Food and Kindred Products 4,335,922 4,869,530 6 Lumber, Furniture, Paper, Printing 2,111,828 5,095,917 7 Chemicals, Drugs, Plastics 3,416,858 5,472,216 8 Petroleum Refinery 568,180 928,370 9 Rubber, Leather, Stone, Glass, Clay 675,244 2,893,848 10 Primary and Fabricated Metals 3,576,180 14,666,321 11 Machinery, Except Electrical 2,909,304 6,661,440 12 Electrical Equipment 527,190 1,573,330 13 Motor Vehicles and Parts 28,679,640 39,769,180 14 Aircraft and Other Transportation 496,338 623,222 Equipment 15 Transportation and Communication 5,382,394 7,262,402 16 Utilities 2,086,750 4,217,498 17 Wholesale, Retail, Miscellaneous 14,185,652 17,880,538 Manufacturing, Including Textiles 18 Financial, Insurance, Real Estate 7,667,766 10,435,098 19 Selected Services 6,869,168 12,230,830 20 Government Enterprises 375,938 1,081,181 Total 144,641,871 71 Administration (FRA) by railroads under the terms of the Interstate Commerce Commission Order 49 CFR, Section 1244. Other documents used to compile the transportation data include EM-5 reports and multiple-car statements . Waybills are a shipping document prepared by the originating rail- road from the shipper's instruction as to the disposition of freight and used by the railroads as the authority to move the shipment and as the basis for determining and settling the freight charges among the carriers involved. The waybills in the sample used are for carloads terminated by line-haul or regular rail-haul as distinguished from a switching move or switching company. As used here, a regular haul is between terminals and stations on the main or branch lines of the rail- road, exclusive of switching moves. The sample also includes only those lines having $3 million or more average operating revenues over a three-year period (U.S. Department of Transportation, FRA, 1980). For the 1 percent sample, waybills are selected by the terminating carrier on the basis of the waybill number assigned by the originating carrier (this number is for the purpose of control and identification). The waybills selected are those numbered "1" or those which have numbers ending in "01." The selection criterion is designed to capture 1 per- cent of the audited waybills. Most waybills represent one-car shipments. Some represent multi- car shipments that vary from two to upwards of 100 cars. A sampling technique for these shipments was developed by the ICC and used by a number of reporting railroads. This technique, EM-5 sampling, provides for the selection and reporting of EM-5 forms, of data for every 100th car from waybills covering six or more carloads. Finally, some of the 72 reporting railroads submitted multiple-car statements listing total car- load movements between specific points for certain commodities such as coal, iron ore, copper ore, pulpwood logs, iron and steel scraps, and iron and steel plates. These statements are sampled by FRA with a computerized algorithm to approximate the EM-5 process (U.S. Department of Transportation, FRA, 1980). Some types of movements are excluded from the waybill data, in- cluding: (1) Canadian and Mexican originations. (2) Mail and express traffic. (3) Less than carload traffic. (4) Shipments weighing less than 10,000 pounds and moving on "any-quantity" rates. (5) Nonrevenue movements. (6) "Dead-head" traffic (company material whose movement is entirely on one road). (7) Empty cars returned for reloading. (8) Movements in substitute services, e.g., truck in lieu of rail. The data received from the railroad industry is processed by the Waybill Information Processing System, which combines manual and com- puter procedures to ensure that resultant tabulated data are accurate and statistically valid. The accuracy and validity of carload data for each commodity grouping is verified by use of regression analysis techniques. The waybill data is classified by commodity using the seven-digit STCC code, which is based on the SIC code. 73 The first step in using input-output to forecast rail traffic is to link the input-output sectors based on SIC codes to waybill samples, based on STCC codes. Table 4.6 shows the STCC categories in each of the 20 input-output sectors. After the input-output and waybill samples are linked, the 1 percent waybills are expanded, by multiplying by 100, to obtain an estimate of total rail traffic in Michigan. Table 4.7 shows the expanded waybill sample for Michigan for the years 1976 and 1980. The Use of Input-Output for Transportation Planning The assumptions of input-output stress the consistency of the economy; the outputs of all sectors must be consistent with all the .other sectors from whom it buys inputs and to whom it sells outputs. This feature of input-output analysis emphasizes the interdependence of the sectors of the economy. The value of the output of final goods and services can be traced back through the intermediate flows of goods and services necessary to produce final demand. This flow of goods and services for both intermediate and final demand must move in space from sector to sector and firm to firm. It is important that the assumption of consistency in the input- output model be met in the transportation model. Corresponding to each dollar flow between sectors in the input-output model, originating and receiving sectors can be identified showing the type of product that is being physically moved. The initial assumption is that economic activity in a sector generates physical movements of products in each of the sector's inputs, in proportion to the change in the sector's output. The input-output 74 Table 4.6 STCC Codes in Each Input-Output Sector STCC Code Contained Sector Industry in I/O Sector 1 Livestock and Products 014, 015 2 Other Agricultural Products 011, 012, 013, 019, 084, 086, 091 3 Mining 10, 14, 111, 112, 122 4 Construction 5 Food and Kindred Products 201, 202, 2032-2035, 2037-2039, 204-209 6 Lumber, Furniture, Paper, Printing 24-27 7 Chemicals, Drugs, Plastics 281, 284-286, 289, 282, 283, 287 8 Petroleum Refinery 29 9 Rubber, Leather, Stone, Glass, Clay 30-32 10 Primary and Fabricated Metals 332, 333, 335, 336, 339, 331. 3431. 3432. 34331, 34333-34336. 34339, 344-346, 341, 342, 348, 349, 191, 192, 196 11 Machinery, Except Electrical 351-359, 34332 12 Electrical Equipment 36 13 Motor Vehicles and Parts 371 14 Aircraft and Other Transportation 372-376, 379, 193 Equipment 15 Transportation and Communication 16 Utilities 17 Wholesale, Retail, Miscellaneous 22, 23, 38, 39, 211, Manufacturing, Including Textiles 212, 401, 402, 441, 412, 421, 422, 431, 441, 451, 461, 462, 471 18 Financial, Insurance, Real Estate 19 Selected Services 20 Government Enterprises 75 Table 4.7 Expanded 1 Percent Waybills, Michigan, 1976, 1980, by Input-Output Sector Sector Industry 1976 Tons 1980 Tons 1 Livestock and Products 0 0 2 Other Agricultural Products 1,211,200 1,545,000 3 Mining 31,641,100 35,463,300 4 Construction 0 O 5 Food and Kindred Products 3,476,700 2,667,800 6 Lumber, Furniture, Paper, Printing 4,607,900 4,021,100 7 Chemicals, Drugs, Plastics 4,180,300 2,713,700 8 Petroleum Refinery 2,520,400 1,679,000 9 Rubber, Leather, Stone, Glass, Clay 2,592,600 1,798,500 10 Primary and Fabricated Metals 6,025,000 3,684,100 11 Machinery, Except Electrical 242,300 111,400 12 Electrical Equipment 134,000 97,000 13 Motor Vehicles and Parts 13,204,600 8,170,800 14 Aircraft and Other Transportation 165,100 129,300 Equipment 15 Transportation and Communication 0 o 16 Utilities 0 o 17 Wholesale, Retail, Miscellaneous 5,502,600 4,224,800 Manufacturing, Including Textiles . 18 Financial, Insurance, Real Estate 0 0 19 Selected Services 0 O 20 Government Enterprises 0 0 Total 75,503,800 66,306,300 76 model of Michigan will be linked with commodity flow data to determine how movements of inputs flow on the state's rail network. By changing the final demands in the 1976 input-output model to reflect subsequent years, such as 1980, it is possible to estimate changes in the total output of the 20 sectors. With these changes, it is then possible to forecast changes in rail traffic due to the change in commodity produc- tion and the demand for the movement of these commodities. This pro- cedure was outlined in Chapter I and will be tested in Chapter V. CHAPTER V TESTING THE USE OF INPUT-OUTPUT FOR RAIL TRAFFIC FORECASTING The procedure used to test the forecasting ability of the input- output model is summarized in Figure 5.1. Beginning with the 1976, 20-sector input-output model of Michigan, final demands for the base year of 1976 and the test year of 1980 are estimated. The 1976 final demands are those estimated when the two-region input-output model was constructed. The 1980 final demands were estimated using as much Michigan specific data as possible. Most state input-output models rely on sharing techniques that apportion national changes in the final demand sectors to the state level. Estimating 1980 Final Demands There are six final demand sectors in the Michigan input-output model: (1) Personal Consumption Expenditures. (2) Gross Private Capital Formation. (3) Net Inventory Changes. (4) Net Exports. (5) State and Local Government. (6) Federal Government. All 1980 final demands are in 1976 dollars, deflated by implicit price deflators for gross national product indexes, found in the 1981 Economic 77 78 Figure 5.1 Flow Diagram of Rail Traffic Forecasting Test 1976 20- Sector I/O 1976 F05 m 1980 F05 W Leontief In- Leontief In- verse Matrix verse Matrix Times 1976 Times 1980 Final Demands Final Demands 1980 Total Output 1 Sector Per- cent Change: 79 1976 1% Expanded Waybills 1976 1% Expanded .._____+’ Waybills ' 1980 Sector Percent.Change Times 1976 Expanded Waybills 1 Projection of Rail Traffic 1980 1980 1% Projected 1980 Minus Actual Expanded *——) 1980 Divided waybills by Actual 1980 Test of Model Project 80 Report of the President (pp. 236-237). Therefore, the dollar value for 1980 final demands represents the real change in demand, from 1976 to 1980, and can be linked to the tonnage change on Michigan's rail net- work. The year 1980 was chosen as the test case because it represents the most recent year in which data are available. Following is the method by which the 1980 final demands were estimated. Personal Consumption Expenditures To calculate the value for personal consumption expenditures, Michigan's retail sales tax collections were used as a proxy for changes in consumer Spending. The Michigan Department of Treasury col- lects retail sales tax information in eight broad categories (building materials, general merchandise, automotive, food, apparel, furniture, miscellaneous retail and nonretail) with subsectors in most categories. This information is reported in the yearly Economic Report of the Governor (1981, p. 33). The percentage change in sales tax collections between 1976 and 1980 was calculated and this percentage change was used to increase or decrease the personal consumption expenditure figures used in the 1976 input-output model. These figures were then deflated to 1976 dollars by the total personal consumption expenditures index of the implicit GNP deflator. Gross Private Capital Formation The value of gross private capital formation represents the sum of receipts for new construction and capital expenditures of the manufac- turing sector. Michigan's data for 1980 was acquired from the Office of Revenue and Taxes, Department of Management and Budget. The data is from the reported capital acquisitions from the Michigan Single 81 Business Tax. Capital acquisitions at the two-digit SIC code level were reported for depreciable assets of firms in Michigan filing the Single Business Tax and represents a value added tax on investment. Not all investors in Michigan file the Single Business Tax; for example, nonprofit organizations do not. Farmers are also exempt, but estimates of farm investment were obtained from the Michigan Department of Agri- culture. These estimates of investment were then deflated to 1976 dollars, using the gross private domestic investment index of the im- plicit GNP deflator. Net Inventory Change Net inventory change was obtained from various Survey of Current Business reports. The percentage change in net inventories between 1976 and 1980 was used to increase or decrease the data on net inven- tory contained in the 1976 input-output model. These figures were adjusted to 1976 dollars by the GNP deflator. The percentages used for net inventory changes are based on national figures and not Michigan-specific data. No Michigan data could be located for this category; however, since it is the smallest part of final demand, the use of national figures should not appreciably alter the results. Net Exports Agricultural export information was obtained from the Michigan Department of Agriculture and fur manufacturing sectors from the Michigan Department of Commerce, Economic Development Office. Import information for the same categories was obtained from the U.S. Depart- ment of Commerce, Foreign Trade Statistics section. All data was de- flated to 1976 dollars by the export index in the implicit GNP deflator. 82 Federal Government This includes the total federal government disbursements minus the disbursements of the following federal government enterprises: post office; farm income stabilization; rural housing and public facilities; agricultural land and water resources; maintenance of housing and mortgage market; and veteran's benefits and services. These categories are included in the government enterprises sector. Data is from the 1980 Michigan Statistical Abstract, deflated by the federal government index of the GNP deflator. State and Local Government This includes the operating expenditures of state and local govern- ment agencies except: liquor stores: water transport and terminals; parking facilities; urban renewal; airports; and transit. These are included in the government enterprise sector. The state and local in- dex in the GNP deflator was used to produce real changes. This data was obtained from the 1980 Michigan Statistical Abstract and the 1979- 80 Government Finances. For both federal and state and local govern- ments, additional data was obtained from the study by Scheppach in State Projections of the Gross National Product, 1970,,1980. Once the two sets of final demands are estimated, the following steps represent the testing procedure: (1) 1980 final demands are multiplied by the Leontief Inverse Matrix of the 20-sector,input-output model to obtain 1980 estimates of total outputs. These total outputs are also calculated in terms of the percentage change from 1976. 83 (2) These 1980 total output estimates are then used to obtain projections, by sector, of 1980 rail traffic in Michigan. The sector percentage changes in total output are multi- plied by the 1976, 1 percent expanded waybills. The results are 1980 rail projections. (3) The 1980 rail projections are then compared to the actual 1 percent expanded waybills for 1980, yielding both the difference in total tons and the percentage difference between the actual total movement of rail traffic in Michigan and the projections of rail traffic. Testing the Rail Traffic Forecasting Method To project total outputs by sector for 1980, estimates of 1980 final demands are used to "drive“ the input-output model. The results of this are shown in Table 5.1, which indicates the projected total output in Michigan for 1980, given the estimated final demands. Table 5.1 also shows the percentage change in total output from 1976 to 1980. In constant dollar terms, Table 5.1 shows, for example, that total output in the livestock and products sector is down 10.59 percent from” 1976, and the motor vehicle and parts sector is down 40.13 percent. On the other hand, total output in the utilities sector is up 16.32 percent, etc. It is with the percentage changes of Table 5.1 that rail traffic for 1980 is projected. The actual expanded 1 percent waybills for 1976 and 1980 are shown in Table 5.2. To project 1980 rail traffic, the actual 1976 expanded 1 percent waybills are adjusted by the 1976-1980 percentage change in total output by sector. For example, in 1976, 1,211,200 tons of 84 _m._~- enm.e~m.m_a .eeop N_.~ amm.~o_._ 4.5.mce meeacacooew ecoEecosoc om Ne.m- Kam.em..._ mme.mmm.o mooascom ooeoo_om m_ _m.m mmo.mmm.cp ome.—mm.m cannon _eo¢ .ooeecsme_ ..aaoeceea m_ mopvuxmh mcwuspuem .mcpeapooeaemz em.o~- emo.mm~.ep www.mme.__ maooeeppoomez ._.ceo¢ .o_amopoez a. Nm.ep m_m.mom.e ~.a.wmm.~ mo_epppe= e_ KN.N- cam.amo.~ oe_.~em.m =o_ecoae=EEou we. =o_bcecoameece m_ gemEQF=cm m_.mN- owe.mo¢ awa.mom eoeeeacoamecce coeeo ea. beecoce< e. Pm.oe- ama.mpo.e~ ema.eo~.ep meted ego mo_oaeo> cope: m. _m.mm- moo.aep._ me~.eee aeoeaasem Feo_caoopm N_ mm.mN- mo~.mma.e Nae.ee_.~ .eoPLSoo_m eaooxm .acee_;oez __ NK.e~- eem.mo..__ eab.moe.m m_eaoz eooeo_coed eea aeee_ca o. am.em- _me.emm._ ema.cam aepu .mme.e .oeoom .coeoeon .coabem a mo.e~- Nmo.MSe owa.aam educate: e=o_ocooa m e_.em- mmc.mpa.m eme.om_.~ moveme_a .mmscn .m_eoaeoeu N am.@. ham.emc.e mwm.oe..~ acaeeeca .coaea .oc=o_=c=d .cooe=n e mo.mp- amm.em~.e _om.aaa.m mooseoca emcee_¥ eee econ m em.Fm- .mm.moo.e mm..omm.~ =o_oo=comeoo e ea.a Pm~._pa._ K_o._m~._ mepe_z m e~.m~ omm.eoa ~a~.eee meoaooca _ec=b_=oecm< cocoa N mm.op- ee~.o~e mme.m_ meosooca nee gooemoseo _ azauao cw usauzo nausea Face; anpmsuzu aouuom omcmgo “smegma cmuuonoca nouuonona Amce__oa camp eo meeameoeev eemaeoez _.m opace "mcovuummoea uaqpao Pouch camp 85 Table 5.2 Expanded 1 Percent Waybills, Michigan, 1976, 1980, by Input-Output Sector Sector Industry 1976 Tons 1980 Tons l Livestock and Products 0 O 2 Other Agricultural Products 1,211,200 1,545,000 3 Mining 31,641,100 35,463,300 4 Construction 0 O 5 Food and Kindred Products 3,476,700 2,667,800 6 Lumber, Furniture, Paper, Printing 4,607,900 4,021,100 7 Chemicals, Drugs, Plastics 4,180,300 2,713,700 8 Petroleum Refinery 2,520,400 1,679,000 9 Rubber, Leather, Stone, Glass, Clay 2,592,600 1,798,500 10 Primary and Fabricated Metals 6,025,000 3,684,100 11 Machinery, Except Electrical 242,300 111,400 12 Electrical Equipment 134,000 97,000 13 Motor Vehicles and Parts 13,204,600 8,170,800 14 Aircraft and Other Transportation 165,100 129,800 Equipment 15 Transportation and Communication 0 0 16 Utilities 0 0 17 Wholesale, Retail, Miscellaneous 5,502,600 4,224,800 Manufacturing, Including Textiles 18 Financial, Insurance, Real Estate 0 0 19 Selected Services 0 O 20 Government Enterprises 0 0 Total ,75,503,800 66,306,300 86 agricultural products moved over Michigan's rail lines. The 1980 total output in the agricultural products sector increased 28.24 percent over 1976 in constant dollars. Thus, it is estimated that rail movements over Michigan's rail system should also increase by 28.24 percent, or 342,042 tons. Consequently, the model estimated that rail traffic in the agricultural products sector is 1,211,200 plus 342,042 tons, or 1,553,242 tons. This procedure is carried out for each commodity sector, yielding the 1980 projections of Michigan rail traffic shown in Table 5.3. In order to evaluate the rail traffic forecasting ability of the input-output model, it is necessary to compare the 1980 projected rail. flows to the actual 1980 rail flows; Table 5.3 shows this comparison.' The projected total rail flow in Michigan for 1980 is 67,099,985 tons. The actual rail flow in 1980 was 66,306,300 tons. (This actual total does not include 1,638,200 tons of hazardous materials.) The model's projection of rail traffic is 793,685 tons higher than the 1980 actual figures, or 1.2 percent over the 1980 actual flows. The same procedure was carried out for the 44-sector input-output model and is shown in Appendix 2. Table 5.3 also includes a comparison of projected versus actual rail traffic for each commodity sector. When this method is used to project rail traffic on a regional or line-by-line basis, individual commodity groups will be important and their estimation will be crucial ,in producing reliable projections of rail traffic. In the Upper Peninsula, for example, rail service is used primarily by the mining, forestry, and related pulp and paper industries. The statewide projec- tions of these industries provide confidence that U.P. rail traffic can 87 mo_.N~ oo~.emm.me o~.p mmo.mmn oom.oom.mm mam.mmo.uo pouch o o oo.o o o .o womanncmucm geesecm>oo ow o o oo.o o o o . mmuF>emm umpuopmm m_ o o oo.o o o a «vegan poem .oucmssmcm .pewucecwa mp mmpvpxmh m:_uapo=H .m:_e:uumm==az mm_.p oom.mmm.~ mm.m om¢.~mp ooqumm.¢ on~.~mm.¢ maom=m~_mumwz .P_mumx .mPemmpozz N— o o oo.o o o o mmwuwauz @— o o oo.c o o o eowumu_==eeou ecu :o—uaucoamcmsh m— . aco5a_:cu me oom.mm mm.¢i mmw.ou oom.mup ~.m.m- cowuaueonceeh cacao ecu Homeucw< op oom.m oom.mnm.m em.m1 mum.mmmu oom.ohp.m .mmm._mm.u mpeoa use mmpuwno> Lope: m— mm oom.mv m¢.~ ¢_¢.~ ooo.mm. epv.mm “cesawsau Fouwcpoer up we oo~.om m~.¢m omm.om ooe._PF mam.-p .pmupcuum_m pamuxu .Amepeumz ~_ «we oom.mwo.p mo.m~ som.m¢m oo_.vwm.m Nom.mmm.¢ mpaumz emHmUwcnmu use Acee_ea OP men oom.mop.P Ne.mi om¢.~m- oom.mmu.p eoo._on._ ampu .mmopw .mcopm .cmguomo .gmnnam m mmm moo.¢ .o.pp mmm.¢m— ooo.m~o.p mmm.mom.p >gm=_emm eaopoeuma m mom oo_._mm me.p mvv.¢m oo~.mF~.~ me_.mmu.~ muvumapa .mmaca .m_eu_am;u m mam oou.mmm.F ~_.m m-.mo~ oop._~o.e omm.m-.e mewucwca .emaea .on=u_=c:a .smaan w New oom.mem.P «m.m_ www.mmm oom.nom.~ mmo.m~o.m mpuaeoea noeucwx can cook m o o oo.o o o o copuuseumcou e Nmm.¢ oom.Nmo.m _m.—i mmo.-o- oom.mme.mm m-.mm~mem . m:_:_: m omF oom.