Mexico - Fourth Railway Project

Document of The World Bank FOR OFFICIAL USE ONLY Report No. 10817 PROJECT PERFORMANCE AUDIT REPORT MEXICO FOURTH RAILWAY PROJECT (LOAN 1929-ME) JUNE 25, 1992 Type: lP -rt . 0 pt Operations Evaluation Department This document has a restricted distribution and may be used by recipients only in the performance of their official duties. Its contents may not otherwise be disclosed without World Bank authorization. GLOSSARY OF ABBREVIATIONS AACRG - Average Annual Compound Rate of Growth CTC - Central Traffic Control FNM - Mexican National Railways FTI - Free Trade Initi7tive INUTD - Infrastructure and Urban Development Department, Transport Division JNR - Japan national Railways LCL - less-than-car-load MGT - million gross tons NdM - Nacionales de Mexico OED - Operations Evaluation Department PEMEX - Petr6leos Mexicanos PPAR - Project Performance Audit Report PCR - Project Completion Report SAR - Staff Appraisal Report SCT - Secretaria de Comunicaciones y Transporte SECAL - Sectoral Adjustment Loan STFRM - Sindicato de Trabajadores de Ferrocarriles de la Repdllica Mexicana UP - Union Pacific Railroad GOVERNMENT OF MEXICO FISCAL YEAR January 1 to December 1 FOR OFFICIAL USE ONLY THE WORLD BANK Washington, D.C. 20433 U.S.A. Office of Director-General Operations Evaluation June 25, 1992 MEMORANDUM TO THE EXECUTIVE DIRECTORS AND THE PRESIDENT SUBJECT: Project Performance Audit Report on Mexico Fourth Railway Project (Loan 1929-ME) Attached, for information, is a copy of a report entitled "Project Performance Audit Report on Mexico Fourth Railway Project (Loan 1929-ME)" prepared by the Operations Evaluation Department Attachment This document has a restricted distribution and may be used by recipients only in the performance of their oficial duties. Its contents may not otherwise be disclosed without World Bank authorization. FOR OFFICIAL USE ONLY PROJECT PERFORM-ANCE AUDIT REPORT MEXICO FOURTH RAILWAY PROJECT (Loan 1929-MEL TABLE OF CONTENTS Page No. PREFAC7 ............................................................. BASIC DATA SHEET ................................................... EVALUATION SUMMARY .................................................. iv I. Introduction .............................................. 1 II. Mexican Railways .......................................... 2 III. World Bank Support ........................................ 7 IV. Project Objectives ........................................ 14 V. Implementation Experience ................................. 15 VI. Project Economics ......................................... 16 VII. Findings and Issues ....................................... 19 VIII. Lessons for the Future .................................... 24 Annex I - Current Productivity and International Comparisons .. 29 Annex II Comments by the Government ......................... 37 Annex III - Comments on Events Subsequent to Railway IV ........ 56 Map - IBRD No. 3655R1 This document has a testricted distribution and may be used by recipients only in the performance of their official duties. Its contents may not otherwise be disclosed without World Bank authorization. i PROJECT PEFORMANCE AUDIT RUPORT MEXICO FOURTH RAILWAY DROJECT (LOAN 1929-ME) PREFACE This is the Project Performance Audit Report (PPAR) of the Fourth Railway Project in Mexico. The US$ 1.5 billion project was partially financed by a Bank loan of US$ 150 million approved by the Board on December 15, 1980. The final cost of the project was US$ 1.2 billion. Bank support remained at US$ 150 million. The loan was closed two years late on June 30, 1Po6. The PPAR is based on the Project Completion Report (PCR) prepared jointly by the Mexican Government and the Latin America and Caribbean Regional Office of the World Bank and issued in 1990.1 In addition it is based on the Staff Appraisal and President's Reports, the loan documents, the transcript of the Executive Directors' meeting at which the project was considered, on a review of project files and on discussions with Bank staff. An Operations Evaluation Department (OED) mission visited Mexico in July 1991 to discuss the effectiveness of the Bank's assistance and to make field visits to the completed project components. The draft PPAR was sent to the Borrower for comments on December 4, 1991. Extensive comments were received on January 21, 1992 and are attached as Annex II. The Bank's Regional staff too has commented in detail on the draft PPAR. Their comments are reflected in this report. In addition, the Region expressed the wish "to be on record that on the key issues and findings [of the PPAR] we are in full agreement." I PC Mexico Fourth Railway Project, OED Report No. 9087, October 26, 1990. 11 PJECT PERFORMANCE AUDIT REPORT MEXICO FOURTH RAILWAY PROJECT (LOAN 1929-ME) BASIC DATA SHEET Key Project Data Appraisal Actual or Actual as % 1Expectatio-n Current Estimate Appraisal Total Project Costs (US$ million) 1526.7 1224.7 80 Loan Amount (US$ million) 150.0 149.9 100 Date of Effectiveness 03/81 06/81 Date of Completion 06/84 06/86 Months for Implementation 39 60 153 Economic Rate of Return 22 Less than 5 Cumulative Estimated and Actual Disbursements (US$ millions) EY81 E. EM EY8 E.M EY8 Appraisal Estimate 14 66 121 150 Actual 7 47 60 89 106 150 Actual as % of Estimate 50 71 50 59 71 100 Project Dates Original Plan Actual First Mention in Files 04/79 Appraisal 04/80 05/80 Negotiations 10/80 11/80 Board Approval 12/80 12/80 Signing 02/81 02/81 Effectiveness 03/81 06/81 Closing 06/84 06/86 111 STAFF INPUTS (staff weeks) Y79 Y8I0 FY81 FY82 .84 Y8I5 FY86 FYa7 FY89 FY90 Total Pzeappraisal 12.8 66.7 79.4 Appraisal 32.8 15.3 48.1 Negotiation 11.8 Supervision 11.2 15.9 15.0 9.5 4.5 8.9 3.2 1.7 5.4 75.3 Other 10.4 8.6 19.0 Total 12.8 109.9 46.9 15.9 15.0 9.5 4.5 8.9 3.2 1.7 5.4 233.6 Mission Data Mission Type Month/Year No. Persons Staff-Days Idenitfication 05/79 3 42 Preparation I 11/79 3 30 Preparation II 04/80 5 70 Appraisal 05/80 5 95 Supervision I 05/81 2 20 Supervision II 09/81 2 20 Supervision III 06/82 2 20 Supervision IV 03/83 1 14 Supervision V 06/83 2 8 Supervision VI 02/84 6 36 Supervision VII 07/84 5 30 Supervision VIII 03/85 4 18 Supervision IX 05/85 4 20 Supervision X 08/85 1 2 Supervision XI 12/85 3 16 Supervision XII 05/86 3 -- Other Prolect Data Borrower: Banco Nacional de Obras y Servicios Publicos (BANOBRAS) Executing Agency: Ferrocarriles Nacionales de Mexico (NdeM) Follow-On Projects: Railways V (Ln 2575-HE) (1985) iv PROJECT PERFORMANCE AUDIT REPORT MEXICO FOURTH RAILWAY PROJECT (LOAN 1929-ME) EVALUATION SUMMARY Introduction Another reason for an expanded Since 1954, the Bank has effort is that recent Government supported five lending operations to initiatives in restructuring the rail subsector in Mexico. The financially troubled public First Project in 1954 was a loan of enterprises has raised the real US$ 61 million (about US$ 340 million possibility that a major in 1991 prices) to the Pacifico restructuring of the railways may be Railway. The most recent (1986) was in the offing. A careful look at the the First Railway Sector Loan of US$ Bank railway experience to date could 300 million (about US$ 456 million in be of use to those planning future 1991 prices) to Nacionales de Mexico operations to support restructuring (NdM). In the thirty-seven years initiatives. (2FAR Paras. 