~nm mm.o ~¢~.m ooo.m¢m.P ~¢~.mmm.p muuauoea ~mc=upaupcm< canoe N o o oo.c o o o . mausuoca ecu xuoumm>wn — opqeom F.wa»a3 «pagom prnxmz augmemempo ouemnmeeao poauu< umuuonoca zoomzueu Loouom ommp cw ommp cw uncured meowum>comno cowumw>ma unaccmum m.m «Pooh Amcohv cmmvgowz :w mueoeo>oz upwemsh Ppmm empomnoea new pmauo< owmp :mmzumm convocanu 88 be predicted accurately. In Table 5.3, the forestry sector is an important component of Sector 2, other agriculture products. In this case, the model predicted rail traffic to within 0.53 percent of actual traffic. The projection for the mining sector was only 1.91 percent lower than actual traffic and the projection of rail traffic in the paper industry (within Sector 6) was 5.12 percent different from the actual rail flow. Most of the commodity sectors were projected within a percentage difference that is believed to yield confidence in the use of the model on a regional or line-by-line basis. Only two sectors--number 10, primary and fabricated metals, and number 11, machinery (except electrical)--produced estimates substan- tially different than actual figures, on a percentage basis. In the machinery sector, while the percent difference between the projection and actual rail traffic is 54.75 percent, the actual tonnage difference is only 60,996 tons. This absolute tonnage difference is not large compared to total traffic, representing only 0.09 percent of total traf- fic. The large percentage difference can be explained by looking at the individual industries that make up these two aggregate sectors. The primary and fabricated metals sector in the 20-sector, input-output model is the aggregation of six of the sectors in the 44-sector model (see Appendix 2): foundries and metal products; blast furnace; basic steel; metal containers, heating, plumbing, fabricated metals; screw machine products; metal stamping; and other fabricated metals. Of these industries, large differences between actual and projected rail flows occur in three: metal containers, etc.; screw machine products; and metal stamping. The actual traffic in 1980 for metal containers was 6,400 tons, while the 1980 projection was 90,883, or 1,320 percent 89 higher. In the other two sectors the actual traffic was zero, down from 1976. What occurred in these three industries between 1976 and 1980 was an extremely large drop in rail traffic--to zero in two cases-- that was not predicted by the input-output model. In these industries, rail traffic dropped much more than the estimated total output for the sectors. The problem appears to be in the waybill sample. Small absolute totals of rail traffic are less likely to be accurately re- flected in the 1 percent sample. While this presents problems to decision makers in a few cases, the 1 percent waybill data is the best available to rail planners and must be used in lieu of more complete rail information. 'While the input-output model can predict rail flows.in industries that exist on the lines between 1976 and 1980, it is unable to forecast traffic from some level to zero. However, this degree of change would be difficult to predict with any rail traffic model and it is not felt that the example of these sectors invalidates the usefulness of the model. A similar problem occurs in the machinery sector. Sector 11 is made up of the following industries: engines and turbines; farm and construction machinery; metal working machinery; industrial and service machinery; and office and computing machines. Two of the industries show the largest change, engines and turbines and industrial service machinery (see Appendix 2). In the case of engines and turbines, rail traffic in 1976 was zero; so the projection was zero. However, between 1976 and 1980, engines and turbine traffic again appeared in the waybill sample. In the industrial and service machinery sector, traffic dropped dramatically between 1976 and 1980, much more than the projected total 90 output. Again, it is felt that these cases are such that no model could predict such changes. Standard deviations for each sector in the way- bill sample were calculated to estimate the size of errors due to the waybill sampling procedure. In both Sectors 10 and 11, the standard deviation is large relative to its mean (30 percent'and 72 percent, respectively). This indicates that for these two sectors, part of the projection error is due to waybill sampling error. Of course, the large difference between the projected and actual values in Sectors 10 and 11 can also be attributed, in part, to errors in the technical coefficients in the input-output matrix and errors in the estimation of final demands. However, it does appear that a substantial portion of the projection error can be attributed to waybill sampling. In using this or any other forecasting procedure, balance must be maintained between a model's results and exogenous information available to rail planners about the nature of rail traffic. While the model alone cannot predict the change from high levels of traffic to zero, it is believed that rail planners would have additional information to complement the model's results. In any case, the majority of the sec- tors in the test show projections close to actual traffic. Conclusions The purpose of this chapter has been to test a method that system- atically links the projected level of statewide economic activity with the demand for rail freight transportation services and facilities. The goal is to broaden the perspective of decision makers to include a more comprehensive and systematic method of forecasting rail traffic in Michigan. 91 The test was whether the 20-sector, input-output model of Michigan could provide reliable estimates of rail traffic over a four-year period. The test was carried out not only on total rail traffic, but also on a sector-by-sector basis. The ability of the input-output model to disaggregate output projections is the key to its use in projecting rail traffic. This is so because the demand for rail services in a region, or on individual lines, depends on the output of the particular commodities that are produced in the area. The results of the test of the rail forecasting method indicate that it is effective in producing estimates of rail flows. The model projected rail traffic to within 1.2 percent of actual traffic over the total rail network. Additionally, the model was effective in providing sector-by-sector estimates in 11 of the 13 sectors in which rail traf- fic was represented in 1976 and 1980. In the two sectors in which the model overestimated rail traffic by approximately 23 percent and 55 per- cent, it is believed that the waybill sample displayed rail traffic changes between 1976 and 1980 which any forecasting model based on com- modity demand would have difficulty projecting. However, the method suggested in this study is not meant to be used in isolation from other information available to rail planners. Thus, it is believed that this model can be of value, along with additional data, in aiding decision makers in forecasting rail traffic. Since the test of the method has shown its ability to forecast total rail traffic over a four-year period, the next step is to il- lustrate its use in aiding the subsidy-abandonment decision. The next chapter uses the model to predict rail traffic beyond 1980 for Michigan's Upper Peninsula region. Using estimates of final demand, 92 various projections will be used to obtain forecasts of total output and forecasts of rail traffic in the U.P. and on a segment of the 500 Line. This will help determine whether, under any economic growth scenario, the segment of the 500 Line that is under a subsidy-abandon- ment study will be able to generate enough traffic to remain profitable, or, if not presently profitable, whether future rail demand will enhance its revenue position. The next chapter represents a guide to the use of the input-output model for future rail planning decisions. CHAPTER VI APPLYING THE RAIL FORECASTING METHOD TO REGIONAL RAIL PLANNING In this chapter, an illustration will be presented demonstrating how the rail traffic forecasting method tested previously can be employ- ed to aid in the subsidy-abandonment decision. Also included is a case study of the northern 500 Line in Michigan's Upper Peninsula. The il- lustration is only suggestive in nature because the financial data needed to perform a full-scale analysis of the case study region is un- available. The Upper Peninsula rail system is currently under study by the Michigan Department of Transportationl/ as part of the rail ration- alization process. The application of the rail traffic forecasting method will be accomplished in the following steps: (1) A description of the Upper Peninsula rail system will be presented with particular emphasis on those rail lines facing financial difficulties. (2) The Michigan input-output model will be used to project total output and Upper Peninsula rail traffic for 1983, given various estimates of final demands. Additionally, total output and U.P. rail traffic will be projected to 1/The material for this chapter on the U.P. rail system is from: Michigan Department of Transportation, "Michigan Rail System Rational- ization Plan: Tier II, Phase III, Upper Peninsula Report" (Draft), May 14, 1982. 93 94 1986, the last year of the subsidy appropriations in Michigan. (3) The estimates for Upper Peninsula rail traffic will be used to analyze major issues facing rail planners on a system-wide basis, particularly the potential commodity growth upon which U.P. rail lines depend. (4) Rail traffic will be forecast for the 500 Line, between Trout Lake and Dollar Bay, to aid in the resolution of the issue surrounding the need for a north-south connec- tion for 500 Line traffic. In both the entire U.P. and on the 500 Line, the future of the rail system is sensitive to the production and movement of certain com- modities. Consequently, the input-output method of forecasting rail demand by comnodity is particularly well-suited to addressing the issues of the U.P. rail system. In projecting rail traffic to 1983 and 1986, most significant is the commodity-by-commodity change in traffic since 1980. To be determined is which commodities may increase by the largest percentage, and whether those commodity increases produce sufficient rail demand to create financial stability. The Upper Peninsula Rail System The Upper Peninsula rail system consists of approximately 1,061 route miles with freight service provided by five railroad companies: the Chicago and Northwestern Transportation Company (C&NW); the Chicago, Milwaukee, St. Paul, and Pacific Railroad Company (Milwaukee Road); the Escanaba and Lake Superior Railroad (E&LS); the Lake Superior and Ishpeming Railroad (LS&I); and the 500 Line Railroad 95 Company (Soo Line). Three of these railroads (the C&NW, the 500 Line, and the Milwaukee Road) are major Class I carriers, while the LS&I is a Class II railroad and the E&LS is classified by the Interstate Commerce Commission as a Class III railroad, or "short line." Service provided by these railroads links the Upper Peninsula with Wisconsin and the Chicago gateway, with Canada at Sault Ste. Marie, and with the Lower Peninsula at St. Ignace utilizing the Straits of Mackinac car ferry service. The Upper Peninsula system functions mainly as an outbound rail route for metallic ores, forest products, and paper and pulp pro- ducts. Most Upper Peninsula traffic moves southerly into Wisconsin or to are locks located at Marquette and Escanaba. The Upper Peninsula rail system has experienced a steady contrac- tion in the last two decades. In 1965, the network consisted of 1,641 route miles served by eight carriers and three car ferry links to the Lower Peninsula. By 1976, six carriers remained, operating a total of 1,358 route miles, including the Straits of Mackinac interpeninsula car ferry route. The current system serves 1,061 route miles with 43 miles subject to active abandonment proceedings and 198 miles under study for possible abandonment. The map in Figure 6.1 depicts the current rail network in the Upper Peninsula. Abandonments since 1976 and current abandonment activities for the region are described in Tables 6.1, 6.2, and 6.3 [the source of the maps and tables is the Michigan Department of Transportation (MDOT) Upper Peninsula Report]. While years of piecemeal regulatory action have eliminated many light-density lines not essential to future needs, continued uncoordi- nated abandonment may threaten the essential core system of the 96 opzmepcmd song: we» a? momoepwam P.e deemed 9:1!!! .- .333:- 824 ........... 93.. 3233...: I I I I l I 0035083.. .3223 8:3 .- 5233 III l- 2532. :50: .- 039.8 I all: (Jam—52w.— CNAED 5:. Z. mo<0¢.= uoz H3 .moxm4.o new; .mnam voozeocu 322:2 22.8.52 832: t .3: 72 :5 85 485232.. as as: can. 35 mucoEEou one: umxuoo cm>—o>:~ cowumuon mew; uncappmm ae__ea Qua ma__z :ovumuppaa< ocoscoocmn< mcpucmd a on uumnnam maze; F_mm epsmcwcma Loan: m.o mpnop 100 Upper Peninsula.g/ A major consideration in U.P. rail restructuring is the extent of proven reserves of both metallic ores and forest products. The objective in restructuring is to ensure adequate access to rail- dependent economic activities and areas of significant future resource development. Service Needs in the Upper Peninsula and the 500 Line As illustrated in Figure 6.2, those lines identified in Tables 6.2 and 6.3 as in possible jeopardy include the entire northern segment of the 500 Line. As a part of the Tier I rail rationalization screening analysis, the Michigan Department of Transportation identified specific segments in the U.P. which were designated as essential core lines. These lines provide service to significant existing traffic bases and contribute to rail system and regional service objectives to a degree consistent with the results of the screening analysis on lines else- where in the state. Segments designated as essential core lines through the results of the Tier I analysis are identified in Figure 6.3. The screening analysis also identified lines which warrant designation as nonessential for servicing existing needs. These lines serve little traffic and exhibit virtually no potential for viability in the absence of significant rail dependent economic growth. These lines are also identified in Figure 6.3. The results of the Michigan Department of Transportation's screen- ing analysis were inconclusive for several U.P. rail segments. While these segments exhibit lower traffic densities and potential for viabil- ity, they provide service to rail dependent industries which generate Z/Michigan Department of Transportation, Upper Peninsula Report, op. cit. 101 zeeeaoon cw apnvmmoa weep; epamcpcma Loan: N.o ocea.a COB-£0; apZuzzoozcn... «a o m 30.30.. .‘WMWI «plaice-25:. oz... 00» 5230.. V‘\M\M\s >0¢<¢Oflq Z. >Jb2u¢¢30 «U23 I I | | I \ x .3032 .o 336 2.2.3.” . _ >oxow om oo.o o o o omN.~Nm.Np _¢N.mwm.s mmu.>gmm umuumpmm mp oo.o o o o NN..¢¢F.N_ mm~.¢ow.m oumumm puma .mucmgam:H ..a_ucmc.. m. mmpwuxoh mcwvapucm .mcwgzuuomacmz mm.m omm.o_ oo~.mm_ omo.mm— Pem.enm.mp em~.~mm.up maomcm—qum_z ...m.m¢ .m_amo_o;2 m~ oo.o o o o omo.mnm.m me.oo..m mm..__..: m. oo.o o o o mom.—mm.~ o-.mwo.c co.»ao.caesou can =o_pm.coam=mch m. pcmsawscm 00.0 o o o v.m.mom o.m..o¢ :o_umpcoamcmch cmguo wen .mmcucw< a. 00.0 o o o mmm..op.mm wmm.nme.mp mucma use mmpuwcm> Logo: m. oo.o o o o «.m.mmm.~ mmo._m¢ u:mEa.=cm .mu.c.um_u N. N... emo.P oo¢.m. ¢m¢.e_ mmo.m¢m.m mm~.o~m.~ .mu..uum.w pamuxu .zcmc.cumz __ mm.m Npm.m oop.mm Npe.¢op www.mqm.mp mmm.cmm.m mpmumz umumuwcam. vcm >Maewca o. a_u m..n mm..- oom.om~ mmm.~om mmm.om~.~ mmo.m~v .mmopw .ocoum .cmzumm. .cmonam m mm.“ new oom.m so..m «mm.mom mom.om¢ xcmc..mm saw—ocumm m av.“ ow_.m ooo.o. ow_.m. .No.mmo.¢ mom.oo¢.~ muwummpm .mmaca .mpau_Em;u N Nc.m s.¢.m- ooq..mm._ “.m.ome.~ o...wq_.m umo.mmm.~ m:.u:..a ..mqmm .m.=u.=c=. ..wnEa. m m~.m omm.¢ oom.mm emn.mm moo.~mo.¢ mmm.mmp.¢ muoscoca um.u:.x new coo. m 00.0 o o o mam.~m¢.¢ mmm.npm.m =o_uu=..mcou e om... “Nm.wmo.~ oom.~mm.sp Num.me.mp mo..eo..~ smo..mm.. mcwcwz m mm.m~ mmm , ooo.~ mm~.~ nm..~mm me~.mmq mpuzuoca .oc=u_=u..m< cmcpo N oo.o o o o mo..omo mom.o. muuzuogm can xuoumm>.. _ mmm. o_.woch ..mm u...m.h uwwwmcp ..mm “sagzo ucmswo .m:.. xgumsncH Louumm -82 E 2.55 :3. 8.: 3.8.2:. 8.3.2:. 3.8.8... omega. a ua.uaho.a ..=.u< m... m... m... meow dwgm_.oo cum. we mucmmaochv u...m.p ..mm ...2 .m.>..u< u_eocoom mo.zmpmum Aommp u Lam> mmmmv cowuomnoca saw: no..wm.h pwmm opzmcpcma swag: ucm auw>_pu< ¢.m m_amp ovsozoum macaw .o cowuumnoca mwm_ 106 l980 would require a high rate of growth over the next year. Con- sequently, if Michigan grows as much as the national economy, it would be considered to be a high growth scenario. The low growth scenario for l983 assumes a zero percentage change from l980; if the Michigan economy in 1983 matches 1980 data, it demonstrates a low rate of growth. Since the low growth projection is the same as l980, no separate table is needed; the percentage changes are zero. The medium range projection of final demand is shown in Table 6.5 and is the average of the high and low projections. Tables 6.4 and 6.5 show the projected final demands and total output given the different growth ranges, in 1976 dollars. The projection of the final demands for 1986 are assumed to represent a high growth scenario and are based on the RSQE estimate of the l982-83 change in real gross national product. The 1982-83 real GNP change is estimated to be 6.0 percent for the national economy. Thus, the final demands for 1983 shown in Table 6.4 are increased by l8 percent to yield a high estimate of final demands shown in Table 6.6. It is recognized that the procedure outlined above is imprecise. However, the purpose of this chapter is to provide an illustration of how the forecasting method may be used. Further use of the method re- quires a better procedure for estimating final demands, as well as an updating of the input-output model itself. Projecting Upper Peninsula Rail Traffic, 1980-1983 and l986 The projections of total output shown in Tables 6.4, 6.5, and 6.6 (projections of total state rail traffic are shown in Appendix 3) are used to estimate the change in rail traffic between 1980 and l983 and l986, also shown in the three tables. Given the high growth scenario, 107' 000.0o—.0 00.0 000.000.0— 000.000.00 000.000.00_ —0000 00.0 0 0 0 000.000., 000.000 00000000000 0000000>o0 00 00.0 0 0 0 000.000.00 000.000.0 0000>000 00000000 00 00.0 0 0 0 000.000.00 000.000.0 000000 0000 .000000000 .000000000 00 mm 0 .5wa 00000—000 .0000000000002 00.0 000.0 000.00— 000.000 000.000.00 000.000.00 000000000000: .000000 .000000003 0— 00.0 0 0 0 000.000.0 000.000.m 000000000 00 00.0 0 0 0 000.000.0 000.000.0 0000000005500 0:0 00000000000000 00 000500000 00.0 0 0 0 000.000 000.000 00000000000000 00:00 0:0 00000000 00 00.0 0 0 0 000.000.00 000.000.00 00000 0:0 0000000> 0000: 00 00.0 0 0 0 000.000._ 000.000 000000000 00000000_0 00 00.0 000 000.00 000.00 000.000.0 000.000.0 0000000000 0000x0 .000000002 00 00.0 000.0 000.00 000.000 000.000.00 000.000.0 00000: 0000000000 0:0 0M05000 00 000 00.0 000.00 000.000 000.000 000.000.0 000.000 .00000 .00000 .0000000 .000000 0 00.0 000 000.0 000.0 000.000 000.000 00000000 500.00000 0 . _0.0 000. 000.00 000.00 000.000.0 000.000.0 00000000 .00000 .000005000 0 00.0 000.00 000.000.0 000.000.p 000.000.0 000.000.0 00000000 .00000 .000000000 .000500 0 00.0 000.0 000.00 000.00 000.000.0 000.000.m 00000000 0000000 0:0 0000 0 00.0 0 0 0 000.000.0 000.000.0 000000000000 0 00.0 000.000.— 000.000.00 000.000.0— 000.000.0 0mm..00.~ 000:0: 0 00.0 00. 000.0 000.0 000.000 000.000 00000000 00000—30000< 1.0000 0 00.0 0 0 0 000.000 000.0_ 00000000 0:0 0000mm>00 0 000— 0.00000 .000 0000000 0000000 .000 000000 000000 00000 00000000 000000 u000_ 00 000000 0000 0000 000000000 000000000 000000000 00020 0 038.00.... 00300 0000 002 0000 0000 00000.00 000— 00 0000000000 0000000 .000 .0.0 000>000< 00000000 000300000 00000 n 000> 00000 0000000000 50000: ”0000000 0000 000000000 0000: 0:0 000>000< 00500000 00000 00 0000000000 0000 0.0 00000 108 mn.mm mmo.¢qm.o ooe.m~m.m_ mme.-m.o~ mce.mm~.me_ .mpOH 00.0 o o o «monomcnp www.mom mmmwgqgmucu acmscgm>ow om 00.0 o o o com.muk.¢_ ¢N¢.m_p.a mmu_>gom cmuum_mm mp 00.0 o o o mmm.mwm.¢. NNo.~mo._p mumumu Puma .mu=~g=m=_ .Pm_u=m=_m m_ wwwwuxmh mcwuapucu .mcvsauummacmz ow.m~ www.mq oo~.mo_ mNM.NF~ m-.m¢m.m_ Noo.-_.m_ maomcmp_mum_z .F_~pm¢ .QPMmm_ogz N_ oo.o o o o mqm.m¢m.o moo.o-.m mm_u___p= a, 8.0 o o o 03.2; 3.8: 8382228 2; 833.63%: .2 pcmaawacm oo. o o o Nev.ooo mmm.m~¢ =o_u~»gogm=agh gaggo ago “*agug_< ¢_ 00.0 o o o mfim.mxm.fim «mm.emm.- wagon new mopo_gm> Logo: mp oo.o o o o. _Nm.m_m._ moo.~mm «cmsa_=cm Pmu_guum_u N— Nm._~ omm.~ ooq.m~ omm.mp mom.Nm_.o mem.¢om.~ _mo_guuwpm ugmuxm .Xgm:_;umz __ m..q~ ~_N.m~ oo_.om ~.m.o_. “mm._~m.cp cmo.moo.¢ m_a»mz cmumovgnmu cam x»~ewgm o_ «.0 mo._~ NQN.om oo~.om~ ~¢a.o¢m m-.~mm.w m_m.eom .mm~_u .mcoum .gmgumm4 .Lmnnam a . mo.m~ oom.~ oom.m ome.o_ Noe._om o_m.mpm xgm=_mma E=m_oguma m mN._N oqm.m_ ooo.o~ www.mm mm¢.mmm.c wa.mmm.~ mu_uma_a .mmaga .mF~u_Emgu N m~.mm www.msm ooe._mm.. qmu.~oh._ om¢.¢~o.m Nom.¢mh.~ m=_u=wga .Lmamm .mg=u_cgau .Lm3534 o .m.m~ _mo.qp oom.mm .mN.mm _m~._xq.m coo.mhm.¢ mausoogg uwgucwx ucm voou m oo.o o o o m_m.omm.m mo_.mmm.m =o_puagumcou ¢ w¢.mm omo.o_o.o oom.Nma.~_ omm.mom.m~ oom.~m¢.~ mmm.¢mm.P mew=_z m mc.mm no“ ooo.~ mo~.~ em¢._fl_._ NN..@Nm muuauoga F~L=HP=U_LO< gmguo N oo.o o o o Nam.~om mem.m_ mpuauogg can xuoumm>va F omm_ memagh __~¢ u_%$mgh uwgmmgh F_~¢ paauao ucmsmo _mcwu xgum=u=_ Loyomm -omm_ =_ mmcmgu __mm omm_ umuumfiogm umpumnogm umuuohogm mmcmzu a umgumnogg _m=uu< mmm_ mmm, omm_ mach AmLmP—oo onm— mo mncmmzoshv uwm$agh __m¢ .¢.= »»_>vuu< u_eo=oum mnpzmgmpm Aomm_ u me> mmmmv :owuumnogm ;m_: uo_w$mgh mem mpzmcwcmm swag: uzm mpp>_uo< owsocoom mumum $0 =o_pomnoga wwmp o.o mpnmh 109 rail traffic in the Upper Peninsula is projected to increase 11.07 per- cent for l983, with the agricultural (including forestry), mining, and lumber sectors showing large increases compared to other sectors. The medium projection shows an increase in rail traffic between 1980 and 1983 of 5.45 percent. Finally, the 1986 projection shown in Table 6.6 estimates a total U.P. rail traffic increase of 32.78 percent. The sector changes in agriculture, mining, and lumber products are particu- larly important in the Upper Peninsula case study. The future of the U.P. rail system is tied to the demand for transportation generated by those industries. In analyzing the service needs of each of the Upper Peninsula rail lines, the Michigan Department of Transportation has emphasized the .potential development of particular commodities. In the Upper Peninsula, the principal industries are highly dependent on rail service. There- fore, an adequate rail system is necessary to provide for the future development of the U.P. resource-based industries. In the rail rationalization process, potential Upper Peninsula rail needs were determined by first identifying areas containing proven economic reserves of iron ore and copper, and those characterized as affording high potential for forest product industries. A second level of potential was addressed by identifying areas having proven sub- economic iron and copper deposits and medium potential for forest pro- duct industries. The areas of proven economic ore deposits and high potential for forest product industries coincide with rail lines already identified in the essential core system or lines which were marginal, but providing service to significant employment generators. The latter 110 set was included in the definition of the essential core system based on both existing and potential needs. The MDOT's analysis indicates that the viability of the Escanaba and Lake Superior Railroad, the Lake Superior and Ishpeming, and the Chicago and Northwestern all depend on either iron ore production or the forestry, lumber, and paper industries, or both. Thus, potential production in these sectors plays a crucial role in determining whether those lines will be able to generate sufficient traffic to remain or become profitable. The rail traffic projections produced by the input- output model indicate that those are the industries predicted to gener- ate the largest increases for all growth ranges (the lumber sector is projected to increase slightly less than Sector l7). As is evident in the waybill projections in Tables 6.4, 6.5, and 6.6, the mining sector makes up approximately 90 percent of total U.P. rail traffic. The pro- jected increase of rail traffic for that sector to ll.30 percent in l983 and 33.48 percent in 1986 indicates that several already strong ore mov- ing lines will remain viable. The largest sector percentage increase is in the agricultural pro- ducts sector. In the Upper Peninsula, most of that sector is made up of forestry products. Further, the lumber, furniture, paper, and printing sector--or forestry related industries--exhibits the fourth largest pro- jected change. Thus, the rail forecasting method provides evidence that, at each growth rate, the sectors that are most important to the future of the U.P. rail system are those that generally show the largest gains. Consequently, there may be an incentive to continue state sup- port rather than allow abandonment. In this manner, the input-output method of rail forecasting can be applied to the subsidy-abandonment lll decision. The disaggregated, commodity-by-commodity nature of the fore- casting procedure does provide the type of information required in the rail planning process. The 500 Line and the North-South Route Issue In this section, a demonstration of the use of the rail forecast- ing procedure is presented to analyze a specific issue related to one rail line in the Upper Peninsula. The decision involved is which of four alternatives for a 500 Line north-south routing would be most ef- ficient. The decision depends largely on the future levels of traffic over the northern tier of the 500 Line. Therefore, the forecasting of rail traffic can aid in determining which alternative should be pursued. The 500 Line Railroad,§/ due to past mergers and changing traffic patterns, must move its northernmost traffic by a circuitous route over long stretches of light density track. Presently, traffic from the Houghton, L'Anse, Ishpeming/Negaunee, Marquette, and Munising areas must be routed east through Trout Lake, then west to Gladstone before it can travel through Wisconsin to major national markets via the Chicago gateway (see Figure 6.4). In order to avoid circuitous routing, managers of the 500 Line have expressed a need for a more direct north-south route. Negotiations with the Chicago and Northwestern ‘U: allow the $00 to gain trackage rights over the C&NW main line from Ishpeming to north Escanaba have been held, but no agreement was reached. The 500 has also negotiated for the possible use of LS&I and E&LS trackage from Humboldt Mine to éjDiscussion of the 500 Line from: Michigan Department of Trans- 'portation, Upper Peninsula Report, op.cit. 112 m=P4 com . mm>wumcgmpp< gusomuspcoz ¢.o ma=m_a 02.22:». .- coEusaa 3:3 .. ........... 93: $234.54.: I. I I I I l 02:60:22 coiwaam 3.5 .- (3233 III: ..l 2552: 5:02 .- oo<9zu II . III I UZ... 00m III I 'I (4:9,:me mug: 3:. Z. ma<0m.:am 3.0m, .3223... 112 mcw4 oom 02.2232 .- c0333 3:: .. .:.... . 93¢ 3.555.: IIIIII. .3323 3.5 .- 3533 III I. 2552.. 3:32 .- 098:8 III- all... .. at... 09 ..IIIIII (Jaw—4.2mm mums: m1... 2. mo<0m.=_um=gmpp< guaom-;ugoz ¢.o ma=m_a -.mwzwmzmmt O O O COG—5°C.: 333. $.Ecomo O U 9.: .236... \mfi. m .~'0—o—. ...\. co 320 3233 o o \I . .Ino/oulc 52532 \UuZK Scam /./ I II I . 33¢ 53/ oIfluJIho.» or 625.301.... .ucz ”new .:....2 . // 80/5903; of] 825‘ {I302 023.3 :02: up 32.3. I 8853! 303:0... 86380\ .:..—cazd 000000 I i ‘ .:.... ............... ...Jf / ...z...\ 36“]. 3-3 ’0: ” 06:20.85; Of .:‘OOCO >3 sank! . 802-033.. 113 Pembine, Wisconsin. The 500 Line may attempt to abandon the light density track between Trout Lake and Shingleton without first securing a north-south outlet for northern tier traffic. This could lead to further abandonment of light density segments ultimately increasing costs to the shippers and adversely affecting potential development. Four alternatives for a more efficient 500 Line north-south rout- ing have been identified by the Michigan Department of Transportation and are shown in Figure 6.4: (l) Shingleton to Manistique: utilizing abandoned rights-of- way now owned primarily by the state and federal govern- ments. (2) Munising Junction to Rapid River: utilizing rights-of- way abandoned by the LS&I and 500 Line. (3) Ishpeming to north Escanaba: utilizing the C&NW line. (4) Humboldt Mine to Pembine, Wisconsin: utilizing the LS&I Line from Humboldt Mine to Republic, then linking with the E&LS from Republic to Pembine. The first alternative, Shingleton to Manistique, would require totally new construction and would most likely not be cost-effective. Most of the route is within federal and state forest areas; however, the southern five or six miles are privately owned. According to the Michigan Department of Transportation, new construction along this route would cost $375,000 per mile, excluding right-of-way purchase on the southern end. The total construction of this route would cost approxi- mately $l3.2 million. The second alternative, Munising Junction to Rapid River, would re- quire new construction to replace trackage which has been removed. . . - - s ..h... ... are rcxnntiudl a tical-Ii 114 Assuming a new construction cost of $375,000 per mile, reconstruction of this route would cost approximately $17,500,000. The Michigan Department of Transportation estimates that this alternative is also not likely to be cost-effective. Use of the C&NW track (Ishpeming to north Escanaba) would probably be the least costly to implement. Information on track condition is not available, although the line is believed to be in reasonably good condi- tion. Via this route, circuitous routing would be eliminated. Serving Forest Center via Humboldt Mine and Channing, over the E&LS, would have similar advantages. The line from Channing to Republic, however, is in poor condition. The cost of rehabilitation for 23 miles of track is estimated by the Michigan Department of Transportation at approximately $5 million. In part, the distinction between these alternatives will depend on whether traffic density on the northern 500 Line will increase over the next few years. Higher densities and different commodity compositions would change the revenue situation for the northern 500 Line and could make any of the four alternatives feasible. It is at this point that forecasts of 300 Line traffic are needed. The 1983 and 1986 projections of 500 Line rail traffic are shown in Tables 6.7, 6.8, and 6.9. The actual l980 rail traffic shown in Table 6.7 is 246,800 tons, carried in 4,200 cars. The 1980 traffic averaged approximately 59 tons per car. The total traffic between Trout Lake and Dollar Bay is approximately 9,300 cars, which includes 1,029 that move on the LS&I between Munising and Munising Junction. The 4,200 cars are those that appear on the 1 percent waybills and it is assumed that the traffic not on the waybills is proportionate among the commodity sectors in the same fashion shown 115 00.0 000.0. 000.00N 000.00N 000.0N0.0N_ 00000 00.0 0 0 0 00N0000._ .0N.000 00000000000 0000000>00 0N 00.0 0 0 0 000.000.00 000.000.N 0000>000 00000000 00 00.0 0 0 0 000.000.__ 000.N00.0 000000 0000 .000000000 ..00000000 00 00000000 mcwuzpucm .0000000000002 00.0 000.0 00N.00_ 00N.000 000.000.00 0NN.00_.N0 000000000000: .000000 .000000003 00 00.0 0 0 0 000.000.0 000.0N0.0 000000000 00 00.0 0 0 0 000.000.0 m00.m.0.0 0000000000500 000 00000000000000 00 000500000 00.0 0 0 0 000.000 000.000 00000000000000 00000 000 00000000 00 00.0 0 0 0. 0N0.000.0N 000.000.00 00000 000 00000000 00002 00 00.0 0 0 0 0.0.00N.0 000.000 000000000 0000000000 N. 00.0 0 0 0 000.000.0 000.00N.N 0000000000 0000x0 .000000002 __ 00.0 0 0 ,0 000.0NN.00 00N.00N.m 00000: 0000000000 000 0M00000 00 000 . 00.0 000.. 000.00 000.00 000.N00.N 000.000 .00000 .00000 .0000000 .000000 0 00.0 0 0 0 000.0NN 00N.0_0 00000000 000000000 0 00.0 000 000.0 000.0 0.N.000.0 000.00N.N 00.00000 .00000 .000000000 0 00.0 0NN.0 000.00 0N0.N0 000.NN0.0 00N.NNN.N 00000000 .00000 .000000000 .000000 0 N_.0 000 000.0 00N.0 000.000.0 000.000.0 00000000 0000000 000 0000 0 00.0 0 0 0 N00.00N.0 N00.000.N 000000000000 0 00.0 000 000.0 00N.00 000.000.N 000.00N._ 00000: 0 00.0 0 0 0 000.000 00N.N00 00000000 000000000000 00000 N 00.0 0 0 0 .NN.000 000.0. 00000000 000 000000>00 0 000— 0000000 .000 0000000 0000000 .000 000000 000000 00000 00000000 000000 -000. 00 000000 .000 000. 000000000 000000000 000000000 00005 0. 00000000 000000 0000 002 002 0:0» ~w00ppoo mnmp 00 0000000200 0000000 0.00 0000 000 00000000 00000000 000300000 ”0000000 0000 0000 cam 0:0 >00>000< uwsocoum 0000m :o_0omnog0 5:000: n.c mPnMP 00 0000000000 mam— 116 .0.0 0.0.00 000.000 000.000 000.000.00. .000. 00.0 0 0 0 000.000.. 000.000 000.0000000 0000000000 00 00.0 0 0 0 000.000.0. .00.000.0 000.0000 00000.00 . 0. 00.0 0 0 0 ~0..00..0. 000.000.0 000000 .000 .00000000. ..0.0000.0 0. 00..0x0. 00.00.00. .00.000000000z 00.0 000.0. 000.00. 000.0.. .00.0.0.0. 000..00.0. 000000..000.0 ...0000 .0.000.000 .. 00.0 0 0 0 000.000.0 .00.00..0 00.0...00 0. 00.0 0 0 0 000...0.0 000.000.0 00.000.000000 000 00.00000000000 0. 00000.:0m 00.0 0 0 0 000.000 0.0..00 00.00000000000 00000 000 000000.< 0. 00.0 0 0 0 000..0...~ 000..00.0. 00000 000 00.0.000 00000 0. 00.0 0 0 0 0.0.000.. 000..00 00000.000 .00.0000.0 0. 00.0 o o o mmo.m0m.m mm0.00m.~ 0000000000 ugmuxu .000000002 .. 00.0 0 0 0 000.000.0. 000.000.0 0.0000 00000.0000 000 0M00.00 00 0.0 0... 000.0 000.00 00..00 000.00..0 000.000 .000.0 .00000 .0000000 .000000 0 00.0 0 0 0 0.0.00. 00..000 0000.000 000.00000 0 00.. 0.0 000.0 0.0.0 .00.000.0 00..000.0 00.000.0 .00000 .0.00.0000 0 00.0 .00.0 000.00 000.00 0...00..0 000.000.0 00.00.00 .00000 .0000.0000 .000000 0 00.0 000 00..0 000.0 000..00.0 000.00..0 00000000 00000.0 000 0000 0 00.0 0 0 0 000.000.0 000...0.0 00.000000000 0 00... 000.. 000.0 000.0. .0..00..0 .00..00.. 00.0.0 0 00.0 0 0 0 000.000 000.000 00000000 .0000.00.000 00000 0 00.0 0 0 0 00..000 000.0. 00000000 000 0000000.0 . 000. 0.00000 ..00 0.00000 0.00000 ..00 000000 000000 .00.0 0000000. 000000 -000. 0. 000000 ..00 000. 000000000 000000000 000000000 0000000 000000000 .00000 000. 000. 000. 0000 00000.00 000— we 0000000000 0.0.00. ..00 00.0 000 00.>.00< 0.000000 00.300000 00.0000000 00.: 00000000 .000 0000 com 000 000>000< 00000000 mpmpm mo 0000000000 mmm. m.o @0000 117 NN.©N 0mm.eo oom.meN omN.__m m¢¢.amN.m¢N NNNON oo.o o o o Nmo.oN¢,_ ¢N¢.wom m~m_gqgmucu Name=Lm>oo oN oo.o o o o com.mNN.e_ ¢N¢.N__.a mmu_>gmm umuum_mm m. oo.o o o o omN.mNN.¢_ NNo.Nmo.__ mpaumm puma .mu:~g=m=_ ..qucacmm N. mmpwuxmh m:_uapu:~ .mc_gauomwzccz om.mN oqm.oN ooN.¢o_ me_._m_ NNN.N¢N.N_ Noo.PNF.m_ macm=~__mum_= .__mNmN .m_cmm_o;3 N. oo.o o o o mem.mvm.o ooo.mNN.m mm_u_.wu= m_ oo.o o o c on.oN_.m mpo.omp.N =o_uauvcaesou veg =o_uougoamcmgh m_ “cosq_:cm oo.o o o .o Nee.ooo mmm.mmc scrumugogmcagh gozuo ucm umeULN< e_ oo.o o o o mNN.mNm._m qmm.¢Nm.NN mugua ace mo_uw;m> Lego: N. oo.o o o o _Nm.mpm.N moo.Nom acmeQ_=aN _Nu_guum_m N. oc.o o o o moN.Nm_.o mem.qom.N .muwguum_m ugmuxu .xgmcwzumz _F oo.o o o o NNN.PNm.e_ moo.mom.e m_ogmz umuou_gnmm tau NMmE_LN op. N_u No._N cNm.N ooq.mm «No.me mNN.NNm.N N_N.oom .mmm_o .ocoum .nguom4 .Lmonam a oo.o o o o No¢._om o_m.m_m Ngo=_NmN E=m_oguma N NN._N mmo._ ooo.m mwo.o mm¢.mmm.¢ NNN.NNN.N muvum~_m .mmago .m_mu_sm;u N NN.NN mm¢.¢N ooq.mw mmN.N_P oN¢.¢No.o Ncm.qu.N mcvu=_gm .Lmamm .mgauvcgam .Lm3534 e m¢.mN mqo._ oo,.¢ mqp.m .NN._N¢.m voo.oNN.¢ muuzuogm umgu=_g cc“ woo; m oo.o o o o N_N.oNN.m No_.mNN.N cowuuagumcou e N¢.Nm NqN.m ooN.m Ncm.N_ ooN.NN¢.N mmm.¢¢m._ m=_=_z N oo.o o o o «NN._NP._ NN_.on muuanogm _NL=N_=U_LQ< Loguo N 00.0 o o o Nmm.Nom mqm.m_ “Nuauoga ecu xuonm>_4 _ ©Nm_ u_NNuLH __~a ovmwmgp uNNNNLP ..ag azauzo ucmsmo _o=_m Ngumzucm Legumm -omm_ =_ macago .Nmm omm. umuumnoga umaumflogm umuumhogm 3:5 N 880?: :32 £2 £2 £2 mach Ange—poo oum— no mucomsochq uwmmogh pwnm mcwa com xu_>wuu< umeocouw quZmumum cowuumfloga cow: uu_mwmgh Pwmm mcw4 com vac xpv>wuu< oweocoum macaw we cowuumnogm camp m.o wpnmh llS in Table 6.7. Consequently, the relative changes in the 500 Line pro- jections should remain the same. The 500 Line traffic shown in Tables 6.7, 6.8, and 6.9 represents only that traffic which runs on the northern tier. As was shown in Figure 6.2, the remainder of the 500 Line is not in jeopardy. The forecasting method has the advantage of being disaggregated by commodity and the use of the waybill samples also allows the isolation of specific rail segments due to the identification of origins and destinations, by county.5/ Given the average of 59 tons per car that occurred in l980, the number of cars projected on the northern 500 Line increases to 4,380 cars in the medium projection and 4,572, given a high economic growth scenario. The number of cars projected for l986 is 5,284. The northern 500 Line between Trout Lake and Dollar Bay, exclud- ing the Newberry-Forest Center segment which will probably be abandoned, is approximately 177 miles. Thus, in l980, the northern 500 Line traf- fic was 23.7 cars per mile. The projections for 1983 range from 23.7 cars per mile to 25.8 cars per mile with the 1986 cars per mile estimate rising to 29.9. Comparing these rail density projections with other rail lines in Michigan that are in financial jeopardy, it appears that the northern 500 Line will not generate sufficient traffic to become profitable. Figure 6.5 shows the frequency distribution of rail density on all fljThe Michigan Department of Transportation retains station desig- nations on the waybill files for applications to circumstances which require greater specificity. To permit analysis at alternate levels of aggregation, MDOT maintains a station to 547 zone equivalence capability, which was aggregated to the county level for this study. The MDOT is currently in the process of disaggregating the rail network to the 2300 zone level. No. of Rail Segments 119 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 --I-------I-------I 1— I ------- 1 ------- I ------- 1 ------- I-- 10.34 51.71 93.09 134.45 175.33 31.03 72.40 113.77 155.14 Cars/Mile ICC Category-System Diagram Classification: l - Carrier anticipates line segment will be subject to an abandonment or discontinuance within the next three years. 2 - Line segment under study and potentially subject to abandonment application. 3 - Line segment for which an abandonment or discontinu- ance is currently pending before the ICC. 4 - Line segment being operated under the 4R Act. Figure 6.5 Frequency Distribution of Rail Density on All Michigan Railroads Classified in ICC Categories l-4 120 Michigan railroads (63 segments) that are classified in ICC categories l-4; from currently subsidized lines to segments that may be abandoned within the next three years. On these 63 segments, the mean rail density is 34.2 cars per mile; the majority of segments have densities below 30 cars per mile. Oh those segments in ICC categories l and 2 (those most like the northern 500 Line), the mean traffic density is 39.5 cars per mile. Under any of the economic growth scenarios, the traffic on the northern 500 Line resembles segments that do not support profitable operations. Further, the rail density on the profitable southern 500 Line was 280 cars per mile in l980, with a similar com- modity composition. In order to make precise statements about future profitability, it is necessary to understand both the revenue and cost structure on the northern 500 Line. Revenues change with commodity composition and levels of traffic. Costs are also sensitive to traffic levels with on-branch costs such as locomotive costs, car-day cost, and car-mile cost dependent on the number of carloads. However, the traffic densi- ties projected for the northern 500 Line do not appear to be sufficient to generate profitability. In terms of the north-south routing issue, as the traffic density characteristics remain virtually unchanged even over high growth projec- tions, revenues will probably not improve. Therefore, given the costs of the various north-south route alternatives, the implementation of alternative 1, 2, or 4 would not ensure a significant improvement in the future financial stability of the northern 500 Line service. This conclusion is similar to the tentative decision of the Michigan Depart- ment of Transportation. In the Upper Peninsula report, the Michigan lZl Department of Transportation concludes that the only proposed alterna- tive which would both improve operations efficiency and overall system density is the joint 500 Line and C&NW operation over the C&NW track from Ishpeming to Escanaba. This alternative uses only track mileage included in the essential core, and would permit the abandonment of otherwise nonessential mileage. The conclusion arrived at through the use of the input-output rail forecasting method indicates that if a north-south route from the 500 Line is to be pursued, the joint 500 Line and C&NW operation would be most appropriate. However, it does not appear that rail traffic in the northern tier of the 500 Line, under any economic scenario, will gener- ate sufficient levels of traffic to become profitable. Consequently, one alternative may be to allow some abandonment on the northern 300 Line rather than to go forward with efforts to encourage a link between the 500 Line and the C&NW. The decision on whether to subsidize or aban- don a line, however, revolves around issues in addition to future profitability, such as employment and noneconomic factors including political concerns. Yet, any service which may be justifiable on a more comprehensive benefit-cost basis, would likely require permanent operating assistance. The results of this study indicate that the is- sue of future profitability is not one that can be used to justify future subsidies. CHAPTER VII CONCLUSIONS The purpose of this study has been to suggest, test, and illustrate the use of a method of rail traffic forecasting that systematically links the level of statewide economic activity with the demand for rail freight transportation services and facilities. The goal has been to broaden the perspective of decision makers to include a more comprehen-. sive framework that predicts rail traffic by forecasting the total outf put of a state's economy. The demand for freight transportation is a derived demand dependent in part on the level of total output in an economy. The working assumption has been that activity in an economic sector generates the physical movement of the aggregate of products in- cluded in its input sectors in proportion to the change in that sector's output. The method suggested in this study has been to link a 20-sector, input-output model of Michigan to 1 percent waybill samples showing the movements of goods on the state's rail lines. The use of input-output as a forecasting method is particularly well-suited to the needs of transportation planning. Input-output does not simply project total economic activity using macro variables, but is disaggregated by com- modity groups-~the same commodity group whose future production will determine the demand for individual rail service. With this informa- tion, it is possible to both estimate future rail traffic and evaluate 122 123 the potential for profitability on currently subsidized lines, or lines in jeopardy. This study has been specifically targeted at the estimation of rail traffic in Michigan and Michigan's Upper Peninsula. The projec- tions used in rail planning decision making have been ad hoc in nature and not related to changes in economic activity in general. In Michigan and elsewhere, rail traffic projections have been based on shipper surveys whose interests are generally to retail rail lines. Con- sequently, there is a perceived need to operationalize a method to forecast rail traffic. In this study, a conceptual framework of the forecasting method was established that links the structure of an economy to the demand for rail freight services. A 20-sector, input-output model of Michigan was linked with commodity flow data to determine movements of inputs on the state's rail network. Each of the 20 sectors in the input-output model was matched to commodity classifications on 1 percent waybill samples through SICs and STCCs. Changes in total output projected by estimates of final demands in the input-output were used to project rail traffic given by the waybill data. The major point of this study has been to test the use of the input-output method in forecasting rail traffic. The test consisted of projecting 1980 rail traffic on the basis of 1976, 1 percent waybill data and comparing that projection to the waybill sample observed for 1980. Starting with the 1976, ZO-sector, input-output model of Michigan, final demands, by sector, were estimated for l980 and substituted for the l976 final demands. Total output, by sector, was estimated by the input-output model through the multiplica- tion of the new final demands by the elements of the inverse matrix. l24 The expanded 1 percent waybills representing total rail traffic in l976 were then multiplied by sector, by the real percentage changes forecast by the input-output model. This procedure produced the l980 rail traf- fic projections which were compared to the actual l980 waybill data. The major conclusion of the test and, thus, of this study is that: A rail forecasting method based on a two-region input- output model is effective in producing estimates of rail traf- fic; the model projected rail traffic to within 1.2 percent of actual traffic over the total rail network. Additionally, the model was effective in providing sector-by-sector estimates of rail traffic moving over Michigan's rail lines in l980. These sector-by-sector estimates are crucial in predicting rail traf- fic in regions within a state where rail demand is based on a few specific commodities. After the testing of the forecasting procedure, the study provided an illustration of the use of the model in regional rail planning decisions. The case of Michigan's Upper Peninsula was used to demon- strate the usefulness of having projections of rail traffic when making subsidy-abandonment, or rail rationalization decisions. In the case study, it was projected that for the entire U.P. rail system, those commodities that are necessary for system viability are those whose production will increase by the largest amounts. Mining activities, forestry, and lumber and paper related industries are three of the four industries whose total output is projected to increase significantly. Therefore, for the case study region, future economic activity may pro- vide sufficient levels of total output