1-4) since the First Project, the Bank will have committed approximately Mexican Railla-a US$ 1.5 billion in 1991 prices for the rail subsector. This will be the The physical dimensions of FNM audit of the Fourth Project in the in 1991 is that of a system of 20,000 five-project series. standard gauge route kilometers employing 83,00 permanent staff to The intention here is to audit operate 1700 locomotives of US the Fourth Project in the context of manufacture out of approximately 40 the entire line of lending and maintenance shops. The locomotives associated transport sector work. pull trains made up from a fleet of This is because the last four lending 47,000 high capacity heavy axle load operations have been very similar in wagons similar to those used in the form, content and policy objectives US. In a Latin Amer1can context, and have constituted something Mexican Railways is roughly similar to a continuous line of comparable to the rail systems of credit for the Mexican rail subsector Argentina and Brazil. In a North from 1971 to the present. It is American context, Mexican Railways is reasonable, therefore, to evaluate comparable to a regional (as opposed the Fourth Project in the light of to a Class 1) railway such as tae larger experience. The fact that Conrail. In a world context it is the Project Completion Report (PCR) physically comparable to the rail and, subsequently, the audit for this systems of the larger European project was prepared almost five countries. It is in that group of years after project completion also medium size railways thnt cluster lends a longer than normal between the large high-traffic- perspective to the exercise - more density continental systems of China, like that of the impact studies that India, the US, and Canada and the OfD undertakes from time to time. pletnora of small, low-traffic- density rail systems that continue to V survive throughout the world. It is period of the time slice of a "serious" long-haul freight railway Investments under consideration. with a significant but not domiuant (PPAR Paras. 24-44) passenger operation. Project Objgativ-es An international comparison exercise reveals that the Mexican The objectives of the Railway rail system is lightly trafficked and IV lending operation were to (i) make overextended, substantially NdeM more efficient in its overmanned and maintains and operates operations, (Li) approach financial Its' locomotives poorly. Given the viability, and (III) provide long average hauls and good sufficient capacity to handle the technology, it is a potentially requirements of a 7.3 percent annual viable freight transportation growth in freLght traffic over the enterprise. At present it is a period 1981 to 1983. (PPAR Paras. 45- substantially more costly freight 46) operation than those of India or the U.S., but well below the costs of IM tAtion Aneriee European operations. The substantial passenger service is a major The 7.3 percent annual gro4ch contributor to the poor finances of in freight traffic over the period FNM. A rationalized Mexican railway 1981-1983 did not transpire. Freight deemphasizing passenger services and traffic for 1982 was 9 percent below placing increased emphasis on long- 1981 levels. For 1983 traffic wac haul domestic and internatianal back to the 1981 level. The ido freight operations would make it one then was to stretch out the period of the low-cost freight covered by Bank support from 1981-83 transportation operations in the to 1981-85. If the 9 percent decline world. (PPAR Paras. 5-23) In 1982 is ignored, freight traffic between 1981 and 1985 eventually grew World Bank Suort at an annual rate of about 1 percent. From 1985 to 1991, freight traffic The 1954 rail lending operation has subsequently declne at an (Railway I) was something of an annual rate of 5 percent. The anomaly in that it was for the projection in the SAR called for a relatively small regional Pacifico long term annual growth rate of 4.6 Railroad. It was sixteen years percent through the year 2000. The later that lending for the dominant stagnation and then secular decline Nacionales de Mexico (NdeM) in freight traffic for the period nationwide system began. Between 1981-91 presumably removed some of 1971 and 1991 the World Bank the urgency for provision of expanded supported a series of four lending capacity. operations to NdeM. Typically, all four lending operations included Despite the substantial support for: rolling stock and difference between actual traffic and locomotives, track renewal, bridge projected traffic, the hardware strengthening, telecommunications and component of the project was signalling, track maintenance Implemented much as specified except equipment, management information and for a five-year, rather than a three- car control systems, and studies. In year, Implementation period. For addition, all had a Plan of Action loconotLns, the actual procurement associated with them covering the was 317 now units rather than the vi originally planned 242 units, an retrospective economic analysis, but inexplicable development given che the sensitivity analysis in the SAR traffic trends and the target for indicates that a 50 percent reduction improved availability. Ee.ght in the assumed traffic growth rate wA&ons, however, were 8,300 new units (down to 3.3 percent per year) would rather than the planned 9,300. Tract yield w. ERR for motive Ror and rehabilitation targets were also Ags, of about 12 percent. Since exceeded with 938 km. of new rail the actual traffic growth rate for rather than the planned 780 km. the last ten years has been stagnant Bridge strengthening was applied to or declining, it Is reasonable to 857 bridges rather than the planned assume that the retrospective ERR for 510. Only 612 km of C[T investmenta this major component af the project, were made rather than the planned 960 using the SARs assumptions, would km. now well below 12 percent and not much above zero. Similarly, a 25 The software components of the percent reduction in traffic growth project were not implemented. The rates in the sensitivity analysis component of the loan for foreign reduced the expected ERR of trac consultants to assist in introducing reDMI Investments in the SAR from modern computeri. 3d planning and 17 to 14 percent. A zero to negative operational methods was not used growth rate for ten years would lower because of "reticence" on the part of the ERR to well below 10 percent even NdeM operational management. The under the favorable (and unrealistic) assertion in the PCR that much of assumptions of the SAR. Given that this work was *effectively done" jy these components accounted for about NdeM staff was not verified by the 96 percent of the total project, the Audit field visit. Audit concludes that the Railway IV Investment program was not an The project objective of economic success. (PPAR Paras. 51- approaching financial viability, as 57) measured by the Operating Ratio, was achieved despite the poor traffic FindJUS ad Issues picture. During a period of strong inflation, substantial real tariff The poor performance of the increases were put into place while Mexican economy between 1981 and 1991 real wages declined.. The objective has had a serious impact on the of more efficient rail operations was demand for rail freight transport. not achieved. Locomotive This factor alone largely explains availability declined substantially the generally successful use of as did wagon productivity. (PPAR investment resources under Railway II Paras. 