to improve the financial viabil- ity of the region's rail system. The forecasting procedure was also used to examine a specific is- sue on one rail segment in the Upper Peninsula; the northern 500 Line. The illustration of the uses of the input-output model emphasizes the flexibility of the method in dealing with individual rail issues. Not 125 only is the input-output model able to disaggregate the effects of changes in specific commodity production, but through the waybill sample it is able to isolate particular county-to-county rail segments. The case study indicated, with its estimates of rail traffic, that the traf- fic along the northern 500 Line will remain similar to other rail seg- ments in Michigan in financial difficulty over all levels of projected economic activity. This conclusion aids in the transportation planning procedure by providing rail planners with the information needed to determine which north-south route alternative is most efficient. The results of the case study further indicated that the traffic density on the northern 500 Line will remain so low that permanent operating as- sistance will be required if service is continued for reasons other than future profitability. The second major conclusion of this study, therefore, is: A rail forecasting method based on a two-region input-out- put model can be effective in aiding rail planners in region rail decisions, as well as line-by-line issues revolving around subsidization or abandonment. Implications for the Use of the Method in Rail Planning When rail planners study a line segment for subsidy or abandonment, the procedure involves a detailed analysis of the costs and revenues associated with a given level of rail service. Usually, future rail traffic is estimated at various hypothetical levels and a cost-revenue analysis is conducted. The ability to project changes in tonnage and rail traffic by commodity is essential to accurately reflect costs that vary by tonnage and car type. To the extent that commodity composition changes as a result of different growth rates by commodity--due to the 126 input-output multipliers--financial projections will reflect changes in the revenue-cost relationship. In order to analyze the future profitability of a rail line, rail planners require an ability to project rail traffic. Not only are rail traffic projections necessary, but the projections must be commodity specific. Revenues are sensitive to the commodity composition on a rail line. To project revenues, planners need to distinguish between changes in high bulk-low value products from high value manufactured products. The input-output method can aid in projecting whether changes in economic activity will have a favorable or adverse effect on the composition of commodities on a rail line. Each sector in the input- output model has a unique impact on the economy. Large changes in spe- cific sectors could cause a change in the commodity composition of a line. Costs are also sensitive to the commodity composition on lines, as well as the level of traffic. The equipment type used on a rail line is determined by the commodity composition. As equipment types change, so do the costs associated with individual lines. The change in tonnage shipped over a rail line affects the required frequency of service and the required locomotive power; higher tonnage requires more frequent trips or more locomotive power. Further, maintenance costs on rail lines vary with tonnage; the higher the tonnage, the more wear on the track. Thus, the ability to project the level and composition of rail traffic can allow more accurate estimates of rail costs and revenues. It is recommended that the forecasting method presented be opera- tionalized in the rail planning process by linking the procedure to the analysis of costs and revenues. The input-output method can forecast l27 rail traffic by commodity and those forecasts can be used to generate cost and revenue estimates based on traffic levels and composition. In many states, the estimates of costs and revenues are arrived at by using various costs and revenues depending on traffic levels and compositions "by hand." This is an extremely time-consuming process that depends on hypothetical levels of traffic. It is recommended that the procedure be operationalized through the linking of the input-output program to a computer algorithm which generates cost-revenue estimates. With this capability, rail planners could analyze any rail line or rail segment using projections of economic activity. They could also change any single commodity projection, or set of commodities, as economic develop- ments take place. This ability would both streamline the rail planning process and make subsidy-abandonment decisions more accurate. Policy decisions could be made with more precision and over a much wider range of economic scenarios than is presently possible. Implications for Further Research In order to use the forecasting method in future years, the follow- ing steps are required: (I) The Michigan input-output transactions table should be updated to 1980 using the forthcoming U.S. and Michigan census data. (2) Efforts should commence to collect as much Michigan specific data as possible. It is suggested that the transactions table itself be based on primary industry data where available. The data gathering capabilities of various state agencies could be harnessed and l28 coordinated to build an input-output table based on more Michigan specific data than is presently the case. (3) In order to project total output as accurately as possible, better forecasts of final demand are required. Research that establishes procedures for collecting final demand data is required. (4) The test of the forecasting method was conducted on only ' one year, l980. Research should continue to test the forecasting procedure over subsequent years as trans- portation and economic data become available. In general, further research is recommended to test the capabili- ties and flexibility of the input-output method; data is currently being collected, for example, from forestry firms in Michigan. This informa- tion can be added to the transactions table to produce a more accurate table. The input-output method can also be expanded to make projections on income and employment changes due to changes in final demands. The input-output method can be further used to simulate changes in the loca- tion of economic activity by reflecting this in estimates of regional final demands. Research has begun on the use of quadratic programming to make the input-output model sensitive to price changes (Harrington, 1973). This research could aid in the use of the method for rail traffic forecast- ing, given changes in the price of energy and the cost of transportation services. Also, the method should be expanded to include consideration of multi-network commodity flows. One of the aims of the rail rationalization process is to suggest various configurations of rail lines best suited to the most efficient 129 use of the rail system. In doing this, it would be useful to have a method to estimate the optimal flow of goods on the state's rail net- work. Research is warranted to link the estimates of rail traffic pro- duced by the input-output model to an optimization program similar to the linear programming process. Finally, the most obvious deficiency of this study has been the lack of data on transportation services other than rail, particularly truck transportation. Not only should research on the trucking industry be conducted, but consideration of the factors that influence the split between modes is necessary. A method to include modal split estimates in the input-output procedure would be useful. Given the data limita- tions at this time, it is believed that the rail traffic forecasting method based on the use of an input-output model does provide useful estimates of rail demand and can aid in the rail planning process. APPENDIX I 1976, ZO-SECTOR INPUT-OUTPUT MODEL OF MICHIGAN WITH AN EXAMPLE OF ITS USE EXAMPLE OF THE USE OF THE ZO-SECTOR INPUT-OUTPUT MODEL Beginning with the transactions matrix, the intersection of Row 1 (livestock and products) with Column 2 (other agricultural products) in Table A 1.1, shows that the agricultural products sector makes pur- chases of $18,059,000 from the livestock sector, while the livestock sector makes sales of $18,059,000 of output to the agricultural pro- ducts sector. The table is divided into the processing sectors, 1-20 and the payments and final demand sectors, 21-23. Row/Column 24 represents the total sales and total purchases in the economy. The technical coefficients are found by dividing each column and entry for the processing sectors by the corresponding column total for the sector. For example, total sales in livestock is $671,181,000 (Column 1, Row 24, Table A 1.1). Purchases from agricultural products are $108,562,000 (Column 1, Row 2, Table A 1.1). The direct require- ments for livestock and agricultural products is .16174773 (Column 1, Row, Table A 1.2, 108,562,000/67l,181,000). Each element in the coef- ficient matrix indicates the dollars of inputs required from each sel- ling sector (horizontal) in order to produce one additional dollar of output in the purchasing sector (vertical). Thus, one dollar of live- stock output requires $0.16 of output from agricultural products. The Leontief matrix of Table A 1.3 is found by subtracting the co- efficients matrix for the processing sector (A) from an identity 130 131 matrix of the order (I-A). Inverting the Leontief matrix, produces the inverse matrix, Table A 1.4. The inverse matrix is read by columns. In the first column of Table A 1.4, it is indicated that livestock must increase its output by a total of $1.32 (Column 1, Row 1, Table A 1.4) in order to sell $1.00 of output to final demand. Other agricultural products must increase its output by $0.20 (Column 1, Row 2, Table A 1.4) for livestock to deliver one additional dollar of output to final demand. All sectors must produce a combined output of $1.69 (Column 1, Row 21, Table A 1.4), the sum of the first column. It can then be said that $1.69 of economic activity is generated by an additional $1.00 of out- put to final demand by the livestock and products sector. The sum of each sector in the inverse matrix is thus an output multiplier. These column sums, or output multipliers (Row 21, Table A 1.4), show the direct, indirect, and induced economic activity that will be generated by the economic system as a whole for each sector to meet an increase of one dollar in final demand. 132 .0000.vv. .0000000 .500700. .0030500 .vmw0 .0.Nh0 ..wm.fi. ..00077 .nm0m0w .0.00.N .00000 .000200 .070N0 .0.00hv .m00nmhfl .NNO.N. .0900. .50000. ..0h.0n .mnh .mmmON. .0h0000 .0 .0 044.01 0m<2~¢m 0. .0 ..0v .0000NON .00000h .0hmh0n .000500 .000n .000Nh .n.700 .90he0 .0050? .nmnhv .05.. .00CN .hnmm ..m00n .0007? .00000. ..000 .0070h. ...005 .nw..0 .w0mm. 00500 ><40 00(40 pm (04 030 0 .00?N.0 .00h00. .numq0v .00.00 .006 ..00.0 .00h.N .0000 ..0hh. .0thv .0 .... .h0m .0000 .0900 .00hn .nwmvn .veOnN .0.~0 .N00 .0000. .00000 .0 .0 >zuZ—Lwc .0000000 .000000. .70me.v .00.0a0 .0000 .h00NNN .00.?0 .00000 .0thm .0.00? .0? ..00. .000. .0NOQN ..0m.0 .N070e ..10.9 .905000 .90.05 .000.0 .0vv0N .90h0fl .00e0 .0Nv muuhm0h. .000N0v .hw0~.v .00050N .hm0 .00N00 .0Nh00 .0000n .0005? .e.0v~ .00. .0em. .Qnuv .00000 .0n00c .00n0n .hmp0 .01000 .0Nn0 .0 .50000 .nw0h0. .0 .0 02.2.2 0 .00.000 .0000. .hhe0.. .n00000 .00 .70h0m .000.0 .50.0N u0nhv .000h .09. .00. .05? .0000 .000N .hn0m .0000 .n.m00 .0.0m .N00 .0005 .090N .?.000 N0000— mpuacoaa o< awzpc N .— ..m_.no wm .0.000 mm .maemm. an ..noee. .m .0 On .¢C_0 a. .0n20 0. .n.90m n. .oa. a. .unmm a. q. .m:. n. ..m. m. .vhm .. .mxn. 0. .27h a .5mm. m .enm. 5 ..mm 0 .nmnow m .«m.w a .0m n .ncmmo. a .Omesw . .mooac a zuopmu>_4 . Jou\3:a 132 .0000.?v. .0000000 .500000. .0050500 .7000 .0.050 ..70.7. ..0007w .000000 .0.00.0 .00000 .000000 .07000 .0.0051 .0000050 .000.0. .0..00. .50000. ..05.00 .005 .00000. .050000 .0 .0 03(502 0(0 0 >042.0& 0. .0000000 .000005 .050500 .000500 .0000 .00005 .0.?00 .20590 .0050? .0005? .05.. .0000 .5000 ..0000 .00090 .00000. ..000 .00705. ...005 . 09.0 .v000. .00500 .0 ..0? 5(30 00(30 50 (03 030 0 .0070.0 .00500. .000v0c .00.00 .00v ..00.0 .005.0 .0000 ..055. .0050? .0 .... .500 .0000 .0900 .0050 .00090 .vw000 .0.00 .000 .0000. .90000 .0 .0 5002.000 130300500 0. .0000000 .000000. .700¢.v .00.000 .0000 .500000 .00.v0 .00000 .00500 .0.00w .09 ..00. .000. .00070 ..00.0 .0070? ..10.v .Q05000 .e0.05 .000.0 .0vv00 .90500 .0000 .00v 00.50<30 00300 8020 5 .0005509 .0000.5v .000.00. .00.050. .79.000 .00000.. .000000. .000000 ...00 .0570 .0050.. .0000?. .50.00 .005.0 .50000. .00000. .0v.0v .00000 ..5050 ..0090 .909 .0 .v.0 .0.. ..50. .03. .00..0 .050.. .00000. .000... .00000 .05000 .0000 .0030 .v0000 .00000 .000000 ..0500. .0500 .0.5000 .v0000 .0000. .009? .055 .00000 .00.0.0 .0 .00..00 52.00 Q<0 05030000 2030 0233 2.: 0 0000 0 0 .0000000 .000.00 .0005.5 .000000. .0.?0 .000.00 .05e0v .500000 .0050 .5e050 .0. .000 .000.5 .0.05.. .0.0005 .000000 .000.v .40000 .000090 .0 .500. .00050 .0000 .0 503050200 1 x.m5<3 020..0<02<05 Louommucm ... < 335 .509505. .00000v .5005.v .000500 .500 .00000 .00500 .00000 .0005v .9.070 .00. .0e0. .900? .00000 .0000? .00000 .5050 .07000 .0000 .0 .50000 . .0?050. .0 .0 02.2.3 0 .00.000 .0000. .55?0.. .000000 .00 .70500 .000.0 .50.00 v0.5.. .0005 .0?— .00. .05? .0000 .0000 .5000 .0000 .0.000 .0.00 .000 .0005 .0900 .9.000 00000. 05030000 0( 00:50 0 . .. ..0..50 10 .0.000 00 .00500. 00 ..0072. .0 ..0 00 .90.0 0. .0500 0. .0.900 5. ..00. 0. .0000 0. w. .0?. 0. ..0. 0. .350 .. .000. 0. .??5 0 .50.“. 0 .950. 5 ..00 0 .00000 0 .90.? 7 .00 0 .00000. 0 .00050 . .00000 0 $00500>.3 . 400\300 133 .000050. .000..00. .0w00000. .0900.05. .0.000.¢ .00.?0.5 .00?000 .00030050 ..0.005. .0007005 00 .menmm .moc¢nn. .momenmm .nmnamum .oaam.m. .acnofip. .oo.nm. .OOcmva .o0¢nqe .OCao..a em ..ouqo. .mmmcec .mmnmnw ..mnmom. .umOWmm .n...a0m ..OQQQ. .mno..mn .memo.n .ewnnmo an .waONQm .mo.n.mm ..mvm.m. .aeemmm. .mOnmam ..aemaa. .OOmmm. .ooammnm .OOmn.e .00mano« .m .333. .oemnm ..omeo ..nmee. .m~0mmu .557.» .amm .mm.na .aa.. .mman on .3a.mn .emeecu .em.ccv .anoneo. .m_swe .scccsn .mqmm. .nnmmo.. .mumhv .m~¢¢e. a. ..e.nn .smcman .omne.p .pmonaw .mmeau .qasam. .«eoo .aeaew. .opmc. .m.¢cp a. ..mep .nOn.qn .mmpmo. .COQnem .seam. .wm:?c. .monnm .nmommm .mVOne ..nnhh. 5. .5593» .emmmh. .00mme .Omna.n ..pmamh . .annc .ensn .opnmm. .vcnm .eeonn c. .q¢2m¢ .nnmpo. .oo~.o .mOan. .q..nq ..c.an~ .mOFw .vammn .mmnn. ..nspe a. .h@ .Omw .moon .opup .o .ccem. .m_.aq ..msn. .aocn .nqmem 2. ..mm .aohmo. .o..m .v.nm .mpw .cean .mpap .pOenssm ..ee. .mocmc n. .aom .womhm .p.mm .noamn .eFOe .neoa. .hnmq. .mmonoe .cmo.n. .Ommam. m. .meON .mhowo. .moann .mmmnm ..ehn .ummOn , .mmpnm .wnnn.n. .meaon .nmonmm _. .ao.. .nw.>m .oewe. .90o5m .ommm .Oanm. .om.0m .msmaenm .pmmma. .e.o.qm :. ..am. .mmmna .meao. .ehapo .uos.. .e.ca. ..op5 .xompna .mwemn .mm.~e : .usmn ..nnm. .en55. .poeue ..sen. .ozsmm .aam .mmmau .vm.. .«nsm x gho.a .mqnv.. .pmwen .Omhwn. ..One .mahp ..onw ..m.nm. .momm. ..msn. 5 .o.>n .memoow ..Omhv .am.m.n ..OQv .mnam. .mane. .so.me. .mnwmm .oome o .o .soemn .noom .ponme .am .mcem. .o .o .c .mmn m .smmmn, .nmmo.. .o-..m ..ao.om .hoa5n. .ssOnc. ..mmm ...moo. ..nnm .m..mm 7 ..m5. .o .nvmm .aOm . .emmom .n.nn .o .0 .me. .o a .un. .nmm.. .mOnmv .o.o_u .o .msme .o .o .o .o m .0 ..am« .ovan .mmou .o ..c. .o .o .o .o . um_aaau_zu mug—>zum uzoo ampumdum .u.a._.u ..uz.u4o:: mu.__4__: a .mz «ope: 4zuz_:u00520>Z. 5202500>2— 00 .ncmmonnw em ...namm . mm .manmmm.w an .mwennm .n .mnwm.n cm .uCCnmpm a. .no.nne> .m. .mmhmmnn. . n. .m.«mnh. o. .o.mo~m. m. .mobmm e. .m.q~mm. n. .somaom m. ..nmsm .. .m_m§n. c. .hnm5wm m .omwm_m a .mowmmon h .mwemmn. c ..m.m.nn m .o w .acmen n .Omomhm m .mnnn. . .zuaxw zamzou my; .0 400\300 135 oooocooo. smopme.n. nomwmon.. capanenn. Onommooo. e.mqomoo. moo.maoo. w~mo..no. efimmmm.o. cmes.m.o. nemom.oo. «Noncono. nomenccc. oom5.nno. coczaom.. amoeemoo. sesw..co. ammwen.o. mmmm.o.o. .momoooo. .weomnooo. «mmmceno. oooocooo. oooooooo. m.<.u: m¢<:.aa 0. 0000<3C>000O0<3C>00<3C>000" 00000000. .0 05000000. 00.00000. 0.577000. 07000700. 07505000. 07050000. .05070.0. 00000070. 00000000. 70.0.000. 00000000. 07000000. 55000000. 0000.700. .00.0500. 05000000. 000.0000. .0500000. 00.0.0.0. 0000050..0 .0 00000000.0 >002—000 130300500 00.000.. 00000000 00000<0<>C>0 0 00000000. 570.0050. 007007... 0050.77.. 05000000. .0500.00. 0.7007.0. 000.05.0. 05..07.0. 00000..0. 570.0000. 00000000. 0.000000. 00500000. 00500000. 500000.0. 00700..0. 0000070.. 000770.0. 00700000. 00700000u 07.00000. 00000000. 000..000. 00.50<30 00300 2020 5 00<3C>0CDC)0000" oooooooo. o.nqnnsn. ...o.0n.. amon.0q~. o-.o~oo. .amhpammo. unnmmm.o. v..m0mmo. mmnmemoo. sumeee.o. nmmmoooo. mun..oco. .mmonecoo. anomoeoc. cmemanwo. nommo..o. meumn.oo. n0m0o.~o. mammahm.. awnmm.oo. «apnoeoo. .mommcoo. momm.mco. ooooooco. .z_za g<¢ 2a:. m::. 2.x a coo. 0 x.0.<2 520—0—00000 00000000‘- 00000000. 05.0c500. 60050770. m.cmnen.. .ammacoo. 0con5300. 00.05000. 0.300000. 00570000. 00:305.o. OCCOOOOO. 00.06300. mcmnccoo. 05070000. 0:305000. n5mo5c.o. mmcmcooo. .507.5oo. 70500000. 73.07700. 7.900000. 70.0. coo. 50000070. am..5n... 05030000 0 Legummuom N.— < m.nm5 0000<3C>0<3C>0<3C>0<3C>0" 00000000. 0000500.. 0000000.. 0.050050. 070.7000. 0077.000. 0.000500. 00000.00. 05000000. ..0000.0. 00000000. 07000000. 700700.0. 5000.000. 70500.0.. 0570.000. 0000.500. 70000..0. 05500070. 00000000. 50..0000. 05007000. 005.7.00. 00000000. 0<00<3C30<3C>0<3C>0000" 503050200 7 00000000. 00000000. 00.70770. 05000.0.. 00070000. 00050000. .0007000. 007500.0. 00070000. 0.5007.0. 000..000. 05005000. 75057000. 00005700. 00500500. 500.0.00. 00500000. 000000.0. 50050700. 00000000. 070000.0. 0500.000. 00000000. 00000000. 00<3C>000000" 02.2.2 0 00000000. 0050.000. 0000000.. 5.050557. 00000000. 07000070. 00000070. 00000500. 70000000. 5000.0.0. 0. 000000. .0070000. 50000000. 00000000. 00.70000. 50000700. 05070000. 00.50000. 00070000. 50770.00. 750000.0. 50000000. 00000.00. 00000000. 05000000 0d 00:50 0 00<3C>0000<3C>0" 00000000. .0000500 .0..20570 007000.0. 00000000. 707500.0. 070..7.0. 000.7000. 05055000. 0000.0.0. 0 7 . 0.0000 005.0000. 00700000 70007000 00000.00. 0700..00. 0.000.00. 00.70000. 70055000. 00000000. 05000000. 05.7OCCC.0 .0 0070700..0 05575.0. 0H0C>00=DC>0C00" A v ”0607'? 1'6 ('4 (W (‘1 C6 @0000) 7. '-o;0-7w30r~0—0 ..00\30~_ 136 33333333. 00000073. 00737503. 00335070. 03700333. 30000003. 007.0.03. .3500033. 77000003. 00070070. 00003333. 30000333. 000.0333. 000.0.00. 59:03.33. 35030.33. .0000000. 00050533. 00037033. 33333333. 330375... 0030.333. 3000.333. 33333333. um.aamw.zu .>33 00 C>00<3C>0CDC>00000" 33333333. 50030.... .0700353. 07.50300. 00000033. 00330003. 03050003. 730.7003. 500.07.3. .amnon.3. 0.703333. 00300033. 0003.033. .Om.5033. .0000033. 50000030. 00033.33. 000.0033. 50.03303. 07000.33. 50005033. .33333333. 5.030333. .3070333. mmo.>uam 33530300 0. 043C>O<3C>000" RAAAAAAAA vuvvvvvvv 0 00000000. 00.00070. 00000000. 7000075.. 0707. 000 . 0.000570. 00000000. 0000.0.0. 50505700. 05000500. 7000.000. 00550000. 05070000. 0000.000. 000.7.00. 00000000. 57005.00. 07050000. 057.0700. 5..00000. 00700070. 70050000. 00005700. 00050000. .u.¢...u 0<3C>0<3C>0000<3C>0<3C>00000<3C>0<3C>0<>C>0‘? .3.3.3.3...3.3n3. .0070000. 3050350.. .7050700.. 00000733. 335mm..3. 00500033. 70000733. 0000000.. 500003.3. 33333333. 70033333. 00.00333. 37.30333. 0.000.33. 00707333. 00070033. 00003.33. 0.303. 33. 707.3333. 0.303003. 500300.3. 33333333. 33333333. 00.....53 0. x.a.<: .zw.u....ou 00<3C>0000000<0C>0<3C>00000" 2200 0 .0200<3C>00IDC>000" 33333333. 03m3305.. 3533003.. Ommnnm7.. .3333333. 00500003. 00307733. .n.~m7m3. 0000.733. 57mnmmoo. 07.50333. 005cancm. .57mmw.3. 00330373. 3090777.. 50000003. 00330333. 70337733. 50003733. 33333333. 03705033. 33333333. 33333333. 33333333. 0517a 32¢ 20> «353: 40<3C>0000000000000C€3CD' ..a. 5% Cr. .0 0. 0. 5. 0. 0. 7. m. 0. .. a: 0 7 7007035 >uuz.:3<2 .. .00x300 137 x—~:oo. nnn.ames. nmmqnnoo. acqmneeo. .ameo.mo. anemmooo. «ocemcoo. acmamp.o. p..nnmoo. wmms.o.o. mammnmmo. oue.5poo. “who“.co. venoooco. ood.pm.ou. an 0<000<3C>0C0<0C>00" 00000000. 20mmwem0. 00Nnvvnp. 00000000. 00000000. w—NNOOOO. 025mmepo. 01000w00. 00000000. v.h0th0. HOONm—CO. NwoohNON. mh.noh00. mam—.Nno. mmme¢000. memenm00. .NNQQOOO. ONOHmQHO. mOBumNNO. «memw—uo. mONhfimOO. mnvmmNmO. mhmm0000. @hnn0000. 0000' 00<0C30<3C>0<3 I . . >zopzu>2_ haoaxw pwz p2uzbmw>z_ mm 00000000. m—hmmcCO Nm000000. no.0Nm00. —®Onm700. QQ—NmOEO. mm.—nfi—. mwpwv_.n. m.mcnhm0‘ omOnmmmO. 06.0..00 .000.wm0. n®.mm.520. fifismn.00. 60.5..210. 00090900. mama-$00. mGNNCNCO samba—m0. m07c5m50. 00000000. m..m..m0€0. enmonwool hVO—mooc. oooocooodooooo— F V 000<3 9n mm mm _m 0m 0. c. ”NM-30053 .Zwaxu S3m200 .N awn 400x302 138 snowmooo.o- e.meowoo.o- sow.omoo.o- enac..no.o- ewmmmm.o.c- cowp_m.o.o- nommm.oo.o- «ponmomo.o- omnmnooo.o- wean.ono.o- oce_ooom.o «mmevmoo.o- schq..oo.o- nemeen.o.o- ammm.a.o.o- hmomcoco.o- wecwnmoo.o- wmmacemo.o- ococoooo.o oococooo.o m4<_u: mam a >a<:_aa 0. 000.0000. 000000n0 ..0fi.0.0 ##0000m0 0 o.- .0: n0000~.0. 0h900h00. .0: 0000b0.0. 00007000. 0h0~0000. .0000.00. 0h0v~0.0. 000.0h.0. .0 07507.00. 90000500. 0000.000. NNNQhGOC. n0h00000. 0000.0«0. 00000000. 0000.000. >440 mm<40 hm <04 00¢ 0| o: 0.: 0| 0| 0| 0| o... 0: 0: 0: 0: 0: 0: 0 0: N.FV?OCO. 07000900 N750h0~0 00500Nh0 00N0050. 00000000. 00000000. 0.: .0: .0: ~70h0000. .0h0w0.0. 00000090. 00000000. 00.0.000. 00000000. 0v000000. h50~0000. 0000.900. .0 nhnmwmno. 00N.0000. .0500000. 00.0.0.0. 0: 0- CI 0 0: 0: 0: 0: 0: 0:. CI 0| 0| .0: 0 0 >302—umm Samdoupmm nh000000. .0500.00. 0.900?.0. «00.05.0. 0h..0¢.0. 000nm..0. BQN.0000. 00000000. 0.000000. 00h00000. 00500000. h00000.0. 00700..0. 00.....0..0. 000770.0. 00700000. 00000000. 0Q.nn000. 00000000. 000..000. 0: 0: 0: 0: 0: 0: 0: 0: 0: 0: 0: 0: 0: 0 0: 0: 0: 0: 0: 0: mo.pm0... 5.00. 90500000. 00000000. v.000000. 70.0. 000 . 50000010. 00?.50... 9:. m puzcaca Z—x a 0000 0 x.¢p<3 uw—.2004 Louumm-om m.. < m.2mh 000.2000. .0: .0: 00000.00. 00?7.000 0.000h00 «5000000 ..0000.0. 00N00000. 02000000. .0: 5000.000. .0: wa0vnw.0 70500.0. 0500.000. 0000.500. 70000..0. 05500070. 00000000. 0h000000. .0: 05000000 00h.v.00. 00000000. 0.: 0| .OI 0|. 0| 0| 0| 0: 0: 0: 0: 0 0 0: 0 h030h0200 v «0090000. .0: .0007000. 00w~00.0. 00090000. .0: 000..000. 0h0N5000. ththOO. .0: 00050000 0.5007.0 0000h700 00500500. .0: 500.0.NO 00h00000. «00000.0. 50050000. .0 .0: 00000000 090000.0 00000000. 00000000. .0 00000000 0: CI 0| 0| Cl 0.: o: 0.: O! 0.: 0| 0 0 02—2.: 0 00000000. 02000N00. .0: .0: #0000000. $000.0.0. 0.000000. .0090000. #0000000. 00000000. 00.70000. .0: .0: 000N00w0 00000500 #0000700 0h070000 00.50000. 00070000. .0: .0: 50200.00 .Qh0000.0 hN000000. .0 000000m0. 90070000 0.: 0:. C: 0:. OJ 0:. o: 0.: 0| 0| 3| 0:. O! mpuaacaa c4 zu:_c « .0..A0.wnvnuavn.. ..n...~.0.m.0 010..9.0 00m.70n0. uhwhhmco. 0000.n.0. 07.03300. nm>.~:guc. 00.0.50: ammo-0:0 000N0.00, 0900..e:0. .0 «.0m0.00 00.9:mC; vm055330. 00:00:30. mhnnmscc. N h. .0000 nn>e~.0.. 00000000 . C: o: 0 0 o- O .0 0 0 0 C 0.” 0. : . 0. I 0. a. : m. : .— : 0. - C : 0 - o - a .0..—can. 0 x00.mw>.4 000u301 139 70000000 000075.. .0 00000000. 007.0.00. .0500000. 77000000. 000 70070 . 00000000. 00000000. 0N0.0000. 000.0.00. 50000.00. 05000.00. .0000000. 00050500. 00007000. 00000000 . 0: 0: 0: 0: 0: 0: 0: 0: 0: 0: 0: 0: 0: 0: 0 .0: 0000.000. 0000.000. 00000000. 0: 0: 0 00.0:00.20 5>00 00 00000000. .0 NOOhNflflO. 900.70NO. 500.07.0. .00000.0. 0.700000. 00000000. MWN0.000. .00.50CO. .0000300. 50000030. URL—.56. QC . fl:$.mGCO. 50.00000. 07L00.00. 50005000. 00000000. 5.000000. .GDQNOOC. ..000700 OI OI 0| 0| 0: 0| CI 0| 0| 0| 0.. 0.. 0: 0A 0:. 0: 0 0: 0: 000.>000 00500400 0. neme.coo. «.mnopeo. .o onmm.o.o. pm~aheoo. ohmamhoo. .o- mopsmooo. osmemooo. coao.noo. omo.e.oo. wooamooo. Feces.oo. nemhnnoo. whe.weoo. ~..ocooo. noeonmeo. emupmooo. mnempeoo. mmmsnnoo. 70.7..00 7000.000 uw-«~._.‘ 0. 0:. 0| 0| 0| 0| 0: 0: 0: 0: 0: 0: 0: 0: 0: 0: 0: 0: 0-. encaseoc. acmumamo. amnoaeno. enmhomwm. mmmOwn.o. semoomoo. nooOeooo. a.e.mcoo. mqmno.oo. opsaanoo. ofimeqmoo. osmomeoo. .mopnaoo. ma.no~oo. «mo>.n.o. .amnmmoo. .nemnnoo. mmcnoooo. macaw.oo. .0. 0000.000 02¢! .00.! .500.04023 5. 00000700. 00500..0. 00500000. 70000700. .07577.0. 500000.0. 00000000. 70000000. 00.00000. 07.00000. .0.000.00. 00707000. 00070000. 00000.00. 07000.00. 707.0000. 0.000000. 500000.0. 00000000. 00000000. mu...4.53 0. 05700000. 0000.000 .00700.0 500077.0 .0550000. .0 0007.000. 0.075000. .0: 07..0000. 0:0050.0. .0: 00005000 .0000000 00500000 00500500. .0: .0: 00000.00 m.....mm. .uavau 00005500. 00500050. .0.00000. 00000000. 0.000000. 2100 CI .0: .0: .0: 0| O! 0:. 0| 0.: o... 0| 0| 0| 0:. o: 0 .0..—<05 0. x_m5<2 00..2004 gouommuom 50000000.0: .0000000.0: 0.070000.0:.00500000.0: 007.0000.0:.00007700.0: 00000500.0: .0.00700.0: 00707700.01 0000.700.0: 00.00000.0: 57000000.0 0N070070.0 07.50000.0: .00.7000.0: 050..005.0 000005.0.0: .57000.0.0: 005.0700 0: 00000070.0: 5050050..0: 0000777..0: 0.070000.0: 500N0000.0: 7000..00.0L 00000000.0: 70005000.0: 70007700.0- 000005.0.0J 50000700.0: 00000000.0 00000000.0 05500000.0: 00705000.04 00000000.0 00000000.0 00000000.0 00000000.0. 00000000.0..00000000.0 000 02(05 050.00502 7. 0. 00.05000. .0: .0: 00570500 007000.0 00000000. 00007000. : 00000500 00.05.00 00050000 .7005000 00050000 00000000 00000000 N. 0.: Cl 0:. .0: .0: 00000700. .0 00000.00. 757.7000. .0: .0: 00000000. 700.00.0. .0 .0.0.000. 0.000000. 00000000. .0 . n. . 00 0 4(0—050040 0| 0: O... 0.: o: 0: 0: 0 00000000 0: 00 000.70.0..- G. 0.050000.0: 0. 00000000.0: 5. .0000700.0- 0. 00050000.0: 0. 0005.000.0: 7. 00.00000.0- 0. 0005NCNC.0: N. .0007050.0 .. 000000.. .0- 0. 7..0...a...: 0 0700..00.0: 0 05005.00.0. 5 07005000.0- 0 00.70000.0- 0 000.0000.0: 7 00000000.0 0 00000000.0 0 00000000 0 . >002.IU¢<2—mm 0— 000'<3C>000<40 mm<40 hm (ma 033 m 013€30<3C30000auZ—uua 000<3C>0<3C>0<3C>0<3" nonONmmm. QQNMOGOO. ch.mhc0. Q—WVOFNO. ON—NCOGO. hn0_aom0. ennuma_0. .vvnmCOO. m.hmm000. 0n0n>.00. n_hvcn¢0. woowm-70. whom00m0. onOOm.0. cvmanm—. anvnmhno. VQNQHQCO. «Nov—NpO. vmvflee—O. moemnnoo. 0N007p00. m0_hm0000<3C>0" QOcmh.mm. hameQOO. mmm~wee0. nncmfinHO. vwmwmweo. e.9m.m.0. anvaomwc. @0660000. wow—mmoo. 960c0_00. mmwvwopo. O—Owcn70. cmwmhh.0. nmaemwoo. owomnmno. 0N5honmn. .vmennoo. QQO.@0—0. GOG.0¢00. .NO—QNOO. #9595000. hzuaa Q0<3"000002— Legumm-o~ «.9 < mpnmh 0<300<3C>0<3C>0<30<3C>0" 03000030. vmh.0n00. m—Nmn.s0. mmONQCNO. mamm9350. .s050n¢0. vpvehwm0. «$900000. QQBQOQOO. .hOmmm—C. h.hhne00. h.hO—Gn—. .hm—vmho. mamnmoOO. mO—Oosno. n—wwnhoo. .7-e.—00. .unehmco. 000000.0. OOOOHNOO. Nnhnn000. hunapmzou v 0¢000<3C>000000000<3C>0<3CJOCDOCDCDO" n.m~n:mc 900mm_00. moan—hno mhm0$m30. m.;.c~;uuqu. mcontxvc. WOWQDmNO‘ $035.030. hcwc.;gyu. om>c..€0. whom-000. wa¢3¢m.0, 00.50500. Conn; 030. mam.m030. m.mnh._0. Q?QF?Cm.. fiOfi~—.H.O. —®fl7mficc‘ enmeéaap. vecnsmn—.. . mv_a..~.n. A“ x00_mu>—g dCU.\332 . C(3220" C30 C:DC)0mmmo. mmnmmnwo. en.mmoco. mncmamoo. snmnwnoo. wsw.mnoo. mnemmwmo. «hummm.o. mammaeoo. pawn—1.0. wmecha.o. .mmnmoco. vwmscmu.. schemeoo. heemhcoo. ammunooo. wm_zaawp2w . _>cu 0N 000<3C>0<3C>000"" mnthOmn. mmvnnmoo. 0mN_0—00. .mon.070. memOmno. Ofincnnm0. hm.w?m.o. OOVQNOOO. h-whmn.0. fiQ-OBROO. on.hmn.0. cc.mhm_0. m—h00—.0. Cum—hNOO. cwnméw.o. m..mm0~0. 000¢5w00. «hmomm.0. mmo.m.00. QOQQONOO. HQNNQOCO. mw0_>aum 0wp0wamm m— 000<3C>0000<3C>000000<3C>0<3C>00’(DCDO" Q—mnha.m. www.m0c0. n.605mm0. 7.thm_0. hmmwen.0. cm.a0>nw. mpnmnm.0. «Om-.000. QmNm—.00. OnnhmnOO. .mOmOmOO. flammmfi—O. bmmm0000. «00.nm00. hm09n900. cmmowhoo. 00m0w000. .vnOcan. «wannhmO. on.0n000. emm0_000. mw—p_4—bD @— C>C0000<3"02— cam-WNHQN. .mhm.N—0. unmeech. rathmmO. c0>mh.m0. nmm.am.0. mh.mnm60. .nhcww00. 0nm00000. .QmwanOO. maenh000. w.mmmmw0. Oah.m000. thcecOO. hnm00w00. enmn.O—0. nammnnOO. 70.606~0. OOOmmNOO. nw.:«.00. 03000000. 00<3C>0<3C>000000"C>00<3" .ummem.o. .nwmeaoc. w.mmm>mo. omeomm.o. mo.mm.mo. oo—mch.o. cemmemmo. v.n_m.oo. «mm.oohn. .hhnmmno. «mo—mace. .mmFOmww. weennmao. hhmh.moo. OO..-.0. OwnOQOmO. n.0awooo. CON—00.0. mmcchcoo. hymnecoo. mnmm.ooo. mpz mops: 4000000<300"C)O00" 7mm.hmom. uhOcnmco, .fifimm.m0, N00—?m_0. mém.ehn0. mQOmOH—O _m£chm—0. mOOcméco. awh—fiOnC. Omfihnhmo. «Ccmwmm.. wevOc.0—. HOOC—x—O 0¢hh.mCC. mflpmom00. manmnw-C. CommeCO 6090055XV .mzvcmoo. 0n0mn000. mnnm¢000. ¢ 0 C300O 00 —N CH 7— c. h. @— @— ”CV57? >3m2—IUmu 0000 0h0000¢a mpuaccma 20 0(0 .42 0 02.2.1 2.! 0.000005 .maoca 024 00000 x<0.02_<¢0 00>zwmwum >3—(0 .(02 .b00¢.0200 0.0% 00300 0 .<.03 .001 302.0 X00.mu>.a 0. 0 0 h 0 0 9 0 N . 400x30a x_z.<2 020..U<0Z<1— Legumm-¢¢ ..N < m.nch 143 .000000. .000000 .000000 .0000.0 .0... .0u009 ..9000. .0.000 .005 .0.00 .5000 .0099. .0000 .00.. .000. .500 .00 .0000 .0500 .000 .0 ..000. .0009 .000. .000 .0050 .0.0 .0050 90950 .000 ..09 .00.0.. .0.090 .000 .90500 .000 ..009 .00000 .0 .09 .0 ... .5000 .9009 .0 ..05 .0 ..09 002.404 0 000000 00 .0090.0 .00500. .000909 .00.00 .009 ..00.0 .005.0 .9050 .500. .00... .0.00 .00509 .000 .0 .... .500 .0 .0090 .00. .000 .0 .0090 .9 .0 .00 .050 .000 .99 .0000. .00000 .090. .0 .0 >002.mwa 800400.00 0. .mmccm. .ooapmn .oOnnnv .oou.a .wnm.¢ .m..vv .ooha.. .ooawo .ccm .ae. .o.~mm .uso.. .oonn. .mo.n .«n.«. .npu. .aon .ann .nmn .eum. ..aoa ..omu .oqov .qmmq .amo. ..m .o ... .o .o .em .mv .o .o .mmm .«n.« .qm. .unm .vm .em .o . .o .«eam .aoe ...p .on .o . .o .o .m. .o .o ..o. .mwo. .nm.n ..cu ..mom .vopv .«mv .mmm. .mnmnn .m .ohm .aqoo .om.vn ..vnao .aom ..p ..qso. .ummn. .o .o .0 .ma. .«o. .om. .na. .0 .n... .0 .men .0 .me« ....o .:o«. .o .ecnn .mmmu .o . .o .o .oa .na ... .ow. .o ..‘x .z>m 00000 0 00.b0<4t u.44<.8020 0. . 5. ..00.990 .000050. .000900 .000009 .090000 ..0050 .000000 .0000.9 .0000 .0005 .0.009. .00599 ..0000 .0005“ .00000 .5000. .0900 .000 .00000 ..50. ..0000 ..590 .550.0 .09.“. .0000 .0009 .90 .0. ..00. .0 .900 .000 .00 .n. .0055. .055 .000. .900 .500. . .00 .0 .0 ..0009 .0050 .0500 .99 .0 .99 .00 .0 .0000 ..09 ..090. .590. .00.0. .0 .9000. .0090 .00500 .000 .0900. .0. .590.9 ..90. .00.000 .00.0. .090. ..n.05 .00009 .00.09. .0 .00. .0000 .00. .000. .000. .90. .0 .0050 .0 .0000 .0 .0590 .0 .5500 .0 .50.9. .0.00 ..05 .0 .5500N .0 .9000 .0 .000 .0 .0000 0 02.20.4004 0. 0. .00.055. .009000 .00.000 .0000.0 .00.. .00099 .9990. .00009 .00.9 ...00. .0.00. .0..00 .905. .0 .0 .900 .99. .0990 ..009 ..00. .0 .0500. .05 .0 .0. .0000 .900. .0090 .5000. .0..9 .000 .000.. .00000 ..500. .000009 .55. .09005 .0005 .0 .0.00 .000 .0 .0 .0090. .0 .0999 .0 .0 .0086 00.4 .02..2.¢Q 49 0 00045 9. x.a.<4u.u 20pm.mm«4u .N 400\108 1115 “acm.w. ..mommu. .Oopanvu .oao0a .p.nmuo .mnmnnu .Oo.pn .onnmau .vpmmuu .u.mm .nwnoo. .omnpnc .oOeOm .ao.a« .asemo. .omo« ..~.m. .omqmn .OOm. .nnma. .uumn. ...v .o.«. .ammm. .om. .«n .cnnm .oo.v .nnpqu .oovmmu .nomm .«q. ...npwv .eun .qnnn .«.qan .o _ .o .o .o .o .o .o .0 .«~« .n .nm .a.0q .o .o .o .ppn .o .on.n .o .o .o .o .o .o. .o .o .ov. .hmu .o .mwo. .o .o .o .o .o .o .c .o .o. .0 .nm«« .woo« .o .o .nop .0 .0 .nm .nm .oo .npm. ma. .0.» .onmu. .o .o .o .o .0 .mn. .o.m .o .pvmn .o .om. .m». .m.. .omo .ammu .o .o .o. .o .o .00 ..o .wnn .npv .o .o .o. .o .o .o. .o .o .00 .o .o .o .o .o .«. ..mnn .aa..v .puooo .o .qaoOo .o .o .o .oo .o .o .o .o .o .o m >¢<_.z2u .09800xu \004 024<> \090008. 0040200302 00920 9>00 000 .00400 9v 0' 0? .000090. .00000 ..00v0. .000000 .000. .00.00 ..0.00 .0000 ..000 ..09?. .00000 .9000? .9.0 .90 ..00 .000 .09 .000 .01 .0.0 ..00 .900 .000 .0. .0. .0 .000 .00v .000. .0900 .00. .0 .0909 .0070 .09. .00. .0 x.09 4.¢_u¢ a .u.u._.m u4w0 l<0.02.<¢0 oooooooooooooooooooooooooooooooocoooo00006006664 00000000.. 000e0v00.0 500.0v00.0 005.?000.0 . 00050000.0 00000000.0 v00.v000.0 500.0500.0 5.000000.0 000ev000.0 05000000.0 00000000.0 .00000000.0 00000000 0 «smaq.oo.o on..oooo.o mphuoooo.o oooooooo.o o.omwpno.o oooooooo.o oooooooo.o oooooooo.o oooooooo.o oooooooo.o mo.qmm.o.o nq.ommoo.o moqnqooo.o .ohnoooo.o pomsmooo.o .nmnowoo.o «anu.ooo.o manmppao.o oooooooo.o onomuooo.o «ohmnooo.o .ov.unoo.o ou.on0mo.o «omvnOno.o usnmu.oo.o. «msmcnno.o «.numnoo.o oooooooo.o 0005 00>¢00u¢a 00000000. 05500000. 05000500. 00000000. 0.000000. 000v00.0. 00050500. 050000.0. .0.00000. 00000000 05005000. 07000050. 0000.000. 00000000. 0.0..000 00.00000 00000000. 00000.00. 000.0000. 50000000. 00000000. 05.05000. 00000000. 00000000. 00000000 00'00000 v0v5v000. 01.10000. 00000.00. 00000000. 00000.00. v.00..00. 005.0.00. 00000.00. 00000000. 00000000. 00000000. 000...00. 0..0.0.0. .0v..000. 00'00000. 000.5000. 00050000. 00000000. 00.00000. 0.001.00. 005.0000. 05000030 50—10 5 ooooooooooooooooooooodddooooooooddoooodoooooooo— 00000000.. 00550000.0 .0500v050.0 000..050.0 0.0??000.0 0000?0.0.0 5..00700.0 000v05.0.0 00.50000.0 50v00000.0 09005000.0 000.0000.0 0000.000.0 00000000.0 00000000.0 00.90000.0 00500000.0 00000000.0 05000000.0 00000000.0 Q00.0500.0 .000v000.0 00000000.0 00000000.0 .0.00000.0 00509000.0 0.0000.0.0 00000000.0 c0000000.0 00000000.0 00000000.0 50090000.0 000..000.0 0.0?0000.0 0.000000.0 09000000.0 00000000.0 5500.000.0 00000000.0 v0000000.0 050000aa 5(08 x—¢5<8 .20.0—00000 oocooooo. mmmnpam.. «noown~.. 0.050050. mam.eooo. mnev.0mo. m.nno.oo. .m.woomo. peoo.ooo. mommoooo. 5.000000. ..mnom.o. somanmoo. .ommooooo. memnoooo. camcn«.o. oooooooo. emameu.o. mmn.oooo. ..menooo. .omnn_oo. mommmn.o. ammo.ooo. «.memooo. omo.w.oo. onu~on.0. .osooauo. m.m.suwo. «mm.emoo. mnao.hoo. 00000000. scoooooo. comma..o. oep.oooo. commmaoc. amoosmoo. whopmnno. ..mno.oo. 00000000 00000000 00000000. 00000000. 00000000. 50..0000. 00000000. 05000000. 005.9.00. 00000000. .500050200 Chum 00:80 gouumm-¢e <3C>0<3C>00<5C>000<3C>00000¢DC>0<3C>00<3C>‘ N.N < wpnmp 00000000.. .5000059.0 00.0010..0 0000005..0 0Q500000.0 000000.0.0 09000000.0 .0000500.0 00005000.0 0000v000.0 55500000.0 00005900.0 00050000.0 00000000.0 00000000.0 0090.000.0 00000000.0 50500000.0 00050000.0 00000000.0 90005000.0 05050000.0 00000000.0 00000000.0 00.05000.0 009.0000.0 00000000.0 0.500v00.0 000.0.00.0 .0050000.0 00000000.0 00000000.0 05000500.0 05500000.0 0 00000000.. 0500900..0 00000000.0 50000000.0 00005000.0 000.0090.0 0'500vq0.0 00v050.0.0 .0050.00.0 00.000.0.0 00000000.0 005000.0 0 0000.000.0 0050.000.0 00.0..00.0 05500.00.0 00000000.0 00.0.0.0.0 00000000.0 0000.0.0.0 .vq00000.0 05050000.0 00000000.0 0.0.5.00.0 50000000.0 0000.000.0 000?0000.0 00590000.0 00007000.0 .0000000.0 00000000.0 00000000.0 05000000.0 00.50000.0 00.00?00.0 00000000.0 005V0000.0 0000..00.0 001000.0 0 00000000.0 00050000.0 00000000.0 00000000.0 02—2.! 2—8 0 .1508 00000000. 0050.000. 0000000.. 5.050550. 00000000. 09000090. 00000010. 00e50000. 50090?00. 90050000. 000.0000. 5000.0.0. 00000000. 0.000000. .00v0000. 50000000. 00000000. 00w5.000. 00000000. 0.000000. ..00.000. 000.0000. 00000000. 00000000. 00000000. 00000000. 00000000. 000.5000. 0.000000. 05090000. 00000000. 00000000. 00.50000. 000.0000. 5009v000. 00000000. 00000.00 00000000 00000000 905v0000. 00000000. 00v00000. 00000000. 050000.0. 00000000. 50000000.0 00000.00.0 00000000.0 050000¢m oooooodddooooooooooooooooooooooooooooooooooo- ~00¢ 00150 00000000. .0000500. .0900590. 00¢000.0. 00000000 901500.0. 090..¢.0. 00.00000. 0000.000. 00090000. .0.00000. 0000.0.0. 00000000. 09.00000. 005.0000. 00v00000. 00000000 0000.000. 00000000. 0.000.00. 00000000 00000000. 00000.00. 55.0.000. 00000000. 00000000. .9000000. 0000.000. 00050000 00.05000. 00000000 01.00000 0000.000 05000000 00000000 05.10000 05515.0. 0000000.. 06666dddodooooooodooddddddddddodooooooooooodooo- w 'cwn'3UIOI~¢:0<)-cn0-7u50r-0¢nC>'cn0-7ul0r-0 9"ssg:!gq£=:2532¢000“(VGINCVGIfltflfifl(flfiifltfifl’fliflfl”!!!1'11PQ'VQ "flfl'DUFOO .000¢a 02¢ $00500>ud . .00\)0¢ 148 00000000. 000.500N. 0900005.. meann.na awq.aooo 000500m0. 0«000000 50«900«0 0«.00000. ««9«0.00 0..0n000 nn0000.0 05000000 00.00000 0««000...0 50000000. 50«90000 0..n0«00 00000000 000.9000. 00000000. 009««0.0. n0000000. 0.500.00. 09909000. 0««00«00 03000000. 50000900. 900m00.0. .0009000. «90.2000. 0009n«00 500009m0. 5n099000. 000000.0. 00000000 . 00009000 05000««0. 00000000. 00000000. AXXXXXYXY 50000000. 09505000. 500«nn00. 00000000. 009.0000. 00000000. .n90«0¢0. 0025(04 0 000000 0« ooooooooooooooooooooooooooooooédoddddododdodooo- 00000000. 00.000... 05000000. 00.00060. «.519000. 09000900. «9505n«0. 90000000. 00009.00 0000.«.0. 000n0000. 000«0090. 0500n000. 00000000. 70.«.000 000««000. 00000000 0«9n5«00. 9000«000 09595000. 00000000. 00005000. 00900000. 00000000. 00000000. 0000«000. 90090.00. v«.5?«00. 55000.00 .0«.n5«0. 00«.0000. 0099.000. .05000«0. 00000000. 00050000. 00000000. ««090000. 00090000. 00000000. 00000000. 00000000. 00000000. ««090000. 0«.0.0.0. 0009000.. 0000.«00. 00000000. 00000000. >¢wz.mw¢ 800400—00 ooooooooooooooooooodooooooooodoEodooooodooooooo- 00000000. .o.m.m«0. «0090000. 0009009.. ..000000. 09¢¢5n00. 0.5000.0. .9em.m.0. 00000000. 00.00000. 00.00n00. 55000000. 5.0.9.00. 00000000. 00000000. .0500000. 00000000. 00n55000. «www.000. .0500000. 00000000. vmmowwoo. 50n00000. 00000000. 00000000. 00000000. 05005000. mnnmmmoo. 0.«5n500. 50000000. ..959N90. www.5000. mononmwo. 900n5000. 005nvn.0. 00000000. 00000000. «m5~.000. 5500.000. «0.9.500. anoonoco. ma00n000. .5un0.00. 00000900. 00000000. 00000000. 0n0..000. 0050«000. 00000 OCDO<3C30<1C>00<3C>O<3(50¢)C>0<3C)0<3C>O<3C>0<30O<3C>0<3CIO<3C>0¢3C>0" 00000000.. .00000v«.0 0000.0«..0 «0«V505..0 00009000.0, 0000.000.0 090.00.0.0 500000«0.0 090«0000.0 0«00«900.0 500.0500.0 ««005«.0.0 0Q«0.000.0 000«0000.0 00000000.0 0.000000.0 500n0.00.0 0000«000.0 00000000.0 000«n.00.0 05.00000.0 05000«00.0 «n000000.0 090.0«.0.0 00v00900.0 «0n.0000.0 9«0900.0.0 «00«050«.0 0000.000.0 0005«900.0 00000000.0 .5v.0000.0 0«n5.«00.0 00000000.0 00000000.0 00000000.0 vn00n«00.0 00000000.0 0.000000.0 00000000.0 .0050000.0 000«0000.0 00000000.0 .54! .250 0 00.50<.¢ u..4<.xw:u 5. 00000000. 00509500. 0900.«0.. 5v0.95n.. 0n««0000. 005.0000. 590.09.0. 900000.0. «0090.00. 05900000. 00n«0000. «0.00«.0. 5««n0000. 000.0000. 05«99000 38.?5 . 00000000. oaammfioo. meanmooo. enn.uooo. oooooooo. cweqqp.o. «paumooo. oooocooo. moonoooo. apoemnoo. canompoo. homunmoo. pfimsucoo. anhmnm.o. p.~mnnco. mmhmoh.o. qamhana.. oopomooo. anomn~.o. oooooooo. osmnaooo. moonwooo. p.5uoooo. cummnuco. anacouoo. omc.o.oo. ommmo.oo. a.mopmoo. «.poaooo. «moon..o. acocn.oo. «ppo.ooo. x.05<8 520.0.05000 OCDCDO(DCDO<0C>0<3C>0<3C>0<3<3C>0<3C>O<3C30O<3C>0O000<3CDO<0C>0<3C>00<3C>0<3C>0<3€>0<3C>0¢0C>0<3C>013C>0<3C>0<3C>00<3" 00000000. 0.00950«. 00090N9.. 0900.05.. 50000000. 505009N0. 500e00.0. 0«0n.9«0. «5000«00. 00050000. 0«nn5000. «050?.«0. 00000000. 00000000 00000000 00..0000. 50.00000. .v5«0000. 00900«00 «0005000. 00000000. 0fl.««000. .0090000. 00000000. 050.0000. 0n9«9.00. 00.05000. 00000.00. .0«00000. 950.0«00. 0090.000 09500000. ..0.0.00 50«05000 0900050« 00000000 .0v000n0 0009.900 00000000 ««000900 00000000 00000000 00000000 5«000500 00000000. .v.00«00. 00000000. 00000000 .0006 00.. .0:—52.00 .4 0 00040 v. doooddddddddddddodooooooooooodoooddoooooooooooo- 00000000. 0«99900«. 00000.«.. 0.0«500«. 05000000 «nn..0«0. 00050000. .5~0nn«0. .00..000 nn.e..cc. ouumunoo. 5005m0.0. u.mwmnoo. $2303. c.nmnooo. acqeo.oo. qamo.coo. apoquaoo. anamouoo. 09.09.00 00000000 0«n00«n0. n0.«5.00. «0000900. 9000«.00. 00500.00. 0v0mn900. 0.09.000. 9«m000«0. «0950000. 00000000. 50500000 .000n0.0 .000. 000 00050..0. «0000.«0 5000.000. 00.«.0.0. 00000000 00000000. 00000000 9095.000. 00000000 «.0nfln00. 00000000. 00000000. 00000000 00000000. 0005—2000 02¢ 000804 a. odooododocooéocOdooooooocooddoooooooooooooooooo- 00000000. 0955000«. 5009.00.. 6900.0n«. 00500000. 0000.0.0. 00000500. 90000.«0. 00.9.000. «0.00.00. .00.«000. 005900.0. .«5«0000. .000.000. 05..0000. 000«9000. 0N000000 90055«00. 00n00000. «55.0000. 00000000. 00990000. 5.000.00. 0«On«.00. «5990.00. 9n«00«00. 0.«00000 «n50n000. 00.00«00 .0.5n«00 00000000 0.«00000 9000.«.0 .000.000 000.0000 00000000 00.n0w0.. 00000000. 050.0000. 00000000. 00000000. 00000000. 00000000. .0«v0900. 00000000. ««500000. 00«00nvO. 00000000. coooooo000006666666dodooooooooodooooooooooooooo- 0000 000) 00000000.. as mnm.moe«.o sq .naoqnwn.o we n..0mm~n.¢ me umn.mooo.o wq ns~.cn.o.o nq cmomqpoo.o «a «acoon.o.o .9 o.w0qooo.o ca nowqo.oo.o an npnmo«oo.o on .omoohoo.o ha swnunnoo.o on mnmoooao.o on ac.moooo o qa qwmq.ooo.o an 00000000 0 an nmpmp.co.c .n n.oeaooo.o an «.omoooo 0 ma oooocooo.o on hpqemovo.o «a hemacooo 0 ma oooooooo.o mu «mooooco.o on on.ooooo.o an «3.2.00.0 an «nmnq.oo o .« ppnwenoo.o ca qpuanooo o a. oooooooo.o o. «.onmono o p. pmo~caoo o m. opnmnooo.o n. seawahco.o a. ecomuooo.o n. nnm.nooo.o u. cacom.n..o .. a«.ooooo.o a. am.emooo.o a encooooo.o o ooooooco.o p oooooooo.o o «pm5a_oo.o m oooooooo.o q moo.oooo.o n meqmmooo.o « «nnn~.oo.o . 4u¢<¢a< ¢ u...xu_ .. Jouxaoa 149 00000000. «0.0.000. 00005.50. 0.0.0000. 05050000. .77000.0. .00000500. 500077.0. 00000000. 05000000. .70.0000. 00.00700. 00000000. 05500000. 0.70.000. «00000.0. 00070000. 00.00000. 07000770. 00.05000. 00000000 007500.0. .5070.00. 00050700. 00000000. 70000000. 00000000. 00000000. 70070000 0005..00 00000000. .0.00000. 00007.00 70050000. 05000000. 05050000 00. 57000. .0570000. .YXXXXXXO 00:30:03. 00003000. 00336300. 00009.00. ..00. ..000. 39.5900 nXXxXzXXY 09 5300c . 00000000 20¢! 02 .3003.<.08 00 0<0C>0<0€D€>0<3€>013<>00<0C)0‘0CDO<0C)0<>C>0<0C>0<0C>00¢0C>0¢3C>0¢3C>0" 00000000.. 0500000..0 0000000..0 .57.005..0 00005000.0 ..70.000.0 000.00.0 0 55000000.0 07m00000.0 07005.00.0 00050000.0 50000000.0 77705000.0 ..050.00.0 0.00.0.0.0 0050...0.0 00000000.0 .0000050.0 000700.0 0 ..000000.0 00000770.0 0700.7.0.0 .0707000.0 50000500.0 .77.0000.0 00077050.0 07700500.0 00500700.0 0570.0.0.0 00.00.00.0 00000000.0 .XXXXXXXvAv 07000.00.0 0700.000.0 00000.00.0 00777000.0 00000.00.0 50000000.0 00000000.0 00000000.0 00000000.0 nXxXXXXXYn. 00000000.0 057.0000.0 00000000.0 2048 50200 0 8¢0<3€>0<3C>0<3CLOCDO<3CD00<3CD0<3C>0<3C>0€DC>0<0<>0¢3C>00<3€30 00000000.. 75050050.0 0050500..0 00.0050«.0 50007000.0 5.000.00.0 00000000.0 070.50.0.0 00700000.0 00005.00.0 0.005000.0 00.70000.0 50700000.0 070070.0.0 .0..0000.0 07050500.0 0.000000.0 00000..0.0 .0000000.0 00000000.0 00705000.0 00075000.0 77007000.0 ..0.0000.0 0700.000.0 .7.0000..0 00000770.0 0500.000.0 05050000.0 .5005000.0 00000000.0. 0000.000.0 0.5.00.0.0 50507700.0 00005000.0 0000..00.0 00750000.0 0080 .450! 0C>0<0C>00<30>0<0C)0<3C>00<3C>0<3C30<3C>0000" 00000000. 0700000.. 0.07055.. 0000500.. 75075000. 00000000. 00.00000. 00000000. 00077.00. 700000.0. 0.0.05.0. 00000000. 00.07.00. 05050000. 00000000. 50.00000. 0007.000, 000.50.0. 00000000. 070.0000. 00000000. 000000.0. 00000000. 00000000. 0005.000. 00000000. 00000000. 50007700. 00000000. 00000000. 00000000. 07..0000 «05000.0 00000000 00.05000 55000000. 0050.000 70000000 00000000 00000000 00000000 00000000 00000000 000000.0. 00000000 07070500. 00000000. 00000000. ooododddddddoddddoooooooooooooooooooooooooooooo- ..0<502<8 00 00000000. 0705.0.0. «500.70.. .5.75m00. 70000000. 55007000. 05000000. .00000.0. «0007000. 55055700. 00.00500. 50005000. .0000000u 70050000. 70000700. 5000.700. 00000000. 00570..0. .0..7500. 07700.00. 700.0000. 00000000. 00000000. .0055000. 00000.00. 00000000. 50000.0.. 000.0.00. 00700.00. 00000000. 00000000. 7.000000. 500.0000. 07000000. .0.0..00. 05000000. 00050.00. .0050000. 00000000. 50700000. 00000000. 00000000. 00000000. 00000000. 00000000 00750050. 00000000. 00000000. 0000 0(508 00 0<0C>CiO0<0C>0¢DC>0¢3<)0<3C>0<3C>0<3C>000<3C)0W00O' 00000000. 00000000. 00.000... 000705.0. 005.0000. 00500500. 005.07.0. 000000.0. 00007000. 005005.0. .0050..0. 70000000. 00070.00. .00.0000. 00000000. 00057.00. 00000000. 00700000. .0700000. 00070700. 00500000 .vuqnuoo. 0075.000 70.00000. 00005000 07500500. 5.770000. 05007000. 00000500. 50007000. .000.000. 0000.000 500.00.0. 50000000. 055.5000. 00050.00. 57000000. 50700.00. 00000000 700.0000 00000000 7707.000. 00000000 50.50000. 00000000 0000.070. 00000000 00000000 ddododoodooooooocoooooododooooooooooooooooooooo- -00700500 54.0.0 .00-002005 20pm.mm(.0 .0 400x23: 150 00000000. 0.m0—h7- 00900nwn. QNO—mhoo. wannowmn. 05.05v.0. mum—76.0. —mn00«00. 000npuoo. mat—Omno. «OmmmmHO. ncowww00. 00000000. 00000000. 00000000. m..~0000. 00000000. thQCNCO. 00000000. 00000000. 00000000. flhthNOO. 00000000. 00000000. 00000000. 00000000. 00000000. 00000000. 50119000. whooc000. 00000000. 0000. xx 0 . mammonOO. Omwfi0000. momm0—00. 00000000. 00000000 NNDGHOOO. 00000000. 00000000. 00000000. 00000000. 00000000. 9.w.n.m0. 00000000. 00000000. 00000000. 00000000. m >a¢-20<3C>00<3C50<30<>C>0<>0<>0(DC)0<3C>00C>00<3C>0<3C>0¢DC30<3C$' 00000000. 0009—.mm. .m000v0.. nqunwn—. chummo.0. «mu—unpo. HNRQOQOO. hpomoooo. .0v.0000. vamhm.0. m7000000. mop..«00. 00000000. 00000000. 00000000. «.000000. 00000000. .hhnv.00. 00000000. 00000000. «0090000. hunnqooo 00000000. 00000000. 00000000. mmmmoooo. 00vvm000. 00000000. 00000000. ohhw_000. 00000000. 00000000. 00000000. 00000000. 00000000. mem—mNO. 0.0Nm.m0. 00000000. 00000000. 00000000. mu— an oooooooooooooooooodoooooddddddddooooooooooooooo- 00000000 Ovmnommn .cwmvumo he.nnmh. F.9mmhh0 nuanqouo ONO—hmoo @0wnq000 0000.«00 0000050. hum-0.0. vm0@.0.0 ONHOODOO. 00000000 nm-—.000. wmhvw.00. 00000000. ..hoON—OO. 00000000. 00.n0000. 00000000. «mVFGOOO. 00000000 00000000. 0.?00000. vmnnmooo. 00—N0000. OhnNOOOO. mmvvwooo. 0.uo.m00. 00000000. hhmm0000. Nvmmvpoo. hhem.000. 0h00..00. 0.?00000. @mNNOOOO. mafia—000. 0—v00000. 0.900000. 00w.0000. 00000000. «@900000. «huawwnO. 00000000. 00wn0000. 00000000. 00000000. ooooooooooooooooooooooooodoooooooododdd666666664 muuh—Jup: on 00000000. «owunhvw. anchumou. .QQOOQwN. nhvnamOO. ammo—Ono. .NOVOQpO. huwflnn.0. a.mh¢000. ovmhm_00. noouwmoo. Oonvnnno. .omceooo. Ooav.a00. O—nVBOOO. .0o0mN00. 