47-50) and Railway 111 (1972-79) and the generally pvor yield of investments Proect Economics under Railway IV and Railway V (1981- 88). No analyst could be expected to A critical analysis of the have prepared a forecast of ten years investments actually made under of stagnant/declining traffic as the Railway IV indicates that a great basis of an FN investment plan for deal of the program was not based on 1981-91 since no macroeconomic adequate empirical evidence and/or projections to support such a view was clearly premature. The Audit were ever provided. The question does not have the detailed then needs to be asked as to why, in information to make a comprehensive the face of declining traffic, the vii procurement of surplus locomotives have a major impact on labor. and wagons continued. For locomotives, part of the answer was The potential for major cost long lead times and the previous savings through disinvestment and commitments to buy from foreign down4izing is enormous. Over 96 manufacturers. For wagons the chief percent of the freight traffic is reason was the need to keep the confined to 13,000 km of the 20,400 government-owned wagon manufacturing km system. This traffic could be facilities operating during a period handled with 800-900 locomotives of extreme domestic economic operating out of no more than five distress. Had the wagons been maintenance shops. The labor force manufactured abroad, the procurement required wculd be 20-30,000 would almost certainly have been cut employees. Ins- ,ad, FNM empluys over back. 80, 100 empl-ees, and operates 1700 locomotives out of about 40 The generally poor action plan maintenance shops. FM must maintain achievements under Railway IV and 7,000 km of track that carry Railway V can also be largely virtually no traffic, and hundreds of attributed to the much less than marshalling yards and freight expected traffic without a stations that have had no function commensurate decline in locomotives for the last twenty years. and rolling stock procurement. Idle capital is reflected in the declining The surplus infrastructure does productivity rates. The steady much more than add to the FNM increase in the rail labor force over maintenance burden. A principal the same period of economic distress reason for the low locomotive is also understandable given the very availability figures is the operation high unemployment rates during the of about 40 locomotive maintenance recessionary period. As in the case facilities (a legacy of the steam of the wagon manufacturers, the era) where only about five are pressure to maintain or expand public needed. All are seeking to justify employment during an economic crisis their existence by working on costly is difficult to resist. diesel locomotives. Hence, the railway adage that "workshops eat The brief period of financial locomotives". The other half of the improvement under Railway IV has adage states that "yards eat trains". passed. The simple expedient of net- Given the increase in unit trains and revenue-generating real increases in the disappearance of LCL traffic, freight rates is increasingly tht-e is no longer any need for most problematical given the simultaneous ot -he marshalling yards that remain deregulation of long haul trucking in the EM system. The most and the down sizing or --rivatization effective initiative for improving of previous governmer.- monopolies average train spee&- and train (steel, foodgrains and fertilizer) productivity is to disinvest in that had accounted for a certain marshalling yards. Every time a amount of guaranteed traffic for the train enters a marphalling yard for railroads. Railway tariff increases any reason a large block of time is must now be considered in the light lost. Keeping locomotives out of of potential traffic diversion or workshops and trains out of yards is cessation. This raises the much more the low cost way to rapid gains in difficult issue of cost cutting and productivity. One way to achieve efficiency mepsures, most of which this is to dispense with the surplus viii facilities and labor working rules government policies are as much a that consume capital in an effort to part of the rail environient as the justify the existence of a large physical terrain between Mexico City amount of redundant labor. and the seaports. As long as it was poussible to Justify rail ir.vestment3 The cost of a single Mexican that avoided uneconomic diversion of rail employee is about US$ 4000 per traffic to highways (Railway II and year. The disappearance of say III), the "second best" argument had 40,000 workers would save the some merit. However, once the railroad about US$ 160 million per traffic began to decline for reasons year, solving about half the other than adequate rail transport financial problem. The other half capacity (Railway IV and V), the would feasibly come from the down- traditional Bank-supported sizing of the infrastructure - investments in track renewal and workshops, lines, yards and passenger rolling stock lost their equipment and facilities. The two justification. The debt incurred sources of savings are not under Railway IV and V then became independent, however. A great deal part of the problem rather thcn the of Mexican rail labor (except for the solution. (PPAR Paras. 58-71) small amount of absentee workers) only becomes completely redundant L2 forLthe-1qture when downsizing and disinvestment are effected and work rules are changed. Increasingly the Bank is Current operating procedures and work involved in programs of restructuring rules are effectively designed to technically bankrupt and inefficient keep a large work force and its public enterprises such as FNM. It associated capital underemployed. dominates the scene in operations in the new Eastern European member With the notable exception of countries. How fast can the viable the very large loan to the del enterprise be extracted from the Pacifico in 1954, the Bank has been bankrupt one and what sort of "actively" supporting new technology ownership structure and policy and operational changes while environment will it operate in? "pass4vely" addressing the issue of work rules and downsizing. This was Fortunately for Mexico, the because the Government wished to public sector is not nearly as large "modernize" the railroad without a share of the economy as is the case addressing the thorny issue of work in most of the former Communist rules or redundant labor. As a countries. The distortions have not result, the perfectly sound been so great and the basic future operational changes supported by the policy environment of competition and Bank have not been adopted and the free pricing already exists. Also, technology investments have not been the rail mods in Mexico is not nearly fully utilized. as important in the total sector as is the case in Eastern The question can then be asked Europe. FNM is also well-endowed whether or not the Bank should have with adequate and modern supported, with four lending infrastructure and equipment thanks operations over twenty years what, in to ten years of large investment retrospect, was a "second best" programs and stagnant to declining approach to rail reform. The case traffic. This latter factor means can be made that deeply rooted that the instment reuirements ot ix the FNM restructuring effort will not What is particularly worrisome beat this stage is that, despite an explicit recognition of the poor This raises the possibility operational and financial performance that in terms of lending operations of FM and the need for doing things it may be most appropriate for the very differently in the future, the Bank to do nothing at this stage of Structural Change Program continues the process. This is because the to include very large sums for fundamental pressure for change must investments for the sort of general come from within the Government - not modernization and improvements that the Bank. And, given the experience were included under the previous four in other countries, the origin of the investment programs supported by the pressure for reform is . the Bank. As pointed out in Section VI intolerable financial cost of an of this report, the track and bridge unreformed railway. Action plans and CTC components are highly suspect implying structural change simply do investments. The large component for not get pursued seriously if the completing ambitious and costly SCT- political will is not there to take initiated rail construction projects extremely unpleasant initiatives. needs to be scrutinized very These initiatives cannot be taken carefully. It isn't enough to without a great deal of preparation continue to put money into new in which the