00000000. ..ohh_00. 0.0m0000. nah-0000. —nnmmooo. 0.05m.00. 00000000. .OF—0.00. 00000000. 00000000 ahmmhn—O. .awvmooo ovam.00. mohnnhoo. 00000000. 00000000. «0600.00. oom0h000. nmnnq¢00. 00000000. two—0000. mommhooo. mvmnoooo. waw0000. 00000000. ONpQONOO. ummmmooo. mmhmwnuo. @0N_0000. 0'000000. 00000000. wpnNOOOO. .8800 ha ‘I€3C>O<3C>0¢3C>0¢DC>0<3C>0<3C30(DC)0<3C$O<3C>0<3C>0<3C>OWDC>0¢D<3C>0¢3€>0<3C>0<3' 00000000. mhmhmomw. mnmmhnmc. QQNh0mmN. hwuhhooo. O—mwmmmo. wanemmoo. hooma.n0. Gama—000. mhflomooo. Owawmwoo. .Omhhwoo. b.0nmmn0. vvme0.00. nan0—a00. buoe09.0. OOOmmOOO. warp—900. 00.01n00. n~wmm000. 00000000. nap-.N—O. mmanhvoo. mouwvvoo. ma—Omooo ovnvun.0. ansnuuwo hvmwwvoo vehmmn.0. «00hm000. mhnnmpoo. Gnommwoo. vnwmw—po. .vcmopoo. vwouwwno. .0hhnw00. 0vv>N900. 0.02mm00. 00000000. FVN00000. Nana—000. 00000000. Nhhmhooo. anamwwoo. 00000000. ammmnooo. vvm0v000. 00000000 on x—¢h000<>0<3C>0<>0¢5CDO0<3C>0<>0<3C>0<3C>0<3C>00<>0<3' 00000000. Ohm7060-. ww.mwmnn. vuh.uvnw. hmmmwoco. 0.010nn0. mmwumeO. mac-«ONO. .an.m000. .00-0000. efion—noo. mn—NOQOO. c00—.000. NhnmnOmO. .mw—vmoo. mnOmnh-O. h0000000. mnohmwno. mambo.—0. n.020n00. aewwmv.0. vomnhfluo. unpuhmoo. —mh—m500 mewOnfluo. moonmnNO. mOmaOunO. vom0vn00. .nnnomoo. wnw0—.00. 00000000. 60000—00. 0.1m0.00. anew—000. 66.00.00. cnmwonoo. 700m.w.0. onmmv.00 mhhm0n00. 00000000. 00000000. 00000000. .0. 00000000. a-30u 2(2» .:po.puu< no oooooddddddooooooooooooodoooooooooooooooooooooo- gogumm-¢e 00000000.. 00000000.. 00000000.. 00000000.. 0' m0¢000hp.0 oO-00mm«.0 mnmhm00-.0 nh—DOO—n.0 h? 0h00000..0 «mama—O—.0 .mnvnvun.0 0&700m0—.0 0v Omannmr..0 own-nnvfl.0 cmth—On.0 «Hommmmu.0 m? .00a0000.0 00.0h000.0 mum00000.0 n_N59000.0 a? mmhmmmn0.0 OnhechN0.0 aw00.nn0.0 90.fimo.0.0 01 GOOMV900.0 mm9600.0.0 .ow«o~«0.0 Ono-0-0.0 NW hnwonw.0.0 @mfiewon0.0 000mhmn0.0 whammm.0.0 .q 00001000.0 06059000.0 m—nh0000.0 .000.000.0 07 .0N00.00.0 007w..00.0 m0m0v000.0 00n00.00.0 an ohnth00.0 Gannon00.0 «en—0.00.0 wmwwan00.0 on hvonmmoo.0 000m-00.0 .Nsovmoo.0 unnohm00.0 kn «Beam—00.0 www.0700.0 nooma.00.0 O..wen00 0 0n mv—hnooo.0 N0.0h—00.0 nm0m0000.0 0mmm0~00.0 mn nuhoonww.0 Onwmn900.0 N.0fla000.0 FOOmomOO 0 9n vhqwmn—0.0 .nhnnwho.0 N9090090.0 nhhnmon0.0 an va¢00000.0 thow.00.0 .mwfiw900.0 @0mh.000.0 «n O—mmhhu0.0 vmo.nm00.0 whaomm.0.0 OmOthw0.0 .n Nwmmom00.0 ounmv~00.0 new..0—0.0 mwonmn.0.0 On 0.pnvN00.0 ..nm0000.0 hhmm0000.0 flammm000.0 ON ¢.mm.h00.0 e00.nn00.0 00000000.0 men0900.0 ON 0006e.N0.0 h.—NN«.0.0 a-Oemm00.0 chommn—0.0 ha Ouvwnmv0.0 Oohnno.0.0 v0nnn¢00.0 .n—mm000.0 mu 000v0000.0 BOO—N500.0 h9000fl00.0 Oompmw00.0 mu ov—h0000.0 vamvn.00.0 mh000w00.0 mnom0..0.0 QN wooavnno.0 .NO.FNN0.0 .0.0.000.0 «mefifihno.0 nu mnmmohno.0 mooom—00.0 «ohhv000.0 .00—won0.0 an mmvmhmoo.0 .mnmo—p0.0 N9050000.0 n~monw°0.0 —N hnmwmn—0.0 0mmwocoo.0 m-hmOQ00.0 .hmnm000.0 0N om000000.0 00mh0000.0 omaan000.0 O-wnn—00.0 O— 00000000.0 00000000.0 00000000.0 00000000.0 up amnwn000.0 N0w60n00.0 hnm¢m000.0 achn0000.0 h— mauv0700.0 00000000.0 nn—0m000.0 o«m~.~00.0 @— hhfl?0000.0 065.0000.0 enmom0.0.0 Gem-000.0 m- m000.u00.0 «nae-m00.0 .Ncm0m00.0 —OONQN00.0 v. .N.N0—00.0 9Nm00700.0 oo~00000.0 9.0.m000.0 a. vohh5000.0 0vmmeO0.0 006v.000.0 nmnhm—O0.0 Np «vommwoo.0 0060m000.0 nmNh.000.0 wnfimv000.0 .— 00000000.0 00000000.0 00000000.0 00000000.0 0. 00000000.0 00000000.0 00000000.0 nmm00000.0 m 00000000.0 00000000.0 00000000.0 00000000.0 o 00000000.0 00000000.0 00000000.0 m0000000.0 h 00000000.0 00000000 0 00000000.0 «m—m0000.0 0 nmvuhN00.0 .0w«.n00.0 «eunun00.0 hQOfimN00.0 m 00000000.0 00000000.0 00000000.0 00000000.0 v 00000000.0 n.000000.0 00000000.0 00000000.0 n 00000000.0 00000000.0 00000000.0 00000000 0 « 00000000.0 00000000.0 00000000.0 00000000.0 . whats d m—Dcw 20¢! 6800‘ 20¢! .02— .:w> CObOI A40—Cp0waw 00¢.w0-u50 Zwu 0 Oman tn on an ‘0 400\30u 151 00000000.. 00000000.. 00000000.. 00000000.. 00000000.. 00000000.. 00000000.. 09 050.0.m..0 90mmweN0.0 m.huoa00.0 nmmmu090.0 h0fl00....0 m~.«cnqm.0 vwonmwmu.0 he Ohm00m00.0 00wnwen..0 ammmmmnn.0 0NQO?FO0.0 .ow~00~0.0 Omanwowo.0 sn90.w00.0 0w N00550Q0.0 00000000.0 hn.wuo00.0 nn¢0>nflm.0 09.9m0mm.0 v0~0mw~..0 Oanwmmww.0 m9 000...00.0 00000000.0 .mOnmv00.0 00e00000.0 cmnNnm00.0 «www.000.0 n.m~.m00 0 we 0.N@.0N0.0 @.«n0000.0a oe.un0m0.0 00mmmwno.0 GowOHmwO 0 «.0N0h?0.0 0.honw@0 0 av vuomo.00.0 095mm7.0.0 mm..nh...0 omq.w.N0.0 «Onhnnn0.0 @wommm00.0 bhhmmcn0.0 n1 Owhme00.0 @nnm0w00.0 b.0mw.m..0 flamen000.0 mncvn.w0.0 no.00m00.0 wmhnm9.0.0 .v 'n000000.0 00000000.0 whe00000.0 emn.mN00.0 flenvwn00.0 an.mw.00.0 known—00 0 09 .QNOmN00.0 00000000.0 000@.N.0.0 wnoshn.0.0 abumhn00.0 ocheh000.0 no.0nN00.0 an .ovnnn00.0 00000000.0 vvwh.m.0.0 vmhnmm70.0 nmm.n000 0 hwomn«00.0 nommn000.0 on Oh.mmhm0.0 9.5005N0.0 wQONmmN0.0 noNQN010.0 .mnmmn.0.0 0hNNQhO0.0 Umfi000.0.0 hm 00059.00.0 m.nmnn00.0 N0m00m00.0 00NON000.0 .mw.9000.0 .QN00000.0 Nflh.n.00.0 wn hN0.Nh00.0 600nm_00.0 wn.0..00.0 cmN00000.0 «.7m0000.0 7600.000.0 .on.9000.0 an flhn.mwev.0 «NOMBNON.0 .000.9N0.0 0muwm000.0 000mm600.0 mmhhoOO0.0 0.059000.0 fin amoenn00.0 ms.nc900.0 nw.mnw00.0 Gum—0000.0 nun0_n00.0 @hm1a000.0 mnwhn.00.0 nn 50005000.0 «wehm000.0: h5900000.0 m0050000.0 NO0Ce000.0 00Nen.00.0 0009~000 0 an ahvman00.0 Onum0000.0 @08m5000.0 n.000000.0 N0m0m700.0 NmOh..00.0 «0071.00.0 .n 0000.0«0.0 @wbh.m00.0 «Nan-000.0 n0nv0000.0 thmN000.0 DOQN.000.0 Gammn000.0 On vahm..0.0 900.00.0.0 ¢0600000.0 unmhm000.0 o.mmo~00.0 nwcnw000.0 N0m0w000.0 mu nomvmmoo.0 .hhvom00.0 om006000.0 hn000000.0 .0m7Q000.0 ohw«.000.0 ONNWN000.0 on 000.h¢.0.0 vuamhnoo.0 bah09000.0 N00hn000.0 00090w00.0 @00ma000.0 00v.e.00.0 bu 00m60600.0 vstom.0.0 00.m0.00.0 vmvmn000.0 00000000.0 thm0000.0. a.mnN000.0 mu nuo.hnoo.0 6.0mw.00.0 w.h02000.0 v0000000.0 vonwhN00.0 «vmw0000.0 00050000.0 mu 0.0.9000.0 0000n000.0 00¢9N000.0 .09.0000.0 00000000.0 0n5N0000.0 0m.N0.00.0 an 60h.0000.0 mah0v0.0.0 0n000000.0u hwv00000.0 00000000.0 mmhmn000.0 on.w.000.0 nu 00000000.0 v059m0.0.0 mm.00000.0| whoan000.0 010N0000.0 monon000.0 nvwhN000.0 «a ovumv000.0 00005000.0 0.059N00.0 «1050000.0 mn.NnN00.0 nvmhn000.0 movme.00.0 .n Oohnm000.0 n.09m.00.0 hmwm.000.0 annow.00.0 Dum00100.0 a..N0000.0 mommhn00.0 0n v00m0000.0 .nhov000.0u amen—000.0 .m«owu00.0 0m000.00.0 hv00h.00.0 OhmowN00.0 m. anmnvhoo.0 GOO—mn00.0 mhmenm00.0 0.0N.000.0 0.0m«000.0 v.N.0000.0 mnemn000.0 o. whom..00.0 n00..000.0 mesa0030.0 00000000.0 00000000.0 mmmm~000.0 vmm¢0000.0 h. hanmnm00.0 n.evm000.0 hhmhnhN0.0 mnwv9500.0 nhmwn000.0 0hh00.00.0 whowwn00.0 w. 05¢0.n00.0 «“0..000.0 @0050000.0 nno.nu00.0 wannwheo.o ..mwmn00.0 Q500>.00.0 m. GOnNm000.0 9000.m00.0 000nwm00.0 N05?..00.0 00.00N00.0 oom00.00.0 wmwommoo.0 1. unm.0n00.0 .an0v.00.0 m..na000.0 00000000.0 00000000.0 «hem0000.0 mnemN000.0 n. 00mo.000.0 navemn00.0 NNQQN.00.0 00000000.0 nah.0000.0 .Nmo.000.0 00nn~000.0 N. oum.v000.0 N00nvm00.0: OQthv.0.0 nO~mm000.0 n.0m0000.0 ..00u000.0 .mu00000.0 .. mov00000.0 omommm00.0 mwonmo.0.0 00000000.0 0hmm0000.0 nhonn000.0 «0006000.0 0. .mecm.0.0 00m.0900.0 000.0900.0 00000000.0 0.mnn000.0 N0000000.0 N0005000.0 m «nm«.000.0 mn.hvn00.0 nOvOmm00.0 00000000.0 00000000.0 9m000000.0 06009000.0 O .Oh.nN000.0 0009..00.0 hoav>000.0 00000000.0 0000m000.0 hoaa.000.0 unab0000.0 h anu0000.0 nv0«N000.0 hvwamm.0.0 00000000.0 OOVON.00.0 NhN00000.0 wmwh¢000.0 0 Oaamnmn0.0 0065vm0n.0 00000000.0 00¢0¢h...0 hounhm00.0 0090n0v0.0 unhnm000.0 n 00000000.0 .thum90.0 00000000.0 00000000.0 00000000.0 mmbhu000.0 00000000.0 v 00v.h500.0 v.h60500.0 @vam000.0 0N00.000.0 00000000.0. wmvmn000.0 50160000.0 n «VhON.00.0 mumm0000.0 qnomn«00.0 0nnu.000.0 h.m00000.0 00009100.0 600.9000.0 a Qv«00000.0 05600000.0 590.w000.0 00000000.0 .00?N000.0 0005N000.0 00000000.0 . hZMIhmw>2~ mp¢0axu mum—ca wmu.> a—<.ua d \00< w04<> \mbcona. m0402uw001 CUFZu h>00 cum homawm .u.¢...u uddmeOI) h? 0? a. Q. a? «Q .V 400\)0a .253. 5.323on Legumm-¢¢ 151 oooooooo. o.o.m.n.. ohmnomno. «oohhuao. aoo...oo. o.um_ouo. voono.oo. oupqmuoo. vnoooooo. ..«Omuoo. .ovunnoo. a..mmpmo. oaapc.oo. .«o.«poo. n.n.ooeq. nomennoo. poochooo. opeouuoo. coco.o~o. vmopm..o. noovmmoo. noo.pv.o. coonOnoo. «no.5noo. o.o.cooo. non.oooo. comooooo. nvunqooo. oopnmooo. cooomooo. onmnVFoo. whom..oo. panmnmoo. chao.aoo. nouamnoo. onm.oeoo. oomo.ooo. oom.vooo. aoqnoooo. .Ommem.o. unm«.ooo. a..nnooo. .vueanooo. «commune. oooooooo. oo...~oo. «qpou.oo. 'vuooooo. C>0¢DC}0¢3€>0<3C¥000013C>043C’0<3C30¢D02. \oo< u34<> 59 00000000. 90009900. 0000990.. 00000000. 00000000 0.000000. 095009.0. 00000900. 00000000. 00000000. 00000000. 9.500500. 0.000000. 00000.00. 00005000. 05.00500. 00950000. 00000000. 0055.000. 900.00.0. .5590000. 90005000. 950000.0. 5.000.00. 00000000. 005090.0. 905900.0. 00005000. 0.090.00. .0509000. 000.0000. 000..000. 0.990000. 000..000. 9000.000. .0009.00. 00990000. 00009000. 00000000. 000.0900. 00.59000. 0009..00. 09000000. 00059000. .0500090. 9.500500. 05000000. 05000000. o<3c>o<3c>o<3croo<3c>o<>o<3<>o<3c>o<3c>o<3c>o<>c>o<3c>o<3c>o<3c>o<3c$o<3c>— mpaoaxw \mpaoaa_ 09 x.c.00 Cum ..00.—00 99 0<3C>00O¢DC>0(DCDO<30¢3C3043C>0¢3C>0<3C>000¢DC>0¢DC>0¢DC>' 00000000. 50000.... .0900050. 09.50000. 00000000. 00000000. 00050000. 00090.00. 09090000. 05005000. 000.0000. .00000.0. .00.9000. 0.900000 00000000. 0000.000. 00009000. 00000900. 50000000 0.000000. .0090000. 00090000. 00000000. 90005000. 00000000. 00000000. 09000000. 00.00000. 00000900. 00000.00. 0.000000. 00000000. 05000000. 00000590. 00.00000. 00000000. 005.0000. 0.000000. 05000000. 0.000000. 00000000. 00000000. 00900.00. 50005000. 00000000. 00000000. 5.000000. .0090000. 000.) 09 0000(DCJOCDCDOCDOGDCDO<3C>0¢3C>0<3C>0¢3C>013¢;0<3C>0<3C>0¢DC>0<3C>’ 00000000. 00.00090. 00000000. 9000095.. 0909.000. 0.000590. 00000000. 00.00000. 00.00.00. 00595000. 59000000. 05000500. .9000000. 9000.000. 00550000. 05090000. 00090.00. N005..00. 0090.000. 00009000. 0500.000. 00000000. 05000000. 09090000. 00500000. 00500000. 00000000. 09050000. 0..00000. 59005.00. 9.0.0000. 00000000. 05500.00. ..000n00. 00000.00. 05900000. .000.000. ..000000. 05000000. 00000000. 90000000. 5000.000. 05000000. 00900090. 00550000. 00900000. 00005900. 00050000. .u.¢...u 09 0(3C’0<3C>0<3C>0<3C>00<300¢DC>0<30>0¢3C>0<3C>00(3C)0<3C>0<3C>0¢3C>' 00000000.. on vqonmom~.o .9 5nwo.nwo.o on oouwomqq.o me n.m~.moo.o «a c.5ounmo.o nq nsnmmcno.o «a umpnmw.o.o .q .aomm.oo.o 0v o¢.onnoo.o an nowwnooo.o on onqooo.o.o 5n nu~.n.oo.o on .on.eooo.o mo o.o5qooo.o an mnqsn.oc.o no acmemooo.o an «mqu.oo.o .n owmmnooo.o On «omoqooo.o an museuooo.o on 009.9.oo.o 5w m.mnmooo.o on woesoooo.o mu om.uo.oo.o an on.o.ooo.o nu neqpnooo.o au moone.oo.o .n mowmpuoo.o cu ohmwmnoo.o m. moemaooo.o o. enmwocoo.o .. onowonoo.o w. epoch—00.0 m. manommoo.o a. mowmnooo.o n. mannsooo.o «. .mnOoooo.o ._ «Owenooo.o o. «mmahooo.o m «noonooo.o a onusmooo.o . wnwswcoo.o o onhnmnoo.o a oooooooo.o o sneuoooo.o n non.qooo.o « oooooooo.o . 4.4.uu a w4w0 0. 000.0000. 000090«0. 00000000. 05000500. .«00.000. 0.000«00. 00000000. «00«0..0. 000«0000. 00000000. 00000000. 000.0000. 00000000. 50.00.00. ..900000. 055..000. 00000000. 05000900. 00000000. 00000000. 00000000. 00000000. 00000000. 000.0000. 00000.00 0««9«000. 0.550000. 00000000. 00000500. 00000.00. 0.0«05.0. 00000000. 00000000. .0900.00. 0000..00. 09050500. 000500.0. .«005000. 005«0.00. 05500000. 00000000. 0«009000. 9.000000. 00000000. OI OI On 00 0| 0| 00 0| On. On. 00 O! o- o- C>O<3C>O 0n 0- 0o 04 0 0n 0- 0- 0. 0 0- 0: 0 0| 0- 0o 0- 0 0n 0- 0 09000.500 190.02.900 «0059000. 00000000. 900.9000. .0- 5.000000. .0: .0- 00000000. 0«000000. 00000000. 00000000. .«0«0000. 00000000. «5««9.00. .0: .0: 00000000. 0.000500. 00000000. 00000000. 00000000. 00000000. 500.0500 00099«00 0500««00 0«.50000 05500000 00.900.0. 09.00000. 00909000. .0500000. 50090000. .0000000. «000.000 0«0055«0 90500000 .99.0000 0.000000 905000«0 «.0N0000. .0 00000000. .0- .0- 00000000 0000000. 00.00000 0| 0| 0| 0| 0| 0| 0| doooo 0| 0| 0| 0.. 0| on. .0: .0: 00000000. 0«00~000. .0- .0: 0«.05000. .0 05000.00. .0- 0 0: 0n 0- 0n 0 coo. ou>¢umucc 0.000000. 000900.0. 0005«500. 950000.0. .0.0«000. 0000««00. 05005500. 09000050. 00«0.000. 00000000. 0.0..000. 00.00000. 00000000. 0000«.00. 000.0000. 50000000. 00000000” 05.05000. 00000000. 9«0«0«00. 00000000. 00000000. 009«0000. 90959000. «9.90000. 00000.00. 000«0000. 00000.00. 9.09..00. 005.0.00. 00900.00. 00000000. 00000000. 000««000. «00...00. 0..0.0.0. .0- .09..000 900909.0. 000.5000. 00055000. 00000000. 00.00000. 0.0«9.00. 005.0000. OI OI 0. 0| 0| 0| 0| 0| 0| 00 0| O0 0| 0| 0| 0 0a 0: 0: 0a 0: 0o 0- 0| 0| 0- 0 0 0: 0a 0: 0 0. 0a 0 0. 0a 0: 0502000& 5 50.90 0.099000. 000«90.0. 5..00900. 0009«5.0. 00.50000. 50900«00. «9005000. «00.0000. 0000.000 00000000 00000000. 050.0000. 00000000. 00000«00. 00.90000. 00500000. 00000000. 05000000. 00000000. 00000000. 00000000 00000000 0000«000. 900.0500. .0009000 00000000 00000000. .0.00000. 00509000. 0.0000.0. 00000000. 90000000. 00000000. 00000000. 50090000. 090..000. 0.«90000. 500..900 09000«00. 00000000. 5500.000. «0900000. 90000000 05000000 5902 0- 0c 0- 0. 0a 0: 0: 0| .0: .0 0 0n 0 0- 0- 0- O 0: .0 0 0 .0: .0 0o 0- .0- .0 0 0| 0- 0| 0 0| 0: 0- o... 0. 0n .0 0n 0 0: 0a .0- mem.eooo. maee.0mo. m.eno.oo. .0.00000. .woa.ooo. mammoooo. ..Omnooo. ..maom.o. .omunuoo. omuocooo. aemnoooo. «waen~.o. oooooooo. «mnmq«.o. mmo.oooo. ..mwnwoo. .omnn.oo. mmmawn.o. 0000.000. «.mqmooo. ouo.m.oo. o.unon.o. .w.ooano. m.m..uwo. «om.omoo. mans..oo. oooooooo. oooooooo. qcnmm..o. cc..oooo. e9¢mmuco. nmow.moo. m.o.mnno. ..mno.oo. oooooooo. oooooooo. ocoooooo. oooooooo. oooooooo. n.omwmmo. oooooooo. n.mmvooo. mm..w.oo. oooooooo. 0| 0| 0| 0.. 0.. 0| 00 0.. OI OI o- OI OI 0| 0| Cl 00 0| 0| 0| 0| 0| 0| 0| 0| 0| 0| 0.. 0| CDC : I O<3C>O<3C>OO 20 .500050200 x.¢598 50.5200. Lapumm-¢e m.N < m.nm5 nemaaooo. nanoom.o. 09900000 .0000500 00005000 55500000 00«n5900 090 592 On 0.. .0- .0: .0- 00009«00. 0| .0- .0- 0005«000. 00000000. 00000000. 0090.000. 00000000. 50500000. 0005«000. 00000000. 90«05000. 05050000. 00000000. 00000000. ««.05«00. 009.0000. 00000000. 0.500900. 000.0.00. .0050000. 00000000. 00000000. 95000500. 055«0000. 00w00000. 00000000. 00000000. 000..000. 00000000. 00000000. 00000000. 00000000. 00000000. 00000000. 00000000. 00000000. 00000000. 00000000. 0| 0| 0| 0| 0| 0| 0| 0| O.- 0. 0: 0n 0 0| 0... 0| GGDCDO¢3C>OO<3C>O~4 . 400\30¢ 153 007. 0000 00090000 00000000. 90070000 00.00000. «2.00.00. 0..00000. 000000.0. 09000000. 00.00000. 00500000. 90000000. 90NZ¥X¥0 0. . 0090C. 00000000. 000.7000 .XXXXXXX3 007000.0. 00000000. .0- 07707000. 0.900.00 00000000 00000000 07097000 .0007000 90000700 000000.0 00000000 09000000 00000000 00000000 90000000 07909000 .0- .0- 0- 0- 0- 0. 0- 0- 0- 0: o- 0- 0- 0- 0- 0n 0 0: 0- On .0- .0- 90000700. .0 .0: 000.0000. 00070000. .0: 90077000. .0: .0: 00007000. 0| Cl C- 0| 0: .0: .0 00000000. .0 o: On .0- 90000000. .0 .0- - .0 .0700000. Co 0.. 00:.002_uw¢ lauaombuc 0 ..000000. 07009000. 0.9000.0 00000000 00090900 70009000 009070.0. 00000000. 00000000. 0090.000. 9900.000. «0.7.000. 00000000. 00000000. .9000.00. 00000700. 00000000. 00000000. 000..000. 00900000. OI o... .0- .770.0.0. 00000000. 70.00000. 00.00000. 99000000. 9.0.0.00. 00000000. 00000000. .0900000. 00000000. 00090000. 0000.000. .0900000. 00000000. 70000000. 90000000. 00000000. .0 00000000. 09000000. 00000000. 0.090900. 90000000. .0 000.9000. 00009070. .0: 0- 0: On 0: 0n 0: 0 0 0- 0 0- 0- 0- 0 0: 0a 0 0 0a 0: 0- 0. 0| 00 0| 6 o O! 0| 0| 0| 6A 00 0| 6 6 0| 0| 00000 00007000.0- 0000.000.0u 070.00.0.0a 90000000.0c 07000000.0c 00000700.0o 900.0900.0: «00090.0.01 0700.000.0a 00000000.0n 00000000.0 0000.000.0u 00000000.0 0.000000.0u 90000.00.0c 00000000.0n 00000000.0 00000.00.0n 09.00000 00 09000000.0| 00000000.0| 070.00.0.0u 00700700.0u 000.0000.0n 09000.00.0 00000900.0u 0000.000.0a 00090700.0u 00000000.0 .97.0000.0u 0009.000.0u 00000000.0 00000000.0 00000000.0 70000000.0n 00000000.0 0.000000.0| 00000000.0 .0090000.0: 00000000.0n 00000000.0 .h48 .2>0 0 00.90<46 9. 00 n «0000. 009.0000. 970.07.0 700000.0. 00070.00 09700000. 00000000. 00.000.0. 90000000. 000.0000. 09077000. 00007000. 00000000. 00000900. 00000000 709.0000. 00000000. 007779.0. 09090000. 00000000. 00000000. 09070000 00000900. 90090000. 97090000. 009000.0 9.000900. 009009.0. 00700900 00900000 000009.0. AxXXXXXXY 09090000. 00007000. 9.900000. 00000000. 99000000. 007.0.00. 00000.00. 0.009000. 0.900000. 00000..0. 00000.00. «990.000. .0000 0 0.4.4.! 0. D. 0.. .00 0| .0! 0.. o.- 0.: 0| OI o.- OI OI Oi o.- o... 0 Cl C0 0 0.- .0l 0| 0| CI .0... 0| 0| .0 .0! o.- 0 0| 0| 0| 0| 3| 0| OI 0| 0| 0| 0| Cl 0. .0 000.0000. 90700000 070000.0. 700.00.0. 90000000. 00099000. 000.0000 .0000000 00000000 00..aXXXy. .XXXKXKXY 0000.000 09000000. 00.00000 00000000 900.0000. 00000000. 00700000 00.00000. 00.00000. 00000000. 90000000. 00000000 .XXXXXXXY 00909.00. .7007000. 900.0000. 79.90000. .0000..0. 00000070. $030090. 0000.000. 0700.000. 00000.00. 0.. .0.. 0| 0| 0| 0| .OI .0l .0.- 0| CI O! .0! on. 0.. .OI O! o.- 0| .0! o: o... 0| 0| 0.. .o- o.- 0.. GI 666666 .0 . .0 02—20. .00.. 0. x.0.<8 mu..200. goaummuee 90000000. 90900700. 900700.0. 0000.700 09000000 00009000 00000000 .7900000 00700000 00000000 00.00000 00000000 00707.00 70.09000 07900000 00000009 00000000 90000900 .7.00000 0: On 0| .0: .0: 00090000. Cl .0: «0907.00. 00000000. 00000000. .0 00..0000. 90.00000. .0- .0- 70009000. 0- 0a 0 0| 0| 0| .0 .0: .0070000. 0| .0 090.0000. .0: .0- 00000.00. .0000000. 790.0000. 0070.000. .0; ..0.0.00. 90309000. .0 00000000. .0700000. 0007.700. 00000000. 00000700. .0: 00000000. 00000000. .0- 00000000. 0.. 0| 0| 0| Cl 00 0| 0- 0- 0- 0 0- 0 0 0 .0- .0 00000000. .0 00000000. 0 0 .0003 00.. .02.92.00 .< 0 CumO<3C>O0 02. .xmczacpu! 00 00005000. ..10.000. 000.00.0. 55000000. 01000000. 01005.00. 00050000. 50000000. 11105000. ..050.00. 0.00.0.0. 0050...0. 00000000. .0000050. 000100.0. .0 00000110. 0100.1.0. .0: .0- 000010.0 .0101000 50000500 411.0000. 00012050. “1100500. 00500100” 0510.090. 00.00.00. 00000000. 00000000. .0: 0100.000. 00000.00. 00111000. 00000.00. 50000000. 00000000. 00000000. 00000000. 00000000. 00000000. 001.0000. 01000.00 00000000 00000000. 00000000” OI OI OI OI OI OI OI OI 0| 0| 0| 0| 0| 0| 0| 0| 0| 0: 0- 01 01 Oil 0: 0| 0| 0| 0| OI 0¢00<0<3C>0 00000000r0. 2001.20200. dsiz 50001000. 5.000.00. 00000000. 010.50.0. 00100000. 00005.00. 0.005000. «0.10000. 50100000. 010010.0. .0..0000. 01050500. 0.000000. 00000..0. .0000000. 00000000. 00105000. .0 51000000 10001000. ..0.0000. 0100.000. .01 00000110. 05009000. 05050000. .5005000. .01 0000.000. 0.5.00.0. 50501100. 00005000. 0000..00. 00150000. 00005000. 00000000. 00000000. 0.100000. 100.0000. 00000000. 00010000. 00000000. 0.000000. 00000000. 00000000. 0 ...0000. 4450’ 0(0 50 . .1.0000. 00000000 OI OI OI OI OI OI OI OI OI OI OI OI OI OI OI OI o- OI OI CI 01 OI OI 0| 0: 0: 0- 0- 0- 0: On 0 0 0: 0- 0 0a 0 0a 0 0 .onneooo.o- oooooooo.o emmv.o.o.o- «on..mno.o- mmvmq.oo.o- ounnweoo.o- unmove.o.o- oooooooo.o «mow—poo.o- o.m~oooo.o- mmwchnoo.o- on«.mooo.o- .qnnnaoo.o- .onahnao.o- accomqao.o- m.mmsuoo.o- «o.~mooo.o- onmoncno.o- mna.aeom.o «shown.o.o- anamnpco.o- oooooooo.o oooooooo.o mmmp.moo.o- punhmmoono- mvn«~.oo.o- oooooooo.o opoanqoo.o- oowmnm.o.o- «.momooc.o- .nuumono.o- wpmmnnoo.o- mqnm0u.o.o- mon.e.oo.o- oooooooo. oooooooo. oooooooo. oooooooo. oooooooo. mwwoo«_o. oooooooo. oooooooo. oooooooo. ooocooeo_o. noz.us‘.u aapux cu 0<3C>00<3C> oooooooo .oooooeoo. n.onnooo. mamnoeoo. qwmqnwoo. «www.moo. coco.ooo. anacoooo. mmnmm.oo. .nsmovno. .ao.eooo. .ooa.ooo. monvmnoo. nomamooo. «oceoooo. comma..o. mnemon.o. nmssoooo. .mnv..oo. ohmeaqoo. «memo.oo. annqcmmm. mesmeooo. oooooooo. .nempomu mnm..roo. pfmem.oo. ew.hnooo. oooooooo. ooocoooo. .peq~ooo. «cacoooo. woumm.oo. nsmuoooo. «nemoooo. .m.e.ooo. oooooooo. oooooooo. oooooooo. .nenoooo. oocooooo. nunqn.oo. oooooooo. oooooooo. OI 0| 0| 0| 0| 0| OI 0| 0| CI 0.. OI 0| 0| OI OI OI OI 0| .0: 9.. . 00¢C.0¢§<; g..oo¢u. zu0¢DC>000<3 CI 9 CI 0<>0<3C>0<3C>00<3 m >a0<3C>0<3C>0 ...qu 2(9— .1p0.p&0( 66 Legumm-¢¢ .0060000. 66566060. 60067900. «00660.0. 6.069000. .@«00.00. 056m5600. 67666600. .0: mv.66000. «6066.00 wh«..06h .he«m«.0. 96000000. .0: o.mmhn«0 «0066000. 0..6w«00. @.m0.600. 0006e.«0. 0«706@?0. 00090000. 0?.«0000. .0- 000«q«60 66000660. .0I .0: 06000000. 00000000. .0I m««w0e00. «6670000. .0- .«.«0.00. vmhshooo. «9600000. 00000000. 00000000. 00000000. 00000000. 00000000. m0v«h«00. 00000000. 00000000. 00000000. 00000000. 00763000 hnmwm«.0 00606000 606o.«00 00 0| OI OI OI OI 0| 0| 0| .0 CI 0! 0| 0| 0.. 0| OI. 0| 0| OI 0.. 0| 0.. 0| OGD I I O<3C>0<3CDO0 mh¢ 30.08 96 06.06000. 06heos«0. .0: 009000n0. 0006.000. 00en..00. 66«66600. 000m«~00. 90«.0e00. «0.06.00. M096m0.0 06mm6900 mw«h«6«m. .0: wma.6000. 0«609600. '5600.00 ..666000 Q00.««00 660.«500 .«c.n««0 06000000 «0«60600 00000000 6«h.0000 «6«?.000 00000000 00000000. 00000000. 00000000. 00000000. .0.«.600. 00000000. 6.600000. 00000000. 00000000. OI CI 0| 0: 0- 0- 0- 0- 0- 0n 0 CI OI .0- .0- h..«««.0. 005660.0. .0- 90676.00. .0- 0006m.60. .0600..0. 0| 0| OI 0| 0| .0: 00650000. 00000000. .0- .0- .0. .0- 96m00#00. 07m00000. 06«00000. OI CI 0 I 0<30<3§>0<3C>0<3 n.00u 4(0-3—0uaw 66 onooooo. Mo- .eomonno. OOOmano. .o- 6.00. ««0 0.6~0000 $0609000. 99«.0.00. .«hmvmoo. «6006.00. 66060000 «.06«000 «70q60h0 menmmm6m wh«mmm.0. .0: «hamOOOO. 00000000. 0.090600. 906«6000. «9660«00. 06000«00. .0.m.000. «ahhemoo. «wohwooo. m.h00w00. 6ww..0.0 0m««6000 .«mmOmOO 00000000 00000000. 20¢: 0800. 0600900000 0| 0| 0| 0| 0| OI OI .0I .0- .0- .0 0| 0| 0| 0| 0| 0| OI 0| 0| 0| .OI 00000000. «Gm.0000. 66.00000. «66060.0. .0I 00500000. 0069.000. 6a«h.000. 00000000. 00000000. 00000000. .00000000. 00000000. «v.m««00. 00000000. 00000000. OI 0! 0| 0| OI 0 I UU<.UU.550 ch 6 Guam «6 m.«he000.0u 1w 00.900.0.0I 6e mm0.~0.0.0-. «w mhnm«m.0.0I .9 .060.000.0I 0* 00600.00.0- mo 96200«00.0I 66 «666h600.0- an o..ne~00.0I on 06600«00.0- mm 5606mm00.0I e6 6hhnmo«0.0I 66 0065.000.0- «6 «066e«60.0 .6 ommnm6.0.OI 06 «0«60000.0u 0« mmmnovoo.ca o« 960m06.0.0I h« .n.cm600.0- ma aom.me00.0- ma m«omw..0.0s 9« «m999560.0- 6« .00—0060.0- «« 6h6¢6000.0I .« .smnmnoo.01 0n 0.066.00.0I m. 00000000.0 a. «ah60000.0- h. a«mh.«00.01 w. mam—«000.0- m. .60«9«00.0u 1. 9.0.m000.0- n. mm«~m.00.0. «. W6vm1000 0- .. 00000000.0 0. 60000000.0- 0 00000000.0 6 60600000 0- h «6.00000.0- 0 «0060600 0- m 00000000.0 e 00000000.0 6 00000000 0 « 00000000.0 . :01! .02. .6 400\)0¢ 156 A.¢5(8 uu~52004 .Lopoom-¢¢ vammOmma Onwmmwno ome.o.uo anuenooo cmn.maoo. wuofihn.o. «whammvo moneuoeo ownomooo. omawoooo. omaaoooo. cam.wooo moohoooo. n.wonooo. acnqoooo onmpmooo. puomaooo. «connooo. vmemnooo. cnaooooo. .Oe.ocoo smeooooo. abounooo «connooo ounce.oo .onomnoo. o.on.ooo oocooooo. mameehoo. nac.nnoo. nose..oo. oooooooo. oooooooo. neummooo. oooooooo. oooooooo. oooooooo. oooooooo. oooooooo. ooQOq... .0 .0: .0: .0: 0| 0| .0: .0: 0| o- O! .0- 0| 0| .0- 0| 0| 0| 0| 0| .0- o... .0: .0: .0- 0| .0- or O. OI OC>0¢DC>0<30 ”0| 00000000e0 0000.000. 0000.000. 00000000. .0.. .0: .0 000585 thzu.b>00 99 00000000. .0 ..000900 00050000. 