political, human, construction efforts simply because technical and financial elements of so much has already been invested. reform are developed in considerable Painful as it may be, an analysis detail. Things have not yet reached assuming sunk costs may show that it this stage in Mexico. is better to defer further investment indefinitely. The clear message of A start is being made with the this component of the Plan is that an proposed 1991 Structural Change all- important consensus on the need Program from the management of FNM. for major rail reforms in Mexico has It indicates that the costs of doing probably not yet been reached. A business as usual have become nearly substantial contingent of the rail intolerable. The need for a drastic renaissance element remains in SCT reduction in the rail labor force is and FNM and any future upturn in now recognized explicitly with a traffic and revenues could easily put proposal to cut 28,100 jobs from the this element back at the center of 1991 level of 83,900. This would national rail policy. leave a labor force of 55,800 by 1995. By the year 2000 the numbers Flowing from the foregoing this would be as low as 45,000 - a 46 audit offers the following percant reduction in the total rail recommendations on future Bank labor force. It is now acceptable lending for railways in Mexico. within Mexico and within the Bank to discuss this most crucial element of First, any new Bank loan for solving the railway problem. the Mexican railways should be However, the important related contingent upon problems of work rule changes and management of the Operating (i) implementation, prior to Board Department have not yet received the presentation, of a number of attention they must have if movement substantial steps and measures is to take place. to rastructure FMN which should have been discussed and agreed with the Bank, and x (ii) full agreement, between the the railroad (disinvestment, staff Bank and the Government, prior reduction) and key operational and to the Board presentation, on a policy changes. detailed and comprehensive program and timetable for the Third, and finally, the audit completion of the restructuring feel that any new Bank railway of FNM during project project should make adequate implementation. (generous) allowance for Bank staff time to enable careful and frequent Second, in addition, the size monitoring of the implementation of any new Bank loan could (should) progress of the restructuring program be considerably less than that of (including a comprehensive annual previous railway projects given the review) during project execution. emphasis to be placed on downsizing PROJECT PERFORMANCE AUDIT REPORT MEXICO FOURTH RAILWAY PROJECT (Loan 1929-ME) I. Introduction 1. Since 1954, the Bank has supported five lending operations to the railway subsector in Mexico. The First Project in 1954 was a loan of US$ 61 million (about US$ 340 million in 1991 prices) to the Pacifico Railway. The most recent (1986) was the First Railway Sector Loan of US$ 300 million (about US$ 456 million in 1991 prices) to Nacionales de 11exico (NdM)../ In the thirty-seven years since the First Project, the Bank will have committed approximately US$ 1.5 billion in 1991 prices for the railway subsector. This will be the audit of the Fourth Project in the five-project series. 2. The intention here is to audit the Fourth Project in the context of the entire line of lending and associated transport sector work. This is because the last four lending operations have been very similar in form, content and policy objectives and have constituted something similar to a continuous line of credit for the Mexican railway subsector from 1971 to the present. It is reasonable, therefore, to evaluate the Fourth Project in the light of the larger experience. The fact that the Project Completion Report (PCR) and, subsequently, the audit for this project was prepared almost five years after project completion also lends a longer than normal perspective to the exercise - more like that of the impact studies that OED undertakes from time to time. 3. The second reason for OED's larger than normal input into this post evaluation exercise is that it is to form a portion of the planned Bank\Mexico Relations Study under the direction of a multi-sector OED team. The Study will cover the Bank's experience with Mexico in all sectors, of which Transport will be one. The railway subsector element of this work can be done both naturally and economically in the context of an expanded railway audit. 4. The final reason for the expanded effort is that recent Government initiatives in restructuring financially troubled public enterprises has raised the real possibility that a major restructuring of the railways may be in the offing. A careful look at the Bank railway experience to date could be of use to those planning future operations to support restructuring initiatives. To serve these broader objectives it was necessary to develop the background material that would allow the Fourth Project to be seen in the broader historical context of which it is a relatively small but important part. 1/ An important nomenclature used in this report is that of NdeN, del Pacifico, and FN. Bef re its recent consolidation into FW, Ndex was the dominant national tilway system accounting for 76 percent of national route mileage. The del Pacifico was an important regional system accounting for 12 percent of national route mileage. The remaining 10 percent of route mileage was made up of minor feeder lines. As of 1990, railway statistics of the previously separate administrative entities have been consolidated into one set of statistics for the entire FM system. 2 II. Mexican Railways 5. In approaching the problems and potential of Mexican National Railways (FNM) in 1991, it is essential to have some familiarity with the history of over 100 years of development and operation and the recurrent themes that run through that history from the late 19th Century to the present.j/ These themes are: (i) geography and geopolitics, (ii) modal competition, (iii) the power of labor, and (iv) financial crises. The System Today 6. The physical dimensions of FNM in 1991 is that of a system of 20,000 standard gauge route kilometers employing 83,000 permanent staff to operate 1700 locomotives of US manufacture out of approximately 40 maintenance shops. The locomotives pull trains made up from a fleet of 47,000 high capacity heavy axle load wagons similar to those used in the US. In a Latin American context, Mexican Railways is roughly comparable to the railway systems of Argentina and Brazil. Together these three stand out as the dominant systems in the region. In a North American context, Mexican Railways is comparable to a regional (as opposed to a Class I) railway such as Conrail. In a world context it is physically comparable to the railway systems of the larger European countries. It is in that group of medium size railways that cluster between the large high- traffic-density continental systems of China, India, the US, and Canada and the plethora of small, low-traffic-density railway systems that continue to survive throughout the world. It is a "serious" long-haul freight railway with a significant but not dominant passenger operation. The characteristics that distinguish it from the rest of the world's railroads will emerge in course of this report. Geography and Geopolitics 7. Mexico City, with a current population of close to 18 million is located on the high central 1lateau at an altitude of about 2300 meters. The principal seaports of Vera Cruz and Tampico on the Gulf of Mexico and Lazero Cardenas and Manzanillo on the Pacific cannot be reached without passing through extremely mountainous terrain over relatively short distances (Vera Cruz on the Gulf is 310 km by air from Mexico City, while Lazero Cardenas on the Pacific is 350 air km.). Getting to the principal sea ports by ground transport from Mexico City is df fficult but especially difficult for railway where eleven degree curves and 3.4 percent gradients are common.11 Travelling north along the central spine of the country is less arduous. The air distances to the northern border are long (1750 km to Nogales, 1520 km to Ciudad Juarez and 900 km to Nuevo Laredo) and the drop in elevation less precipitous (1300 meters elevation at Nogales). As a general rule, railways are an effective means of transport where the gradients and curvatures are gentle and the haul distances great while 11 Unless otherwise stated, the historical information is based ont Los Ferrocarriles de fexic 1837-1987, Ferrocarriles Nacionales do Mexico, Mexico D., 1987. In one 60 km. section of the railway line from Lasero Cardenas the line rises 600 meters limiting the trailing load of a train to only 610 tons. Sees Lazero Cardenase-Central Mexico Corridor Study, Report No. 7964-ME, June 30, 1989, The World Bank, Washington, D.C. 3 trucking technology has the edge with shorter hauls, steeper gradients and sharper curves. The geography of the country therefore gives a special advantage to railway technology for the Mexico City-northern border traffic while road transport has the advantage on the Mexico City-seaport traffic. 8. The early investors in railways in Mexico did not have road transport to contend with and the first projects, largely supported by British capital, were aimed at ocean-to-ocean and Mexico City-to-seaport traffic. The early geopolitics were European oriented. This changed as the century progressed and the emphasis shifted to American investment capital and the extraction and transport of Mexican raw materials for the US market. This lati-er emphasis led to the development of the north-south axis of the national system. The Mexican railroads and the raw materials that moved over them became an important element of the US industrial effort through the First World War. Partial ownership and control of Mexican railways and natural resources by US investors during a period of generally uncontrolled monopolistic industrial activity in the US was of great concern to Mexico. A sensitivity to the extent and nature of trade and transport between the two countries developed early on. This understandable concern on the part of the Mexicans for the economic and political power of their most geographically natural trading partner has had, and continues to have, a great influence on trade and associated transportation. 9. The geography of Mexico has remained largely unchanged, but the geopolitics of trade have changed dramatically since the turn of the century. Anti trust legislation in the US, the rise of trucking, and the decline in the economic and political power of the US railways have changed the environment. The US railways have become reasonably efficient transportation enterprises while raw materials from Mexico are of less interest to US importers. Low cost Mexican industrial labor and agricultural commodities are attractive to US interests as access to US markets and expanded domestic employment opportunities is attractive to Mexican interests.i/ The recently proposed Free Trade Initiative (FTI) between Mexico, the USA and Canada is an attempt to expand trade and its benefits systematically between the three geographically contiguous countries. 10. For both cultural and ideological reasons, the railways of Mexico identified strongly with the railroads of Europe rather than those north of the border.l The most obvious manifestation of this leaning were the investments in railway line improvements to the seaports, electrification, and subsidized passenger services during a period when the US railroads were emphasizing disinvestment in much freight and passenger infrastructure except for that required for heavy axle-load and intermodal freight movements. The FTI has made it both possible and necessary for FNM to expand its relations with the railroads north of the border. A recent operational initiative has been the introduction of double-stack container trains between Mexico City and Chicago, the first such use of this technology by a developing world railway. FNM has also contracted with Union Pacific Railroad (UP) to install and use its most recent software for Al The foreign trade cZ Mexico In 1990 was 80 percent by value with the USA. Mexico has been a dues-paying member of the Association of Averican Railroade for many years but has not participated actively in their meetings until the last five years. 4 car control and management information through a fiber optic cable connection between UP and FNM. New line construction, electrification and expanded passenger services have begun to fade into the background in the light of the attempt to make FNM a modern freight transportation enterprise with an appropriate emphasis on the long north-south hauls. Modal Competition 11. In 1873, the first important railway line, Mexico City - Vera Cruz, was completed. Since this was the only mechanized mode of transport at the time, it quickly became a monopoly operation. However, between 1928 and 1934 the Ford Motor Company began assembling automobiles in Mexico. By 1949 there were estimated to be 275,000 vehicles in the country of which 106,000 were trucks. This rapidly growing vehicle fleet was operating on a road network of about 26,000 km in 1955. The earliest evidence we have of actual modal splits of intercity freight traffic in ton-km is for 1948 and 1956. In 1948, it was estimated that the highway share was only 11 percent of total intercity ten-km. This had increased to 46 percent by 1956.6/ The most recent (1990) estimate has highways at 65 percent, railway 21 percent, coastal shipping 10 percent and pipelines 5 percent.Z/ Thus, the railway share of intercity freight ton-km has declined steadily from about 100 percent in the early 1900s to about 21 percent in 1990. The now dominant highway mode rose from virtually nothing in 1900 to 65 percent of total intercity freight flows. 12. One of the structural reforms undertaken by the Mexican government under the recent FTI has been the deregulation of long haul trucking, including international movements. Expanding foreign trade with the US and Canada requires efficient land transportation links. The challenge for FNM is how to improve its international operations to maintain and perhaps expand its share of this particular traffic.8i Failure to do this will put an enormous burden on the Mexican highway system. The economics clearly favor the long haul intermodal railway operation but until recently this has not been in the forefront of the concerns of railway planners in Mexico. The Power of Labor 13. Until 1908, the railway system of Mexico was made up of numerous concessions granted by the government to local and foreign groups. These groups had financed the construction of independent lines, some of which were in direct A/ An Appraisal of the Development Prosram of Mexico, Report No. Wd-37a, July 13, 1964, The World Bank, Washington, D.C., Vol. 5, p. 2-5. These estimates, made available to the Audit, are based on a recent confidential study for the Ministry of Comaunications and Transport. The PCR for Railway IV asserts, without giving sources, that the railway share has dropped to 14 percent. (PCR Railway IV, para. 7.01, pp. 9-10.) It is difficult to see how even an efficient deregulated trucking industry, US or Mexican, could possibly compete effectively with a long-haul double stack international rail container operation, so almost any reasonable attempt by FNM will capture this traffic. For the shorter haul domestic railway container operation only a well-organized operation will be able to compete with trucking. 