00090.00. 09090000. 05005000. .0- 000.0000 .00000.0. .00.9000. 0.900000. .0: 0000.000. «0009000. .0: 00000000 «0000900 50000000. .0: 0.000000 .0090000. 00090000. .0 90005000. 00000000. 00000000. 09000000. 00.00000. 00000900. 00000.00. 0.000000. 00000000. 05000000. 00000590. 00.00000. 00000000. .0- 0.000000. .0- 0.000000. 00000000. 00000000. 00900.00. 50005000. 00000000. .0 00000000 005.0000 05000000 00000000 5*000000_ u0: 000—). ..0090000 0| 0| 0. Cl 0! Cl 0| 0| OI 0| 0! 0| 0| 0- 0 0 0a 0: 0n 0: 0o 0 0n 0» 0n 0 On 0- 0 0a 0- 0n 0 at Cumnhoudum 09 «909.000 0.000590 90.9..00 00.00000 «0.00.00 00595000 59000000 05000500 .9000000 9000.000 00550000 05090000 00090.00 «005..00 0090.000 00009000 0500.000 00000000 05000000 09090000 00500000 00500000 00000000 09050000 0..N0000 59005.00 9.0.0000 00000000 05500.00 .:000000 00000.00 05900000 .000.000 ..000000 05000000 00000000 90000000 5000.000 05000000 00900090 00550000 0090N000 00005900 00050000 .0..“..0 «9 .0- .0: .0 .0: .0: .0: .0: .0: .0: .0: .0: .0: .0: .0: .0: .0: .0: .0- .0: .0- .0- .0: .0- .0: .0: .0: .0: .0- .0: .0- .0- .0: .0: .0- .0- .0- .0- .0- .0: .0- .0: .0: .0: .0: n.mh.moo.o- we o.~o«qoo.o- nu snsmomno.o- an vvnoomom.o .q known.oo.o- cc oo.onnoo.o- an noownmoo.o- on. omqaoo.o.o- pa «up.n.oo.o- an .ou.eooo.o- ma o.osqooo.o- en onehn.oo.o- no oomqnooo.o- an «avqn.oo.o- .n ouoanooo.o- en «omoqooo.o- w“ oupeaooo.o- on owq.w.oo.o- nu «.mnnooo.o- on woosoooo.o- mu mm.no.oo.o- an nm.m.ooo.o- nu anhuooo.o- «a mowme.oo.o- .n mowmsnco.o- on opmmwnoo.o- m. moqmnooo.o- m. nmmwoooo.o- ». msmewnoo.o- w. cpwo~.oo.o- m: omamomoo.o- 1. moqmaooo.c- n. unansooo.o- a. .maowcoo.o- .. «omenooo.o- o. ammmhooo o- m onconooo o- o onuhoooo.o- n omohqooo.o- o onpnmnoo.o- m oooooooo.o w smenoooo.o- n non.eooo.o- « oooooooo.c . an00C>0"0000CDO(DC)0<3C>0O0<3C>0<>0" 05000599.. 50000000.0 0000.990.0 000009.0.0 00595000.0 0000..00.0 50000900.0 .00000.0.0 900000.0.0 00090000.0 09.0.000.0 00.05000.0 00595000.0 00000000.0 00009000.0. 00000000.0 00000.00.0 00990000.0 00509500.0 00.00000.0 00500000.0 00005000.0 00050000.0 05005.00.0 00500.00.0 0..00000.0 00050000.0 00000000.0 095.0000.0 0000.0.0.0 000.0900.0 00000000.0 00900000.0 00000000.0 05000000.0 90900.00.0 00000000.. 090.00.0 0 00000..0.0 00050.00.0 59050000.0 00009000.0 00000.00.0 00.900...0 00000000.0 . 39020.500 000<¢u>w0 190.02.480 umm..noo.. mammonoo.o «monum.o.o enen.m.o.o oOanOmo.o sumam.oo.o c.0amvoo.o ohnqsmoo.o oomomwoo.o monomooo.o o.m.mooo.o .0005000.0 psfin..oo.o nu.ooooo.o ..oupnoo.o swmcm.oo.o .mmmo.oo.o aconnooo.o anau.nvo.o no.5mooo.o ananwooo.o «mamo.oo.o .mhno.oo.o osmn0voo.o oeomo.«o.o «.OVQmoo.o vnmufiuoo.o macheooo.o o»«..~oo.o munvn.uo.o mmmmho.o.o snmeqvo.o mamp.ooo.o www.mnoo.o oncom.oo.o oohh.woo.o .oqohhmo.o nmmw.uoo.. .onqsuoo.o m...cmuo.o «.ouoooo.o h.awmooo.o oaoopuoo.o .opoap...o cocoon.o.o 0000 00>¢0000t 000.0.00.. 50900000.0 90005000.0 00.005.0.0 90.0.000.0 05050000.0 00000900.0 00000500.0 00900.00.0 00500000.0 00090000.0 90.00000.0 .00.0000.0 00000000.0 00000000.0 00000.00.0 000.5000.0 50090000.0 .0000000.0 000.0000.0 00009000.0 00000000.0 99000000.0 09.50900.0 00090.00.0 09090900.0 00050000.0 00000000.0 00009000.0 00000500.0 09.90900.0 00000590.0 00000000.0 00090000.0 00500000.0 90.90.00.0 00900000.0 99.09000.0 09000000.. ..0000000.0 00.00000.0 00000000.0 00000000.0 05050900.0 905.90...0 05030003 50.90 5 00.00.05.. 0.0.0000.. 000.0000.0 0000.000.0 00500000.0. 9059.050.0 00.050.0.0 .5000000.0 50059500.0 00009.50.0 900.0.00.0 90000000.0 0.000900.0 90..0900.0 90900000.0 00000500.0 000000.0 0 00900000.0 05090000.0 00500000.0 .9000000.0 00000000.0 09000000.0 50090000.0 .0500000.0 000000.0.0 90.00000.0 0590.000.0 00000900.0 .0590000.0 09500000.0 00000000.0 95050.00.0 00550000.0 00000000.0 .0000000.0 00000000.0 000000.0 0 00900000.0 00500.00.0 ..000000.0 00990000.0 05050000.0 0.000900.0 09090000.0 5.000590.0 00.00.00.0 0.050000.0 00090.00.0 0.09.050.0 09000000.0 09009000.0 50005000.0 99.9.000.0 00005000.0 00090000.0 50000.00.0 00000.00.0 000000.0.0 005.0.00.0 00..0900.0 9.50.900.0 90000900.0 000000.0.0 0990.000.0 0.909000.0 00000000.0 00500090.0 99000.00.0 00.0.000.0 00000000.0 09000000.0 .0509090.0 00900000.0 00005000.0 00900000.0 .090.000.0 00000000.0 00000000.. 00000000.0 00000000.0 00905000.. ..900000.0 90000.00.0 0000..00.0 000000.0.0 05000050.0 00000000.0 05000009.0 90000000.0 05000000 20 5901 .500850200 0 0 Lopumm-qe x.¢.<$ 0030):. ¢.N < m.nm5 00.00900.. 00000000.0 .0900500.0 09.05000.0 500900.0.0 50050.00.0 .00.9900.0 00000000.0 09000000.0 00905000.0 09.00000.0 05000000.0 0509.500.0 0000.000.0 .00900.0.0 0.000000.0 0090.900.0 .0.00900.0 0.095900.0 .0500000.0 00009000.0 50500000.0 .90000.0.0 50.0.900.0 00000000.0 500.0000.0 00000.00.0 0000.000.0 09000000.0 00000..0.0 95005.00.0 00000.00.0 00050000.0 00050.00.0 05000000.0 5..00000.0 505..000.0 00090000.0 99.00000.0 00.00000.0 9.000000.0 00000000.. 00000000.0 .0005000.0 00009000.0 090 592 0 0505 00000 9 09000090.. 00000000.0 .000.900.0 900...00.0 05.00500.0 00000.00.0 0.9000.0.0 09900090.0 05090000.0 99500.00.0 90.00000.0 00005000.0 0000.900.0 0009.000.0 00990000.0 00005000.0 000050.0.0 059.00.0.0 0000.500.0 000.0000.0 0.000000.0 50000000.0 00500.90.0 000005.0.0 05.00000.0 005.0..0.0 .9000000.0 00000000.0 9.000.00.0 00090000.0 00.05000.0 005.0..0.0 00900000.0 0.905000.0 5500.000.0 00000000.0 09000000.0 0000.000.0 0..0.000.0 0.090000.0 00009500.0 0.050.00.0 090..00... 09000000.0 0.000000.0 02.2.! 2.8 0 44508 .0000009. 0009.900. 09000900. 000.0000. 00000090. 05050000. .00.5000. 09559000. 9.0000.0. 00900000. 00009000. 0.090.00. 50000.00. 0050.000. 00900000. 00.00000. 59005..0. 05000.00. 09.00000. .00.9000. 00009000. 09000.00. 00.90000. ..900000. 000.0000. 00009000. 000000.0. 09000000 05000.00. 00095050. 5.000900. 95.50000. 0005.000. .0000000. 05500000. 5.00.000. 00009900. 00050000. 00000000. 00590000. 055505.0. .0000.00. 00000000. 00950.00. 0000..00. 05000000 {3"0¢3C>0<3C>0<3C>0<3C>0¢3C>0<3CDO<3C>043C>0¢3C>0¢3C>OWDC>000<3C)0<3C>00"C>00<3C>0<3C>0<3C>0<5C50<3C>O<3C>02. 90000.09.. .0000500.0 0000.090.0 05.00000.0 00000.00.0 .0.90000.0 55000000.0 990..500.0 005000.0.0 00000000.0 0000.000.0 000.0000.0 00000000.0 .0000000.0 00050000.0 0009..00.0 00009000.0 0900.000.0 559.0000.0 0.000000.0 0.090000.0 00000000.0 05.0.000.0 0050.000.0 00000.00.0 00000900.0 000.5.00.0 00590000.0 90000000.0 00099.00.0 .0009900.. 00..5.0..0 05050000.0 50000500.0 05500000.0 05000000.0 0.900000.0 0000.000.0 00.0.000.0 0000.000.0 0.05.000.0 09.50000.0 .0900.00.0 00005000.0 000.0000.0 .opumm-.. nmnom.o... ¢a~«.moo.o mhaeaomo.o oasannuo.o noaammno.o .onomnoo.o cahpm..o.o ouuo.c.o.o pvhnoeno.o .mmmo~00.o mopnnooo.o o.ao..oo.o oeonh.oo.o avmm_ooo.o «ponmhoo.o onmm.1oo.o onmom_oo.o oammnooo.o phpomm.o.o owweuooo.o vmuvmooo.o mmmnm.oo.o oamv.poo.o commemoo.o o.n_mqoo.o oaavuw.o.o .ome.moo.o nm.q.ooo.o o.mwo..o.o up.mnvmo.o mommmo.o.o awnmms.n.. o.qmvooo.o op.opnmo.o «oonohoo.o oumnoooo.o o..omooo.o he...ooo.o «mo.nooo.o nonoaooo.o smhumm.o.o com...oo.o mommmmoo.o ammouqoo.o ..«nnooo.o .0000 00.4 .02.52.00 49 0 00090 9. 90.00059. 09090000. 00009000. 000000.0. 90000000. 05050000. 090.0900. .5099000. .00000.0. 50090000. 99900000. .0090.00. 00005000. 5050.000. 50005000. 55000000. 90000000. 90009000. 0.0.5500. .0090000. .0.95000. 09090000. 00000000. .00000.0. 00009..0. 50000000. 00000.00. 0090.000. 0.000900. 090000.0. 0.500000. 090500.0. 50005000. 99000050. 50000000. 09.00000. 00900000. 00050000. 00.00000. 00.00000. 005.0500. 000.9000. 5009.000. 595.0900. 00050000. 0005.2000 029 000204 0. 0<3C>0<3C>0<3C>0<3C)’<3C>00<3CV0<3C>0C>0¢DC>000000. human—.0. Goonowuo. pnhnmnoo. QONom.-0. Novonhoo. om—mn..0. 000v.000. Opnommwo. mnFOOG—O. 00’00000. Omuuhpoo. .hhmuhno. Oh00m500. OONOBmOO. .omnhvoo. .mmh.@m¢. 00¢0m000. cachmwoo. ov0nwmoo. OOGOMNOO. one.«000. hunch—00. Vomnwn«0. Opnmmwoo. OOQmOF—O. fishvvuoo. unw.¢000. w.mmm.00. awhp0000. .00—N000. 00000000. wmencooo. nm0m.000. oohhomoo. .m7n0000. Quonmh—O. Gnonmooo. Oh00¢000. 0035 J0<>0WDC)0<3C>O0<3CDOCDO<3C>O<3Ci0(3"0<3C>0<3CDO¢DC>00¢>O" OF-mhnhm.— «thbvoo.0 00mh00—0.0 nmwnnon0.0 NOONGNm0.0 @0050n00.0 ..nmn000.0 n70v.nu0.0 @mhnN000.0 mownh000.0 00.0h—00.0 0mm0n0w..0 «.mhnv_0.0 nah—ha00.0 Nov0m0n0.0 m.—nnOn0.0 Nummm100.0 rpm-«N00.0 wpmenmno.0 h000m0v0.— ?@v¢¢w¢0.0 OanhN0—0.0 cannon—0.0 mo.000.0.0 mmmvvu—0.0 «00507.0.0 900m0n00.0 .Omww000.0 00mmo¢00.0 mmvmnwu0.0 vnwhnN00.0 050.0?90.0 0..00700.0 nv00m0.0.0 0mmnmn00.0 hw0n0000.0 nhnmn000.0 vow—.000.0 OONQ-000.0 u.wm.000.0 «.mano.0.0 hmh00000.0 nomnmfl00.0 om—FN.00.0 n00mn000.0 m02~a80CDCDO<3C>O<>0OC>0(>0<3C%0"C>OOO<3C>O¢3C>0000.0 «o-mmn00.0 vv.@.>w0.0 00mn«¢00.0 Oevmhn—0.0 pmhnm.n0.. .—wmmmm«.0 moans—00.0 an.om0—0.0 hw—pvn00.0 wwhwwn00.0 onmh.000.0 m.vv0—00.0 omvnvn.0.0 amounn00.0 00mma0.0.0 .n’hm~00.0 mmnm0m00.0 m6000_00.0 nsnn0000.0 hhmo.000.0 hm050000.0 h0nn.000.0 Nona—000.0 whenuw.0.0 QFNN0000.0 0vwo«c«0.0 0m0nm000.0 .005.000.0 00mm 045. v“ x.¢»ra umau>z_ . goaumm-¢¢ @0000500. wOONHwOO. ..Nmmm—O. Ohwwnuuo. achwuhmo. m.mwew00. Omwhmmno mmhnwmno. uvv.n0—0. anomNOO. mowmm000. convuuoo. am.mm—_0. unnmnooo. vbph—QNO. v.nv00.0. mononwoo. HGGNONOO. monumnno. anhoooo. Oh—.N—00. «wwovnoo. vv.oho.n. .moovomo. nmnhmw.0. hmwmmnOO. vmmh0m00. a00au000. 000nw¢00. hum.O—no. Gamma—00. mm>amm00 OFDRHOOO. uhwanhoo. 0mm50—00. mmo~0000 no00N000. fine-.000. m~on.000. wwwn.000. n—F—wao. QBOBONOO. nvN0.00-. h00~0000. «huhcooo. ooooooooodooodoooooooo-ooooooooooooooodooooo- .ho(u:Z<8 nu Omnamuvm. 00.09600. cachmm—O. ewnomn.0. Ono-mmuo. hemhmooo. em.hnm00. n.—.mm—0. ov-««c@0. «om-@000. «5059000. nmhnmhoo. nmhvwhoo. .0.000000. oumm.u.0. nhnpnop0. «nomnnoo. O@—.n.00. h090m000. mm9m0000. Onuhvhoo. checOHOO. onmw¢h¢0. «webmwop. 00mmww00. mwwmhuoo. finmwnaoo. qu_—000. mwemnnoo. Nwmhna’O. whwmm.00. Canmmnoo. mnnmaooo. O.Nuwnoo. QNNONOOO. hNo.0000. ohnwnooo. nnmvoooo. FOMNNOOO. mun—’000. Oman—moo. b.00h000. Hannahno. QGQNVOOO. ”Opt—000. 003a 44h”! 0<3C>0<3C>0<3C30<3CIO¢>0MD<>0<3C>0'-CDO<3C>0<3C>0¢3C>O¢3C>0¢3€>0¢3C>0<3C)‘ .mu~¢az:o. zopm.mm<40.u .« JOU\3om 160 .nm.mfiohh. swvnpwwn. awnwmneo. om.amom0. nOonmv.0. aOmmenoo. wowmwnvo. mnmohnho. No.9nnm0. mwoosooo. «n.0n000. m-a~.00. n¢9nan00. vhno.ceo. wwwmhmoo. .wnmqooo. .unn.n.oo. 3.3.5000. nmm.m700. 00H.QCOO. cenw.ooo. armhmmoo. nmnwmmco. 9~0m.moo. vmh¢msoo. .hocnnoo‘ newmmnoc‘ mh00.030. .ho.nmocc. .chnsC.C. m..qnnoo. Ovocneco. 0mmmoooov .wmmmeoo. 00360.30. 2.0.0000. nmmC.CC¢‘ .vaeOOSC. «n00.005. mmwc.ooo. Omwoo.o.. nctmwnoo. NomaONoo. «356.9030. 2.3.3000. m >a0<3C>00¢DC>0043¢)C>0¢DC>00C3C)0<3€>0<3C>0¢3"0<3C>0" vanOnQ..? ..owma.0.0 «mmnwa.0.0 neowwm.o.0 OOQNm300.0 vnm.w000.0 ammwaw.0.. hhhvnu00.0 mnw¢0100.0 mv.hN000.0 15¢?0000.0. 00mv0000.0 mmo.0.00.0 Oahn0000u0 90007.00.0 0090.00020 wh.mm000.0 huh00000.0 Qowom000.0 av.00000.0 @Omm.000.0 cunnmn00.0 ovom0900.0 vowmm.00.0_ enam0w00.0 mmohm000.0 n.wmm000.0 ..mm0000.0 mann.000.0 Commh.00.0 Onmw..00.0 m.mwv.00.0 m»0.~000.0 @0nn1.00.0 0h90n000.0 wm¢00000.0 hflv00000.0 w¢000000.0 hnonOOO0.0 00100000.0 mouhw.00.0 womvpam0.0 .m.n0000.0 wnvvn000.0 anh00000.0 .ww. an m0hnvs05.. obnma.0..0 noewu¢N0.0 000mmm.0.0 «Uncom.0.0 hn.0N000.0. OONQNQM..0 Qm0mmvvn.. oomwflcn0.0 QOmm2000.0 90.0.000.0 @Owh0000.0 Oommmcoo.0 0.0m0000.0 omhww000.0 .nm.6000.0 chow0000.0 hm¢n9000.0 .vOmwa00.0 noonn000.0 .ann000.0 .mno.vm00.0 pmqmnmoo.o o.«omnoo.o n~.nmeoo.o amenm.oo.o nomvmhoo.o mauhoooo.o wonnqooo.o «mmmemoo.o an.mcmoo.o ..n.wnoo.o npmnwooo.o .nonmuoo.o q.anmooo.o nqh.oooo.o mnemoooo.o mpmnoooo.o mowo.ooo.o «ammoooo.o nasqqnuo.o msm.omoo.o «oppe.oo.o oomnqooo.o «vnq.ooo.o nm...:e_: uhvnmmON.. «..0mn.0.0 QNQNQN00.0 vowwmv«0.0 Omnwmo.0.0 Qumhn.00.0 .0vo~nO0.0 wh.no..0.0 voC«.mnO.. 0.mm..00.0 0N¢m1N00.0 ummm.~00.0 ouaumn00.0 m0000000.0 non.hnO0.0 v.09m000.0 ..Q.0000.0 00006.00.0 @mn0m000.0 1v.un000.0 oovnm.00.0 m00.mu00.0 awowOn00.0 «mwvnm.0.0 00m.nn00.0 ow0w0n00.0 .ommvmoo.0 .Qh00000.0 m0009000.0 un.vmn00.0 000v.v00.0 n.0um000.0 nmw-000.0 hmnmm.00.0 naco..00.0 «Ohm0000.0 @000u000.0 061N0000.0 noanN00.0 .mhowooo.0 .0Nn00n0.0 :Qnunv.00.0 mucmn.00.0 bass..00.0 0000h000.0 8300.0. #0 .mowonmn.. vo.nmw00.0 v.momN10.0 @.hmow.0.0 0h0««mN0.0 096mm000.0 «coo—000.0. .mnv0000.0 .mnmmm.0.0 aw.mmw«0.. awhvv.00.0 mmmn~¢00.0. hemmmw.0.0 mmw.hOO0.0_ 90.nmh00.0 vh.mQ¢00.0 0.0.h.00.0 vaum0000.0 mmmnmm.0.0 mwwmmm00.0 eneonm00.0 hmmve.00.0 ”chomn«0.0‘ VQFann0.0 mon.0h00.0 anemnw.0.0 mwwmw.00.0 60090.00.0 wmvwnm00.0 cnwv..n0.0 hasn6600.0 mwmemwno.0 mmoww.00.0 hmwnmhoo.0 Dovvvh00.0 .00N0000.0 hmhwn000.0 ow>.~000.0 «0v0.000.0 wmmmmOO0.0 nunhmwoo.0 wv1m0000.0 .mumwnv00.0 n00nn.00.0 non'm000.0 uz<8 .Um.8 On x—Cpta amau>2~ nmmpwnmm. mo.mp«oo. monocono. .vsepp.o. v.oqq.no. c.0cmooo. nomomnoo. eno.mooo. opon.u.o. .Onpmmpoo. nmmmcnmo. .Cwmpw.o. .hommnuo. cmmeaooo. noesomeo. meOOm.o. emco.moo. mmp.w~.o. apnonmuo. uwnOwooo. oomnsmoo. m.nnmv.o. «.mqnoqo. «acn.mmo. whmOQooo. .mpmmooo. wwwOnnco. weoo.ooo. wnnuenoo. phannpoo. wwhmenoo. ev.oowoo. mnanneoo. vnqoo~.o. .cm.waoo. vmo.oooo. wean.ooo. «canoooo. memo.ooo. mawv.000 000.0500 aOBvQOOQ. whowoooo. Wcm90.00. 9010.000. 1.00u.2(¢b ooooddooooooooooooooooooooooooo000vooooooooo- .:.0.b50( an gouuomuwc nnvap.o.« eccennoo.o n.opsumo.o oaupnu.a.o nc.hdmno.o nunw«.oo.o ommoamoo.o momma..o.o am.«nnno.o ...0nvoo.o «vomq.oo.o «.smqmmn.. acnnonua.o ...0.0000.0 u.«a«ocho manoeh.q.o pheaomoo.o cameo..o.o «cocoonq.o amnesmoo.o oqm.u..ogo .momnnoo.o mhpwuuha.o ooh.mmpo.o n.an.o.q.o .nusn.uo.o a«.~««oo.o .mvo.ooq.o «macsnoo.o omunms.o.o .oamenoo.o qa.«o..o.o noesm.oo.o nmomavoo.o new.m..o.o p..aoooq.o awoanoo.o .nnooooo.o svhv.ooo.o .oamm«000.0 ..pm.o.qto .vummooq‘o no.5uoqq.o cvswmooq.o wasvuooo.o .u»:¢c.c .:ux echo: '0 owmmnnov.., ov.ooaoo.o «mama.eo.o .oqwmo.o.o s.nmm0no.o .qmnwm000.0 m~.suqoo.o maounmoo.o mmmqmq.o.o vmnpumoo.o «amonuoo.o «monamoo.o ohoommoo.. wohw0aoo.o «vvuoq.o.o p.n.no.o.o cano.uoo.o u-«.noo.o ppmn0p.o.o ammowu.o.o «mucmooo.o sap.s~oo,o nomdamno.o ...omcqo.o snoo.o.o.o ommmo..o.o o.h~n.oo.o vnhm.ooo.o «snncmoo.o monmnq.o.o apnmqaoo.o «ocnmw.o.o .wmnsvoo.o vomomnoo.o «qohq.oo.o n.a.oooo.o o~oo~ooo.o mo.woooo.o coo..ooo.o wpsp.ooo.o meopspoo.o «oqnqooo.o an..nwoo.o cunnaooo.o .on..cco.o «.:Ou an Ononsnov. mospnuoo. non.amno. ammm.nno. ;.mooeno. oaosqooo. «..0nnoo. aqnwvmoo. «mo.mo.o. swmcvnoo. unpuwooo. 4~qm0noo. o.oo.noo. wamw.muo. emwcq.ao. womh.n.o. nvmoaooo. na.nmooo. aw.memoo. «.mmho.o. mommmooo. wmquaqoo. .wo«.m.o. cowoam.o mosvnnoo. awqmqmoo. nq0mo_oo. «Napoooo. n.n0s.oo. nempaqoo. «mo.me.o. .cmmo~.o. asmme.oo. mmqoh.oo. «nonpooo. manaoooo. mhsq.ooo. oqmnoooo. no.moooo. uo.«.ooo. «Omaqooo. manqaooo. ceuwsaoo. unannooo. caco~ooo.0 000¢3C>00<3C>O0<3C>0¢5chpC>O<3C>0OOO(DC)O<3C>0O<3C>0O<3C>0OO<3"0<3C>0¢D<>C>0¢3C>O¢3C>' .ru¢:.¢:oua zu00 000 bouqum 99 O‘DOCDOWDCSO(3(’0<30¢3CMO<>0¢3CWO<3OHDCDOGDCMO¢DO<3C>O<3€>0<3CMO0<5C>00OO¢DCIO<3C>00"C>' 000.0.00. 000900.0. 00090.00. .0000000. 0000.0.0. 0900.000. 00000000. .99.0000. 00.090.0. 000.0.00. 0.000000. .0000000. 00000000 00009.00 09000000. 00009000. 00000.00. .0009000. 00000000. 00000000. 00009000. 0000.000. 00000000. 00009000. 00000900. 00000.00. 9.000000. 00009000. 00000000. 00090000. 90000000. 0..00900. 00099000 00000000. 00000000 00000000. 90000000. 90...000. 00000000. 0.00.000. 09000000. 90000000. 00000.00. 00000000. 9.000900. .u.0...u 09 cooooooododooooooooooooooooooddoooooaoo-oo- 0 0 00000000. 000000.0. 90090000. 00000090. 0.0900.0. 00000000. 00000900. .09909.0. 900.09.0. 00.90000. 90.00000. 00000.00. 00.00000. 00.00000. 0.000000. 00000000. 0.090000. 09000000. 00000000. 0.009000. 00999000. 00.00.00. 00000.00. 00000.00. 00000000. 00000000. 00.09000. .0009000. 00000000. 00990000. 000.0000. 000000.0. 09009000. 90000000. 0.000.00. 00090000. 0.000.00. 00009000. 90000.00. 00000000. 000.0000. 00090000. 00000000. 0000..00. 09000000. 9000000900900000090090 NNNNNNMGGGOHGOHHVQVQVV ev0cvu10l~cl00<30¢DCDOO0<3C>009<30<3C>0<3C>0<>0"0¢DOH' " fl 9 0 9€!0"¢I@!~IIO¢D - 4.1.00 0 041004023 .9 Jouxxoa 162 .90.000.99. ..000 00.0 9....99 ooo..mo.. 0.0.0.0 «on.sasuu:u acne:go>o0 99 00.0- om9.mmm.0 ~o0.~mm.~. 00..mem.o moo—>500 uuuuo.mm 09 00.0 can..0m.0 9mm..m9.o. oo...mo.. ouaunm .aoc .oucogamc. ..a.ucac.u N9 .m.o~- mm..~.~.o. ~9....o.m. . -o.o9m.~. . ...uoa 9:. o...m.oz: .9 3...- m8 3...: .0.. “3.23. 0.3.5. as. .:...: o9 99.n~ m.o.~mn.. 0.9.oco.. ~.m.mmm mu......= now an .0... .om.mmm.. mmm.m.9.~ mo9.om... no......= .uu.guuo.u mm 90... c9..~9m.m mmo..w~.. 9mn.~wn.m :o..~u.=:eaou vac .ucmz: .:o.uo.goam=ogh .n 0..0u .00.009 .mm.~9. www.mm9 0:.gauuamacoz mucosa..mum.z on .o.9~- m...mom 990.00. mmm.0o9 u:o&n.=au ca.uaugoamcag. Losuo 9cm u.agug.< mm no.O9n 90...o~... 000.....00 c9o.a.e.0~ «ago. 9:9 mo.u.=~> gouoz 9. 09.9.- m9~..99 0o..~em.. oo...~m .:asa.=cu .mu.suuo.m mm 00.0m .oo... «...00 moo.~9 auo=.:uo: 0:..aaeou can ac..=:ouu< .mu...o mm 00.0.- ~9m.m~9 .wo.~.~.~ 9o~.m9~ xgoc.guox .a.Lum:9:. .ouucmu van uw~..o.uwum .m ...N.. -m.m9~ mo9.9wm.. 0.0.90... .guc.;u~z 0:.xgoz.a.mz on «9...- ow.._mm m........ .m..... .uoc.guoz co..u=gum=ou 9:. a... ow nn.~. .mm.m.0 0.0.o0... .mm.~.o .0:.4L3. 9cm mmc.mcu w. om.o¢- .N..¢om ..c.cco.~ m...mmm nuuavogg ...ox 90......o. .N o~.o- ..N..~m.~ ~.o.....9 moo..mc.~ mm:.asoum ...m: o. m..9.- ..9.mm~ com.om. mom..m~ uuuauog. «e.guoz :mgum m. ....N. o...mm om...0o 9mm..9 uuuavog. 9m.au.gan. .m:.n&=.. .0:..99: 9. 09.00- .00... .m9.~o..n 0....9. m:.s:.ue.:caz .wuum .coL. mu 0..00- oe9.9m ~c..ocm.n. 000.00. maustOL. .auox .mo.guc:o. N. .~.om- ..o... 9m..oc... ooo.o.~ a..u .oco.m...._u .N .0.00- o9..m~m mwm.m~9.. 0.0.9.9 Luau-u. 9cm Lanna. ow .0.00- o~...om ..0.... ow..mmm hum:..u¢ saw.oguwa a. .m..- .o..om. can...» o.m.¢m. “wage m. 0~.mn- ~99.w~ om..m~9 .0..9. n.a.goua: u..ozuc.m can mu.»ma.. .. 0°.o9- am..O9. .Nm.oo..9 cam.mmm.~ muuauogm 90...< .m.ou.eozu o. 00...- .9..m.o om9.m~m.. ~o~.mm. 0:.gm..n:m .0:..c... m. 00.0- ..m.ma. .99.m~..~ omm.~mo uuuaeosa 90...< use Luna. 9. a... o.o.o~. 9.0..9. mmm.0.o ogau.:g=m m. 0~.onu cum... ..0.... oom.m~. uuuauog. coo: vca guess. ~— no.n~- 9.0.0.. m.m.cm9.. 099.... .osoqa< 9cm o...xm. .— am.. ._..m.. ..o.mm. ~mm.o_. mooaga>om o. oo.... ..m.o.9 9mm.mmm.. mam.o.... Lama. ...oxom ..=_o.u m 0..0u 0°9.m.o om..9.o 99m...o coo. 9o>gwmwga m 0.... nmm.n9. .o9.mo. 000.090 uuuauogq 0g.oo . oo.o.- .mm..m¢ ~mo.mm~.. .....o.~.. uuua9ogg yam: o .o..m- mm..omm.~ 9m...mw.m .o..~om.9 co..u=g.mcou m om.~ 9o..o.9 9mo.omm .oo.omm «90 .953.92 9:9 a:o.og.m. 992.. 9 09.0 n-.o~0 own...... .~9..mm uc.:.: .ogo=.: vca .oumz m 0..00 Na..m99 ..N.... mm..o.~ nuuavos. .og:..:u.gm< gmguc N 9N...- mm9.m. oo~.o.o 00.... uuuavog. 9:9 xuouuo>.. . oao.-o.m. mucosa: .a:.. unauao mvcosmo .a:.u auuuzuc. Louuom “sauna emu. a... a... e. one-go . .mgm..oo .o mucmmzog.v mgouomm 99 :mm.=u.z "mac..umnoga .zguzo .nuo. oma. 9cm o.m. 0.0 < m.nm. 163 no..- o—~.no~.p- oon.oon.oo ~mm.~—c.mo _a»o» ac.o o o o mwmwgagoucm ucwscgo>ow cc oo.o o o o . mou_>gam couumpom me cod 0 o o 333 poo”. .3532: £3055... No oo.~ ~Pa.~o coo.mm—.e ~.m.mp~.e ..uuoz can u_amo_ogz .v oo.c o o o mou_>gom Agau_:om ace Loam: ow cod 0 o c 332.5: now an oo.c o c o “own“..ua .ou_guuopu an oc.o o c o cowaou—caaaou we. .ocmzz =o_uougoamcogh Km oc.o~_ 5pc... oo~.o ~_—.m. o:.g:uua»=caz mzooco__oum_x mm mm.e- ovo.m- oom.mup omm.mmp ucoaapacu co,uaugoamcagk gmguo can umcgugw< mm mo.m- cmo.~m~- oom.o~_.o oo~.o_m.s wagon can mwpu_;w> gouo: on —n.e _m~.e ooo.~m .m—._o— “cesawaaw _ou_guuo_u mm oo.o o c c agocwgua: acwu345ou new mc_uc=ouu< .auwwbo Nn oe._m~ nvo.n~ ooc.m mec.~n agoc.guax powgamzvc_ pagmcoc can uo~__u_uoam pm ...aw ~a~.n ooh." “aa.o . xgmc_;ucz mcwxgox_ouoz on om.oq cem.~¢ oom.om oev.mn_ xgmcvcuax comuuagumcou can Ego; aw oo.oo_- ooc.m- ooc.c c mw:_a;=~ new mmc_m:u mm 8.: .36. 25.: _3.3 32.8; .3»: 2:82am 3 om.mmm moo.~ c mao.~ wo:_Q5e.m Paumz oN om.oma ~_m.~ o ~_m.~ muuauogm m:_;uux :mgum mu mo.o~m.— mmv.va oce.o nmm.oo muuauogm umauuvgaom .mcwnaapm .mcwuamx e~ mm.m- osm.oo~- ooh.oec.n o~m.oem.~ m:_gauuo»==oz .emam .:og_ mw mm.om .mm.om_ ooo.omm .mm.o~n muuzvoga .mumz .=o_gnc=om - oo.~_- ee~.oo—- oo~.~um.p www.mmm.P »o_u .ocoum .mma_c .m oo.- mmm.5~ cem.F- m~_.mq~ Locaooa ace Lunaax om hm.m oce.wo. ooo.m~o.— www.cem.p xgmc_‘om s=w_oguaa mp mo.¢m mwm.m oo_.¢ awe.“ “wage a. ~e._n- om_.mp_. oom.m~n o__.om~ m—a_gouez upumzucxm can mu—uma_a ~— o~.m- omm.oa- acm.omm.~ osm.me~.~ muuavoga cw.__< .m—uu_5m;u op om.om- mom- oom.~ coo.~ o=.gm__n=g .a=_u¢_ga m_ o_.~_- www.mmm- oop.eq~.~ mom.o~o.~ «guanoga vo_pp< new swam; c— o~.mo— mmm.mo coo.eo mam.~m~ mgau_=gau «— ~q.m. q-.N- osm.oo~._ ¢-.Nm¢._ avenues; woo: uca gangs; ~— Nm.mm mv_.oe oom.~m meo.Pm— _ogmag< can o__uxmh .p .m.e—- ~m~.mo- oo~.o~e mmv.moe mmoagm>mm op ~e.m- mam.om_- oo~._oc._ m__.mom.~ Leann .xgmxaa .mcwagu a ~e.m_ o¢~.~m ooo.~_m aem.osm coo; um>gommgm m ve.mm- amm.om- oov.em. .eo.mm muuavoga »g_oa s mn.nn mm~.mv ooe.~o mm_.mo_ mausucga “no: o oo.o o o o =o_uuagumcou m oo.o o o o mum pagans: can e:u_cgum¢ ouzgu ¢ mm.~- nwm.~ea- acm.noe.mn N—m.m~o.en u:.cw: pogoc_z new .auox n om.n- “no.qm- ooo.mem.. nom.cmc._ «unavogm .ogaupau.gm< bongo N oo.o o o o muuauoga was mucumm>va _ aucmgau»_o oucogowvvo paauu< uuuuunogm Agumae=_ Lauuom a Amcohv msouomm ee .:mmwgowz cw mucmEm>oz u_$$mgh pvma copumnogm van .msuo< ommp cmmzamm :omFgmasoo m.~ < mpamh APPENDIX 3 1983 AND 1986 PROJECTIONS OF STATE ECONOMIC ACTIVITY AND TOTAL STATE RAIL TRAFFIC '164 no.o_ www.mso.o oc~.oom.uo mom._mo.- som.mcm.o~_ _~»Oh Ik co.° o o o mcq.nc~.p mom.om¢ mom.gaguu=u ”casch>o¢ °~ cc.c c o o om~.-m.~. .¢~.m-.~ mau_>g~m euuuo_om m_ co.o o c c -_.¢¢_.~_ cm~.¢oo.m ouuumu .aoa .uu=.g=m=~ ._a_u=~=.m m. mop.uxuh acveapucu mm:_g:uuauaca: mo.m com.moc oom.e~m.¢ °o~.mnm.¢ _¢M.¢~m.m_ em~.~mm.~_ asco=._.uua_z .._~ua¢ .m.~mo_o== ~— co.c c o c coo.m~m.m pm~.om_.m mo_».__u= o. oc.o o c o mom._-.h o-.mmc.o =o_uou.==esou ecu =o.u~ugoam=~gh m. acasapacm o~.~ ~oe.a oom.a~_ ~o~.am_ «Em.mom csm._ce =o_uau.ogmcagh goguo new u»~gug.< e. .¢.~ o_m.o¢o oom.o~_.m °__.~.m.m mm~._op.- mmm.-¢.m_ was“; was mo_u_;a> Logo: m. m_.m em¢.h ooo.~a emm.¢¢_ ..m.m¢~._ who._me u:m5a.=au .auvguum_u ~_ Nu.“ ooo.m ooe.... ooo.o~_ mmo.mv~.m mm~.o~m.~ .uu.guua_m unauxm .»La=_;u~= .— mo.m .sm.m_n co_.vmo.m c-.~c¢.e mam.¢em.~_ enm.omo.n “.maox ecuau.gnau neg xnas.gg o_ ~_u ah.~ m~¢.am_ oom.wm~._ m~¢.wma._ amm.cm_.~ mmo.o~¢ .mm._w .ucoum .gaguaod .gmgnaz a mm.“ _cm._m_ ooo.a~m._ .om.c_a._ Num.mo~ mo~.mm¢ »L~=_Eo¢ ssmpogaoa m cc.“ m_m.oo~ oo~.m_s.~ m.m.e_o.~ .No.wmo.¢ mc~.oc¢.~ mu.umo_a .mmago .m_.u.sm;u N ~e.m smh.m~m oo_._~o.¢ “mm.amm.e o__.m._.m “mo.mom.~ ac.»=.ga .gaaoa .mg=g_=gam .Lonsad o m~.m mmm.¢- oom.hom.~ nae.mmm.~ ooo.~mm.e mm~.~m_.e muuauoga umgu=_g can coon m oo.o c o c cem.~mq.q mmm.~_m.~ =o.uu=gum=ou ¢ on... mmm.~oo.¢ oom.moe.mm Nmo.cfie.mm ho_.¢o_.~ “mo._mm.p m=.=_z m ao.~_ oeo.mmp coo.mem._ coc.p¢~._ ~w~.~am me~.mme muuauogg .~E=u_=u_ga< Lagua ~ oo.o o c c ao_.omo com.o_ muuaooga we. xucumu>.4 _ mam. u.wm..p .vaz 9.»..Lp u.~mugh __a¢ “gauze sewage _~=.u agam=c=_ gauuom -owm_ c. macazu ..a: cum. coauofiogm umuuofiogm uauuonogg mucmng umuomfiogg .u:&u< mam. mam. mam. -C_o a ~mgo..0a m~¢_ Co ”oceanogp. meow u_uungh p—az .»w_>_uu< u.