5 competition.21 In 1908, the government fused a number of these companies together into Nacionales de Mexico (NdeM), issuing stock in the new company to the mostly foreign owners of the smaller entities and retaining 50 percent of the shares for the Mexican Government. With 11,157 km of integrated network, a formidable monopoly was formed. A strike by railway labor after this time could paralyze an economy that had no other significant mode of transport. In 1926 a general strike did just that. By 1933 a 47,000 member union (STFRM) was formed and sometimes violent confrontations took place with the police. 14. By 1937-38 the railway union had become an important political support group. The government took over the operation of all the railroads and turned them over to the unions to operate, one of the early experiments in worker management. It was not a happy experience since wage increases were immediately awarded and the financial obligation that went with the labor management agreement (an operating ratio of .85) could not be met. By 1940, the experiment had ended. As late as 1959, a general railway strike was carried out requiring the intervention of the army. Since then a peace based on extremely detailed labor agreements between the railroads and the union has prevailed. Work rules have been codified and labor has a major say as to how the senior management positions of the railway are to be filled. A large residue of worker control remains even with the relative demise of the railroad in the overall transport sector.,1/ Financial Crises 15. At the end of the Mexican Revolution (1911-1920) the railroads emerged into a financial crisis. Thirty one percent of the external debt of Mexico at that time was attributed to the railroads. In 1922 and again in 1924 cnferences were held to reschedule the debt. The Government of Mexico eventually agreed to assume most of the debt obligations. An Efficiency Commission was set up to examine wage costs, other expenses and tariffs in an attempt to set up a viable enterprise with 51 percent government ownership. 16. The notion of a private, largely foreign-financed company, with majority government ownership and a powerful labor unions generated many contradictions. The profitability margin declined from 39 percent in 1909-1911 to 6 percent in 1927. The "problem" in 1927 was said to be an excessively large labor force, large salary increases and money losing lines. By 1937 the fiction of a private company was abandoned and the government took over all the assets of NdeM, as well as the debt, and the railway became a public utility. Studies of the "railway problem" were carried out by foreign consultants in 1928, 1929, 1930 and 1948. The "solution" proposed was to reduce personnel, cut wages, close down money-losing operations and otherwise "rationalize" the system. The prescription of sixty years ago is remarkably similar to that being set forth in the national financial crisis of 1982-1990. 21 This explains the common existence of alternative and sometimes redundant lines in the N system. 10/ Eleven years prior to 1988, the President of the Railroad was the President of the rail workers* union. 6 17. Between 1950 and 1974, Mexico enjoyed a remarkable period of high growth, low inflation and moderate external debt accumulation. Real growth averaged 6.4 percent per year and inflation was in single digits throughout the period. This came to an abrupt end in the early 1970s. Rapidly expanding government involvement in the economy pushed up the rate of growth. However, increasing government expenditure was not matched by rising public sector revenues. The inflation tax and external debt became increasingly important sources of finance. A serious but comparatively brief, financial and economic crisis in 1976 terminated following wajor oil discoveries in 1977. The ensuing prosperity lasted until 1982 when the Mexican Government suspended payment of interest on external debt. 18. Over the 1982-90 period, economic growth ground to a virtual halt. The matter was made worse by the collapse of oil prices in 1986. The eight year period of little or no economic growth in the face of a rapidly increasing population and labor force has led to major structural reforms in the Mexican political economy. The intention now is to promote private sector investment and back away from public sector led growth financed through debt. The railways, as a major public sector employer and source of financial losses and public debt, could not escape the scrutiny of those promoting the new policies.I/ 19. Major restructuring of the railroads of the US in the 1960s and 70s resulted from massive bankruptcies. The legal system and bankruptcy law did what economists and transportition planners could not. Similar forces have been operating in Mexico from 1982 on. The relative decline of the railroad in the transport sector, the deregulation of long-haul and international trucking and the restructuring of other public monopolies that had guaranteed traffic to the railroad made it impossible to solve the railway financial problems by simply raising tariffs. Attempting to do so would drive the traffic to other modes or render the shipments unprofitable for the shipper. Restructuring of some sort has become unavoidable. Current Productivity and International Comarisons 20. The few comparisons made between Mexican railroads and other railroads in the past has involved references to the railroads of Europe. This may have been appropriate at one time, but the evolution of Mexican railway to long hauls and unit trains, with equipment compatible with that of the US and Canada, makes it more appropriate to examine the productivity of the Mexican operation in the context of other high-volume, long-haul freight operations.g/ For purposes of this exercise, comparisons are made in Annex 1 with the French, German and Spanish systems, the Conrail system of the US, and the systems of India and China. Conrail is similar to the Mexican freight system in extent and technology and has undergone considerable restructuring in the last ten years. India's is a larger, higher density system with a dominant passenger operation and a very large redundant labor force. China is included simply because it 1, Excerpted from Mexico Road Transport and Telecoumnicatione Sector Adlustment Projects (Part I - The Economy), Report No. P-5254-ME, May 3, 1990, The World Bank, Washington, D.C. pp. 3-7. U/ Average freight hauls for French and Spanish railways are In the 300-350 km range whilst those for Mexico, Conrail, China and India are in the 600-750 km. range. Average freight haul for US Class I railways is about 1100 I=. 7 gives some notion of what maximum physical productivity levels are achievable in a resource constrained environment. 21. The international comparison exercise indicates that FNM has the lowest network productivity measure of all the systems indicating that it is an overextended system despite its low density geographic coverage. This squares with the fact that 96 percent of the current freight traffic moves on the "basic" 12,800 km of the 20,400 km system. This also squares with the fact that the maximum freight density on the system is only 17 million gross tons (MGT) per year on one 120 km section of track. With the technology available to the Mexicans, 100 MGT per single track is achievable. An unusually low locomotive availability figure (around 60 percent compared to 80-90 percent for all other systems) prevents the achievement of otherwise quite respectable locomotive productivity figures. Finally, the passenger service of Mexico has, by far, the lowest revenue contribution figure of all the railroads on the system indicating that the railway passenger subsidy policy is one of the most durable and costly policies of the Mexican Government. 22. More neutral findings are that Mexicanwagon productivity figures are near those of the US, though far below Indian Railways..U/ Mexican labor productivity exceeds that of all the European railway systems and that of India. It is, however about one eighth that of the US Conrail system. The Mexican railway wage bill as a share of total revenues has fluctuated between the high labor cost European systems and the low labor cost US and Asian systems. 23. On the positive side, FNM has the lowest share of passenger output to total output of all the systems except Conrail, maintaining its emphasis on freight operations as its raison d'etre. This is important because the other positive finding is that Mexican railway freight unit revenues (and probably costs) are much lower than European unit freight revenues and closer (though still about 50 percent higher) to US and Indian unit freight revenue figures. Given the very great potential for cost cutting and efficiency in FNM, this means that FNM has the potential for being one of the lowest cost freight operations in the world. III. World Bank Support 24. Bank support for Mexican railways falls into two distinct periods, early lending and sector work in the period 1954 to 1971 and a regular series of loans for the dominant Nacionales de Mexico (NdeM) in the subsequent twenty years to the present. The Early Years 1954-1971 25. In 1954, the Bank made its First Railway Loan (Railway I) to Mexico (Ln 103-ME). It was US$ 61.0 millions for the rehabilftation of the 2,000 km. L/ The Region point out that in tems of ton-h per wago per year that the Hezican perfomance is well below that of the US. The Audit's wagon productivity estimates are in ton-ho per ton of wao sapacity per year. This taes accmnt of the smaller average wagoa capacity of the Meican fleet. See footnote 6 of Annex I for more an this Issue. 8 Pacifico Railroad serving the Pacific coast region from Nogales on the border to Guadalajara. In 1991 prices that would be a loan of US$ 341 millions, an enormous investment in such a small property. The Pacifico had belonged to the American-owned Southern Pacific Railroad (which had purchased it in 1909) until 1951 when it was purchased for US$ 12 millions by the newly-formed and Mexican- owned Pacific Railroad company. The objectives were to modernize (dieselize), rebuild track, cut back on redundant labor, reorganize the management information system, train managers and supervisors, and make sure that regular rate increases were in place to protect the financial viability of the railroad. The loan covenants included inter alia: (i) twenty key supervisory posts that previously had to be filled from the labor unions would now be filled from non union sources,L, and (ii) reduction in personnel to levels requized for running a modern railway. Significantly, no estimates were made of the extent of che redundant labor problem. Nowhere in the PAR is the size of the actual labor force mentioned. At least half the report and virtually all the analysis was devoted to financial matters.15/ 26. No Project Completion Report was ever done for Railway I, but there is a brief statement in the SAR of Railway II: "...agreed financial targets (for Railway I) were not achieved, mainly because labor costs increased faster than revenues." 27. In 1964 the Bank did its First Transport Sector Report as part of a larger effort to appraise the development program of Mexico for the period 1963- 65.16/ It was to be the largest such effort ever mounted by the Bank in the transport sector. It documented the rapidly declining share of railway in freight and especially passengers, the disappearance of less-than-car-load (LCL) traffic to trucks and the rapid evolution of the railway to a long hauler of bulk commodities.IZ/ The major "modernization" effort involved a shift out of steam and into diesel locomotives between 1952 and 1961. The other modernization effort was in telecommunications. The idea was to replace the old telegraph line wire equipment with telephones finally linking the entire system to Mexico City. The sector mission also pointed out how the large proposed investments in Central Traffic Control (CTC) were uncalled for. Much cheaper modern signals and telephone dispatching would do the job given the relatively low traffic volumes. is/ 14/ This was to be the first and last time that labor agreements were negotiated as part of a Bank lending operation. 151 Appraisal of the Rehabilit tion Program for the Pacific Railroad of Mexico, Report No. T.O. 58-a, August 10, 1954, DepartmeAt of Technical Operations. The World Bank, Washington, D.C. 16/ An Aipraisal of the Development Program of Mexico, Seven Volumes, Report No. WH-137a, The World Bank, Washington, D.C., July 13, 1964. See in Particular Volume V, Annex IV - Transportation. IZ/ By 1959, about 36 of the 2,000 freight stations on the system accounted for 74 percent of the tonnage shipped and 61 percent of the tonnage received. 18/ This was the first and last time that the appropriateness of CTC investments was questioned by the Bank. This is an important point that is addressed again in the Findings and Issues section of this report. 9 28. The sector mission also pointed out that the shift in national freight shares from railway to trucking has been rapid and appropriate and that the railroad system, in spite of nagging deficits, was reasonably well run and had a long term specialized future as a freight carrier. A careful analysis of the deficit (US$ 200 million in 1991 prices) showed that the major sources of the loss were: (i) a passenger service that no longer made economic sense, (ii) at least 2600 km of hopelessly uneconomic branch lines, and (iii) railway freight rates for some commodities that were far below costs and could be raised without losing the traffic to trucks. Based on ratios from the railways of Europe and Japan for 1960, there were not found to be any egregious problems with overmanning..9/ 29. In 1971, the Bank produced a Second Transport Sector Report as a prelude to a major railway lending operation.20/ By that time the railway deficit was getting worse. About 80 percent of it was because of passenger services which was estimated to represent 47 percent of railway effort and only 14 percent of revenues. Three new factors entered the picture with this report: (i) labor surpluses, (ii) locomotive availability and utilization, and (iii) foreign wagons. Surplus labor was being perceived to be a problem even by railway management. Tha estimate then was that a 17 percent reduction was called for. Diesel locomotiv. availability was good (89-93 percent) but utilization was not good. Only fifty percent of the wagons used on the system were railway property - the others were divided evenly between private Mexican (PEMEX wagons) and those of US railroads. 30. In order to solve the dominant financial problem in 1971, the recommendation was to (i) formulate a National Transport Policy, (ii) define the role of the railways, and (iii) set up a corporate plan including a "railway philosophy" and explicit subsidies from the government. Railway management was said to be too much operations oriented and too little business or market oriented. Without a corporate plan the performance of railway management could not be judged nor could the social objectives of the government railway policy (subsidized passenger services and uneconomic branch lines) be separated from the economic (high- density long-haul freight). Nacionales de Mexico 1971-1991 31. The 1954 railway lending operation (Railway I) was something of an anomaly in that it was for the relatively small regional Pacifico Railroad. It was sixteen years later that lending for the dominant Nacionales de Mexico (NdeM) nationwide system began. Between 1971 and 1991 the World Bank supported a series of four lending operations to NdeM. The first three (Railway II, III and IV) 9/ Interestingly, the major anticipated benefit of dieselization was labor savings (56 Z) as opposed to fuel savings (212). Savings in freight train kilometers was the other 24 K (p. 30 WH-137a). From 1952 to 1961 the combined national railway labor force actually declined from 77,649 to 69,739 a reduction of almost 8,000 men. By 1976 the labor force was back to the 1952 levels. Since then it has continued to grow steadily to 83,290 in 1990. 