§o=oou mu.:mu~um :o_pummoga Eapumz Aommp u saw» «many uu_ECmLh __m¢ ace apw>wuu< uvsocoum oumgm mo =o_pomnoga mmm_ p.m < mpnmh 165 cc.m mm¢..¢m.m oo~.oon.oo mmn.cmm.ao mm~.e~o.o~_ .auop oo.o c . a a ao~.mmp._ _a~.mpq ugm.ggg~u=u uguecg~>oa ow oc.c c o o amm.¢mm.__ omm.mom.~ mouw>gum vaguo_om m. oc.o o o c om...ae... msa.~mm.m macaw“ ..ua .ou=~.=m=~ ._~.u=~=,E a. “OFVHXQh OCwUQ—Ucu .mcpbzuuukzcflx ca.¢ omc.¢o~ com..-.e om~.m~¢.¢ m_m.¢—_.m_ a-.mm_.~. «soucappuum.x .e..uo¢ .o_~m~.o== N. oc.o c c o mmc.om_.m .me.e~o.m mm_a__wa= o. oo.o c c o moa.m_¢.h mo~.mpm.m =o_u.u.==ssoo can covuuagoamcagp m, «caea_=cu no." Nn~.e oom.m~_ Nmm.¢n_ "mo.mm¢ omo.mmm so.uaug°chagh gaguc ago HC~LULP< ¢_ ma.m mom.m~m com.oh_.m mam.¢oe.m .mm.~mm.m~ omc.~_m.m_ mug.¢ can mmpu_;m> gouoz n. mo.¢ Rom.” coo.hm “ca.co_ m.m..m~._ oo_.mo¢ “caea.=cu _~u_gpuopm ~— mm.m ocm.e ooe._F_ ooh.m__ mwo.omm.e mom.o-.~ _~u_guuu_u ugauxu .»L~=_gu~= __ mm.¢ .Nm.mm_ co_.¢mo.n _~9.m¢a.n m.¢.¢~§.__ om~.oms.n m_~»o= uuuou_gnau van anaspga o_ «.9 om.m .q..o~ cem.ma~._ .eo.moa._ m-.~¢o.~ mmm.m°¢ .mm~_m .wcoum .nguuaa .gonaag a ~o.a m_m.mo ooo.m~o._ o.m.¢¢h.P mgm.m- m_~.~_¢ xgmgwwuz s=a_oguog a _~.m mfio.oo_ oo~.m_~.~ msm.e_m.~ ¢_~.mom.m ofim.ma~.~ mu.um._g .mmago .m.~uweo;u s vm.¢ .No.¢a_ oop._~o.¢ _-.m_~.e m.m.-m.e oa~.-~.~ a=.u=_gg .guaam .og=u_=gam .goa534 o e... mee.o__ com.~om.~ m.~.mus.~ a_~.mme.q om~.¢mm.n finances; umgc=_x ecu coca m oc.c o c w o ~co.m_~.¢ Koo.¢mo.~ =o_uuagam=ou ¢ mo.m mso.moo.~ oom.moq.mn mum.coe.~n om~.~co.~ “mm._a~._ acvgpz n mm.c ~o_.m¢ ooo.m¢m.. ho_.nco._ www.meo mo~.~ov “auguOEg ..E=H_=u_gu< ggsuo N oo.o a a o _-.omo cm..o_ muuzuoga ecu xuogm~>.4 _ ”ma. u.EC~gp __~¢ u_uhugh 0....sh ...z “aauao acosma _~=.. xgum=u=~ segue” -cmm. =_ umcugu ..ag omm. touumfioga umuuonog. umaumfiogm oucmgu. ugauofiogm .aauu< mam_ mam. nua— -..a a . Amgm_,oo o~m_ Ca mucamaogwq mcoh u_CC~Lp ..ag xm*>—uu< u_so:oum muvzouaum =o_uumnoga gov: Aommp u me> mmmmv "uowagh .Pmm cgm »p_>mpo< owsocoum oumgm Co copgumnoga mwap N.m < mPQMP '166 NN.NN omm.vNN.m_ ooN.oom.oo NNN._No.mN NN¢.mmN.ae_ ..Noh oo.o c o a mmc.cN¢.p NNN.Nom mNNNLNsmucm ugusggo>o¢ oN oc.o o . a o Nom.mNN.¢_ ¢N¢.Np_.m mauN>gom coauopam a. oo.c a o a on.aNm.¢. NNc.Nao.__ oaoumu .Noa .oucogamcm ._N_u=~=.m Np mop_uxoh mcpva_u:_ .o:_g=uuum==uz om.mN mmm.Nmo.p ooN.¢NN.¢ NNN.N.N.m NNN.N¢N.N_ Nom._N_.m. “semen—pgum.z ..ruuog .u_umo_og= N. oo.o o c c Nem.m¢N.c omo.mNN.N mg.a...»= o_ oo.o c c c cNm.oN_.m a.e.omp.N =o.u~u.=;eaou ecu =o_guagogm=~sh m_ u:m§a.=cu mo._N mm_.NN oom.aN_ Nmm.Nm_ Nee.oom NNN.NN¢ =o_».»gogm=~gp Luguo age NNNLULP< N. c¢.oN NNN.¢N¢._ oom.oN..N NNN.NNo.a mNN.aNN._N emm.oNa.NN mason uca wu_u.=m> None: N. oN.NN enm._N coo.Na emm.o__ _NN.NFm._ moo.Nom N=NEN_=oN .ou..uum_u N. NN._N NON.¢N cog..p. NNN.mN_ mom.Nm_.e mem.¢oN.N ..u.guuu.m uguuxm .Ngmc.guax __ o_.¢N NNo.omN oo..¢NN.N NN_.¢Nm.. NNN..Nm.¢_ omo.amm.¢ “.ouoz cogau_gnuu can NMNEPLN op «Pu No._N e_a.mNN oom.NmN._ N_N.NN_.N mNN.Nmm.N N_N.ccm .mma.m .mcoum .Laguamd .Lmaaag a No.NN NNN.NNN ooo.mNm._ m¢m.omo.N Noe._c¢ o_N.m_m NL~=_CNN e=a_o.umg N NN._N meo._mm coN.N_N.N meN.eom.N one.mmN.e NNN.NNN.N mu.amN_N .mmaga .m_NuPeN;u N NN.NN om_.mm_._ cop..No.¢ omN.om_.m om¢.¢No.o Nom.qu.N ac.pc.gg .LoNNN .og=a_=gam .gm5534 o .m.mN mmm.ome oom.Nee.N mmN.N¢N.N .NN._Ne.m coo.cNN.¢ muuauogg umgu=_g can Noam m oo.o o a c N.N.omN.m No_.mNN.N =o.uu=gumcou e NN.NN N__.NNN.__ ooN.No¢.mN N_¢.oNN.N. cmN.NNN.N mm¢.¢mm._ u=.=_= N N¢.NN .om.¢am coo.mem._ .oN.NNP.N ¢N¢._N_._ NN_.on muuauoga Passe—au,ga< gmguo N oo.o o a c Nam.NoN oem.m_ mauauogg uca xuoNWNNNN _ ommp upuuas» ppm: u_umagp uvwmas» F’s: uaauso nausea Pug—m xgumsu:_ Louuom -cmap a. mmcacu ..Nz cam. amuumnogg umaumwogN couumnogN oucmgob Nmuuofloga .N3Nu< NNN. cam. mam. N_o N Awgo_.oo on_ No mucmmaoghg mac» Nm_>_uu< u.Eo=oum «v.3NuNNm uwmymsh F’s: Aomo— u gmo> mmmmv =o_uumhogg gap: "U_NNNLN P_NN uca NNN>NN0< uNeozooN mumpm No =o_NumNoNN mam. m.m < «Pack BIBLIOGRAPHY Adiarte, Rosario G. and Ernesto C. Venegas. "State of Michigan Input- Output Model," Mimeograph, Minnesota: Minnesota Energy Agency, 1980. Allen, Benjamin J. "The Economic Effects of Rail Line Abandonment on Commodities: A Case Study," Transportation Journal, XV, l, Fall ”7 1975’ pp. 52-6]. Anderson, Dale 6., Marc A. Johnson and Orlo L. Sorenson. "Rail-Line y, Abandonment in the North Central Region," North Central Regional Research Publication No. 249, Manhattan, Kansas: Agricultural Experiment Station, l979. Baumel, C.P., et al. "The Economics of Upgrading Seventy-One Branch Rail Lines in Iowa," American Journal nggricultural Economics, Vol. 59, No. 1, pp. 61-70, February 1977. Binkley, James K. and Mary E. Revelt. "The Role of Transportation in International Agricultural Trade," Station Bulletin No. 337, w. Lafayette, Indiana: Purdue University, Department of Agricultural Economics, 1981. Boyer, Kenneth D. "Estimating the Price of Elasticity of Demand for Freight Transportation," Mimeograph, East Lansing, Michigan: Michigan State University, 1980. Boyer, Kenneth D. "Using Freight Bills to Estimate the Cross Elastici- ties of Demand for Freight Transportation," Mimeograph, East Lansing, Michigan: Michigan State University, l980. Brink, Lars and Bruce McCarl. "Input-Output Analysis, Linear Program- ming and the Output Multiplier," Mimeograph, N. Lafayette, Indiana: Purdue University, 1980. Cambridge Systematics, Inc. "Freight Demand Forecasts for Nest Coast Corridor Study," Mimeograph, l978. Canadian Transport Commission. "Report of Transport Demand Forecast- ing Program: Phase One," Forecasting Program Working Paper No. 6, Ottawa, Canada, l975. Chenery, Hollis B. and Paul G. Clark. Interindustry Economics, New York: John Wiley and Sons, Inc., l959. I67 168 Chiang, Yu-Sheng. "A Policy Sensitive Model of Freight Demand," Center for Transportation Studies Report 79-7, Cambridge, Massachusetts: Massachusetts Institute of Technology, l979. Chiang, Yu-Sheng and Paul 0. Roberts. "A Note Describing an Empirically Determined Model of Transit Time and Reliability for Regular-Route Trucking," Center for Transportation Stgdies Report 78:11, Cambridge, Massachusetts: Massachusetts Institute of Technology, 9 8. Chiang, Yu-Sheng and Paul D. Roberts. "Representing Industry and Population Structure for Estimating Freight Flows," Center for Transportation Studies Report 76-8, Cambridge, Massachusetts: Massachusetts Institute of Technology, l976. Chung, C.C. and Paul 0. Roberts. "Design of a Structure and Data Analysis Scheme for Intercity Freight Demand Forecasting," Center for Transportation Studiesggeport 75-l5, Cambridge, Massachusetts: Massachusetts Institute of Technology. 1975. Crary, Joan P., Saul H. Hymans and Harold T. Shapiro. "The Michigan Economy in 1982,“ Research Seminar in Quantitativg_Economics Forecasts, Ann Arbor, Michigan: The University of Michigan, 1981. Diamond, J. and D. Chappelle. "Application of an Input-Output Model Based on Secondary Data in Local Planning: The Case of Manistee County," Agricultural Experiment Station Research Report No. 409, East Lansing, Michigan: Michigan State University, l981. Dorfman, R., Paul Samuelson and R. Solo. Linear Programming;and Economic Analysis, New York: McGraw Hill, l958. Economic Re rt of the President l980. Washington, D.C.: U.S. Govern- ment Printing Ufiice, January 1§80. Economic Report of the President, l98l. Washington, D.C.: U.S. Govern- ment Printing Office, January T§§l. Federal Reserve Bank of Richmond. Business Forecasts, 1982, February 1982. Federal Reserve Bank of St. Louis. National Economic Trends, February 26, 1982. Federal Reserve Bank of St. Louis. National Economic Trends, March 3l, 1982. Fishbein, R.E. and S.R. Thompson. "National Income and Employment Impacts of Railroad Branch Line Abandonment: Conceptualization and Estimation," Agricultural Economics Report Ng;_332, East Lansing, Michigan: Michigan State University, 1977. 169 Fast, H.R. "Changing Conditions in Rural Saskatchewan," Canadian Farm Economics, VII, June l972. Friedlaender, Ann F. and R. Spady. "A Derived Demand Function for Freight Transportation," Review of Economics and Statistics, Vol. LXII, No. 3, pp. 432-44l, August 1980. Glassey, C. Roger. "Price Sensitive Consumer Demands in Energy Model- ing--A Quadratic Programming Approach to the Analysis of Some Federal Energy Agency Policies," Management Science, Vol. 24, No. 9, pp. 877-886, May 1978. Hariton, George, et al. "Econometric Forecasting Model, Demand for Freight Transport in Canada," Canadian Transport Commission Report on Phase 2, Ottawa, Canada, June 1976. Harrington, David H. "Quadratic Input-Output Analysis: Methodology for Empirical General Equilibrium Models," Ph.D. Dissertation, Purdue University, l973. Hwang, Henry H. and Wilbur R. Maki. "A User's Guide to the Minnesota Two-Region Input-Output Computer Model: With Reference to the 1970 Minnesota 35-Sector Model," St. Paul, Minnesota: Department of Agricultural and Applied Economics, University of Minnesota, 1976. Hwang, Henry H. and Wilbur R. Maki. "User's Guide to the Minnesota Two-Region Input-Output Model,“ St. Paul, Minnesota: Department of Agricultural and Applied Economics, University of Minnesota, l979. Johnson, Marc A. "Community Evaluation of Railroad Branch Lines: Principles and Procedures," Center for Rural Manpower and Public Affairs Report NpiiBB, East Lansing, Michigan: Michigan State university, 1975. Johnson, Marc A. "Estimating the Influence of Service Quality on Trans- portation Demand,“ American Journal of_Agricultural Economics, Vol. 58, No. 3, pp. 496-503, August 1976. Johnson, Marc A. "Market and Social Investment and Disinvestment in Railroad Branch Lines: Evaluation Procedures and Decision Criteria,“ Ph.D. Dissertation, Michigan State University, l975. Jones, Clifford 0., Jr. Input-Output Analysis Applied to RuraljRgsource Developmept Planning, Washington, D.C.: U.S. Department of Agri4 cUiture, Economics, Statistics and Cooperatives Service-l4, l978. Leontief, Wassily. Studies in the Structure of the American Econom : Theoretical and Empirical Expiorations in Input-Output Analysis, New York: International Arts and Sciences Press, Inc., 1953. 170 Leontief, Wassily. The Structure of the American Economy, 1919- 1939: An Empirical Application of Equilibrium Analysis, New York. Inter- national Arts and Sciences Press, Inc. , 1941. Lipschutz, Seymour and Arthur Poe. Theoryand Problems of Programming, with Fortran: Including,Structured Fortran, New York: McGraw-Hill Company, 1978. Maki, Wilbur, L.A. Laulainen and Mason Chen. "User Guide to Simlab 11," St. Paul, Minnesota: Department of Agricultural and Applied Economics, University of Minnesota, 1977. Maki, Wilbur R., P.D. Meagher, L.A. Laulainen and M. Chen. "User's Guide to the Minnesota Regional Development Simulation Laboratory," St. Paul, Minnesota: Department of Agricultural and Applied Economics, University of Minnesota, 1979. Mathematica. Studies on the Demand for Freight Transportation, Vol. II, 1969. McCarl, Bruce A. and Hal Everett. "Regional Economic Planning: A Methodology for Integrating Linear Programming and Input-Output Analysis," Mimeograph, West Lafayette, Indiana: Purdue University, 1979. Michigan, State of. Economic Report of the Governor, 1978, Lansing, Michigan, March 1978. Michigan, State of. Economic Report of the Governor, 1979, Lansing, Michigan, March.l979. Michigan, State of. Economic Report of the Governor,,1980, Lansing, Michigan, March 1980. Michigan, State of. Economic Report of the Governor, 1981, Lansing, Michigan, March T981. Michigan Department of Commerce, Energy Administration. "A 44-Sector Input- -Output Model of the Michigan Economy," nergy Issue Analysis Series, Lansing, Michigan, March 1980. Michigan Department of State Highways and Transportation. Michi an Railroad Plan: Annual Update, Lansing, Michigan, August 1976. Michigan Department of State Highways and Transportation. Michi an Railroad Plan: Annual Update, Lansing, Michigan, August 1977. Michigan Department of State Highways and Transportation. Michigan Railroad Plan: Annual Update, Lansing, Michigan, August 1978. Michigan Department of State Highways and Transportation. Michigan Railroad Plan: Phase I, Lansing, Michigan, May 1975. 171 Michigan Department of State Highways and Transportation. Michigan Railroad Plan: Phase II, Lansing, Michigan, June 1976. Michigan Department of Transportation. Michigan Railroad Plan: Annual Update, Lansing, Michigan, October 1980. Michigan Department of Transportation. Michigan Rail System Ratiop;_ alization Plan: Tier ngeport, Lansing, Michigan, December 17, 1981. Michigan Department of Transportation. Michigan Rail System Ration- alization Plan: Tier II, Phase I, the Ann Arbor Report, Lansing, Michigan, March 26, 1982. Michigan Department of Transportation. Michigan Rail System Ration- alizationiPlan: Tier 11, Phase II, Grand Rapids to Mackinac Corridor Report, Lansing, Michigan, March 26, 1982. Michigan Department of Transportation. Nprthwest Michigan Rail System Rationalization, Lansing, Michigan, December T7, l980. Michigan Department of Transportation. "Statewide Rail Rationalization Plan," Mimeograph, Lansing, Michigan, June 22, 1981. Michigan Department of Transportation. Summapy of Michigan's Trans- portation Needs: 1977-1989, Lansing, Michigan, 1978. Mi klius, Walter, Kenneth L. Casavant and Peter V. Gerrod. "Estimation of Demand for Transportation of Agricultural Commodities," American Journa1 of A ricultural Economics, Vol. 58, No. 2, pp. 2|7-223, May 1976. Miller, Edward. "Effect of Rates on the Demand for Railroad Services," Proceedings of the Transportation Research Forum, 1972. Oum, T.H. "A Crass-Sectional Study of Freight Transport Demand and Rail-Truck Competition in Canada," The_§ell Journal of Economics, Vol. 10, No. 2, pp. 463-482, Autumn 1978. Oum, T.H. "A Warning on the Use of Linear Logit Models in Transport Mode Choice Studies," The Bell Journal of Economics, Vol. 10, No. I, pp. 374-388, Spring 1979. Patrick, John M. "An Economic Analysis of Improving the Viability of Rail Lines: A Michigan Case Study," Ph.D. Dissertation, Michigan State University, 1977. Patrick, J. Michael. "Improving Economic Viability of Three Michigan Rail Lines: The Ann Arbor, Michigan Northern and C&O-Northwest," Agricultural Economics Report No. 336, East Lansing, Michigan: Michigan State University, 1977. 172 Patrick, J. Michael and Stanley R. Thompson. "The Potential Responsive- ness of Selected Michigan Rail Users to Improvements in Rail Freight Service," North Central Journal of Agricultural Economics, Vol. 1, No. 2, pp. 115-121, July 1979. Pattie, Preston S. "Dynamic Interaction of Location and Transportation: Concepts for Research," Mimeograph, East Lansing, Michigan: Michigan State University, 1978. Perle, Eugene. The Demand for Transportation: Regional and Commodity, Studies in the united States, Chicago, Illinois: University of* Chicago Press, I964. Pratt, James. "Using I/O Analysis to Forecast the Demand for an Indi- vidual Firm's Products," Mimeograph, East Lansing, Michigan: Michigan State University, 1980. Pratt, James E. and M. C. Conner. "A General Computer Program for In- put-Output Analyses," _gricultural Economics Staff Paper No. SP- 76- 6, Blacksburg, Virginia: Virginia Polytechnic Institute and Sn mte University, 1976. Pratt, James E., Dennis K. Smith and M.C. Conner. "Input-Output Analysis: An Introductory Self-Learning Package," Agricultural Economics Staff Paper No. SP-76-l4, Blacksburg, Virginia: VirginiaiPolytechnic Institute and State University, 1976. Pratt, James E. and Stanley R. Thompson. "Integrating a Regional Economic Model with Commodity Flow Data for Transportation Analysis," Agricultural Economics Staff Paper No. 79- 90, East Lansing, Michigan: MichiganState Univer51ty, l979. Rausch, David, L.F. Schrader and Otto C. Doering. "Evaluating Rail Services to an Agricultural Community: A Case Study in Benton County, Indiana," Station Bulletin No. 71, West Lafayette, Indiana: Purdue University, Department of Agricultural Economics, 1975. Richardson, Harry N. Input-Out ut and Regional Economics, New York: John Wiley and Sons, 1972. Roberts, Paul O. "A Set of Simplified Multimodal Cost Models For Use in Freight Studies,“ Center for Transportation Studies Report No. 77- 27, Cambridge, Massachusetts: Massachusetts Institute off Technology, 1977. Roberts, Paul O. "Consolidation: Key to Multimodal Freight Trans- portation?" Center for Transportation Studies Report No. 78-3, Cambridge, Massachusetts: Massachusetts Inst1tute of Technology, 1978. Roberts, Paul O. "Disaggregate Demand Modelling: Theoretical Tanta- lizer or Practical Problem Solver?" Center for Transportation Studies Report No. 77- 29, Cambridge, Massachusetts: Massachusetts Institute of Technology, 1977. 173 Roberts, Paul 0. "Factors Influencing the Demand for Goods Movement," Center for Transportation Studies Report No. 75-16, Cambridge, Massachusetts: Massachusetts Institute of Technoiogy, 1975. Roberts, Paul 0. "Forecasting Freight Flows Using a Disaggregate Freight Demand Model," Center for Transportation Studies Report No. 76-1, Cambridge, Massachusetts: Massachusetts Institute of Technoiogy, 1976. Roberts, Paul 0. "Freight Demand Response to Changes in the U.S. Trans- portation System," Center for Transportation Studies Report No. 77-28, Cambridge, Massachusetts: IMassachusetts Institute ofI Technology, 1977. Roberts, Paul O. "The Structure of the Nation's Future Freight System," Center for Transportation Studies Report Np;_79-6, Cambridge, Massachusetts: Massachusetts Institute of Technology, 1979. Roberts, Paul 0., Thomas B. Brigham and Carol A. Miller. "An Equilib- rium Analysis of Selected Intercity Freight Markets: Trucks with Double Trailers vs. TOFC Shuttle Trains as Energy Conservation Alternatives," Center for Transpprtation Studies Report No. 77-25, gambridge, Massachusetts: Massachusetts Institute of Technology, 977. Roberts, Paul O. and A.S. Lang. "The Trade-Off Between Railroad Rates and Service Quality," Center for Transportation Studies Report No. 78-12, Cambridge, Massachusetts: Massachusetts Institute of Technology, 1978. Schaffer, W.A. and K. Chu. "Non-Survey Techniques for Constructing Regional Interindustry Models," Pa ers of the Regipnal Science Association, Vol. 23, pp. 83-99, 1969. Scheppach, Raymond C. State Projections of the Gross National Product, 1970, 1980, Massachusetts: Lexington Books, D.C. Heath and Company, 1972. - 51055, J., T.J. Humphrey and F.A. Krutter. "An Analysis and Evaluation of Past Experience in Rationalizing Railroad Networks," Center for Transportation Studies Repogt No. 74-54, Cambridge, Massachusetts: Massachusetts Institute of Technology, 1975. Stevens, B.H., J. Kindahl, M. Woelfel and D. Ehrlich. Basic Regional Input-Outppt_for Transportatiog‘Impact Analysis: Handbook ne of RE ional Economic Analysi§_for Transportation'Pianning, Amherst, assachusetts: Regional ScienceIResearch Institute, T979. Stevens, B.H., J. Kindahl, M. Woelfel and D. Ehrlich. "Introductory Input-Output Methods for the Measurement of Regional Impacts with Emphasis on Transportation Facilities," Regional Science Research Institute Discussion Paper No. 122, Amherst, Massachusetts, 1980. 174 Stipe, Sterling H. "A Proposal and Evaluation of a Regional Input- Output Modeling System," Ph.D. Dissertation, Michigan State University, 1975. Survey Research Center, The University of Michigan. "1981 Forecast Issue," Economic Outlook, U.S.A., Vol. 8, No. 1, Winter 1981. Survey Research Center, The University of Michigan. "1982 Forecast Issue," Economic Outlook, U.S.A., Vol. 9, No. 1, Winter 1982. Terziev, Marc N. "Modeling the Demand for Freight Transportation," M.S. Thesis, Massachusetts Institute of Technology, 1976. Terziev, Marc N., F.W. Richardson and Paul D. Roberts. "An Approach for the Evaluation of Freight Network Operations: A Stochastic Supply Model of the Regular-Route Motor Carrier Industry," Center for Transportation Studies Repogt No. 78-7, Cambridge, Massachusetts: Massachusetts Institute of Technology, 1978. Terziev, Marc N.,.et al. "Freight Demand Modeling: A Policy Sensitive Approach," Center for Transportation Studies Report No. 76-6, Cambridge, Massachusetts: Massachusetts Institute of Technology, 1976. ' Torene, Robert. "A General Purpose Disaggregate Data Base with Special Application to Transportation and Energy Planning and Analysis," Mimeograph, U.S. Department of Commerce, Bureau of the Census, 1978. Transportation Research Board, National Academy of Sciences. "Local Rail Services Program: Perspectives of States and Railroads," Circular 209, August 1979. Upper Peninsula Rail Planning Committee, Central Upper Peninsula Plan- ning and Development Regional Commission. "Rail Needs in Michigan's Upper Peninsula," Mimeograph, 1982. U.S. Department of Agriculture, Economics, Statistics and Cooperatives Service. "Regional Development and Plan Evaluation: The Use of Input-Output Analysis," Agriculture Handbook No. 530, Washington, D.C., May 1978. U.S. Department of Commerce, Bureau of the Census. Census Bureau Guide to Transportation Statistics, Washington, D.C.: U.S. Government Printing Office, 1976. U.S. Department of Commerce, Bureau of the Census. 1972 Census of Transportation: Commodity Transportation Survey,Areijeport 3, Michi an and Ohio, Washington, D.C.: U.S. Government Printing Office, 1975. U.S. Department of Commerce, Bureau of Economic Analysis. Surve of Current Business, Washington, D.C., Vol. 59, No. 2, 4, 7, 1%, 1979; Vol. 61, No. 7, 10, 11, 12, 1981. U.S. U.S. U.S. 175 Department of Transportation. "Rail Service in the Midwest and Northeast Region: A Report by the Secretary of Transportation," 1974. Department of Transportation, Federal Railroad Administration, Office of Information and Statistics Division. 1979 Carload Way- bill Statistics: Territorial Distributiop, Traffic and Revenueppx Commodity Classes, December 1980. Office of Management and Budget. CommodityClassjfication for Transportation Statistics, Washington, D.C.: U.S. Government Printing Office, 1972. Verway, David I. Michigan Statistical Abstract, 1979, East Lansing, Michigan: Graduate School ofIBusiness Administration, Michigan State University, 1979. Vollmers, A. Clyde and Stanley R. Thompson. "The Economic Analysis of a Grain Logistics System Serving Michigan's Thumb Area," Agri- cultural Economics Report No. 385, East Lansing, Michigan: Michigan State University, 1980. von Oppen, Matthias and John T. Scott. "A Spatial Equilibrium Model for Plant Location and Interregional Trade," American Journalvpf. Agricultural Economics, Vol. 58, No. 3, pp. 437-445, August 1976. Washington Department of Transportation. "Washington State Freight Transportation: Baseline Data and Forecasts," Mimeograph, l979.