20/ The Transport Sector of Mexico, (in four volumes), Report No. PTR-88, May 13, 1971, The World Bank, Washington, D.C. See Volume 11 for railways. 10 'ere, with minor variations, virtually identical in form and content.IL/ The most recent (Railway V) was referred to as a sector loan while the others were project loans.1I This was because, whereas Railway II, III, and IV included only the non construction investments of NdeM, the new line construction investments of the SCT were included as well in Railway V.2I Typically, all four lending operations included support for: rolling stock and locomotives, track renewal, bridge strengthening, telecommunications and signalling, track maintenance equipment, management information and car control systems, and studies. In addition, all four lending operations had a Plan of Action associated with them covering the period of the time slice of investments under consideration. The basic elements of these action plans, as extracted from the Staff Appraisal Reports, are set forth in Table 1. 32. An examination of the action plan obiectives for the last four lending operations gives a good picture of what has transpired in the railroads from the early 1970s to the present. The most obvious (and in the end the most critical) problem of financial viability received consistent and explicit attention. Over the 16 year period of the action plans the actual Operating Ratio did improve periodically but not for long as costs outstripped tariff and traffic increases. The recurring financial crises have periodically forced the government to take steps but the steps taken have never been structural in that the mechanism has never been put into place that would assure financial viability continuously. 33. The action plans of the last three lending operations avoided having quantitative goals for labor. Instead, studies, plans and discussions were to supplant quantitative targets. The number of staff for all Mexican railroads increased from 74,000 in 1972 to 83,900 in 1990, a period in which most other modernizing developed country railroads of the world were registering substantial declines in their labor forces.241 11/ Anraisal of a Second Railway Proiect Nexico (Railway I), Report No. PTR-107a, May 10, 1972, and Mexicos Appraors of a Third Railway Project(tailway 111), Report No. 957a-ME, March 15, 1976 and Fourth - vay Project Mexico (Railway IV), Report go. 3078b-ME, Nov. 20, 1980, The World Bank, Washingcon, D.C. 22/ Mexico Railway Sector Project (Railway V), Report No. 5464b-MIE, May 14, 1985, The World Bank, Washington, D.C. / An important administrative distinction up until 1988 was the senaration of planning and design of major civil engineering works in a separate department (V a Perreas) In SCT. This organisation would plan, design and contract for works that were then turned over to the railroads for their use and subsequent maintenance. Some notably poor investment decisions were made under this arrangement, of which more will be said later. The Resoo notes that another distinct characteristic of the sector loan was the proviston for Bank/Goveramnt annual reviews of the Investment program. 24/ The railroads of China and India were registering modest increases in the labor force but, unlike Mexico, ware registering substantial Increases in traffic over the same period. 11 ACTION PLAN MEASURES Railway II-Railway 1 Showing Actuals at the Beginning of the Period and Targets at the End of the Period .-... ---.----------.......-----.--...------------------------------- Item Uail- Rail- Rail- Rail- Actual (Year) way II way III way 1V way V -------------------------------------------------w........ ...... Period 1972-76 1976-79 1981-85 1984-88 No. Staff2_5 74k-74k Study Study Study 83k (1990) Op. Ratio 1.4-1 2 1.3-1.1 1.4-1.1 1.1-1.0 1.5 (1990) Loco-km/yr 90k-100k 87k-100k 85k-105k 95k-100k Loco Avail (%) 85-85 85-86 77-81 68-80 63 (1988) Wagon Turn- N.A. 15.5- 12.4- 18.0- Around (days) N.A. 13.7 11.5 15.0 Ton-km/serv. 1450- 1700- 2100- 1700- 1573 (1990) car-day 1750 2050 2500 1900 No. Foreign N.A. 9000- 10000- 3500- 6830 (1990) Cars on Line N.A. 7100 7000 3500 AACRG Freight 4.7 7.8 7.3 5.0- 8.9 (1972-76) Traffic26 3.0 3 . 1 (1976-79) Assumed (%) 1.1 (1981-85) -2.0 (1984-88) -----------------...----------------- -------------..t---------- Sources: Railway II - SAR, Annex 7, p. 1, Report Number PTRO107, dated May 1972. Railway III - SAR, Annex 6, p. 1, Report Number 0957, dated March 1976. Railway IV - SAR, p. 25, Report Number 307R, dated November 1980. Railway V - SAR, p. 23, Report Number 5.944, dated May 1985. fl/ The actual maning figure used in the Rail II action plan was 56,000 for the deN system only. The idea was to hold the labor force constant over the plan period. The Audit has eubstituted the 1972 level of all the railroads in order to ake the 1990 actual figures for M meaningful. I&I AACR - Average Annual Compound Rate of Growth. 12 34. Locomotive productivity in annual mileage declined steadily through 1985 and the elusive :,oal of 100,000 km per year has yet to be achieved.IZ/ Locomotive availability has declined dramatically from a respectable 85 percent in the early 1970s to a poor 63 percent in 1988 Wagon producti,.ity improved steadily through 1981 and then declined dramatically. 35. Traffic for the Railway II period (1972-76) was greatly underestimated but for all subsequent loans it was greatly overestimated. Traffic, which was growing at close to 9 percent per annum for Railway II dropped to 3 percent growth for Railway III, 1 percent for Railway IV and finally a negative 2 percent for Railway V. 36. The general notion of associating an action plan with a lending operation to effect desirable change in a financially and operationally troubled railroad has been given a sustained trial with Railway II through Railway V. Without going into specific reasons at this point it is reasonably clear that the linkage between Bank lending and movement toward action plan objectives has not been very robust.2al 37. A great deal of formal and informal sector work was carried out during this period. In 1979, a Bank mission visited Mexico to study the problems brought on by a large increase in railway freight traffic, particularly food- grain imports from the US. The congestion of FNM yards and line by over 30,000 foreign grain wagons led to recommendatious for investments in i:emoving bottlenecks in the railroad. The grain traffic induced railway congestion plus the brief prosperity brought on by the 1977 oil discoveries led to the promotion of a railway renaissance and the initiation of a large number of SCT-designed civil works that were undertaken at that time and remain incomplete today.L/ 38. The lengthy 1985 Railway V Staff Appraisal Report included the results of the sector work associated with preparation of what was officially recognized as the first railway sector loan. At that time the momentum and depth of the national financial crisis was not yet manifest and the concerns of the railway subsector were the traditional ones of railway costing and pricing, subsidized diesel fuel for trucking competition, and techniques of modern operations, management and marketing. P -yond the action plan, chere were twenty- two discrete conditions of the loan that government and railway management were to undertake in support of the action plan objectives.10/ 39. Under Railway V, a major effort was undertaken in the area of costing and pricing. The basic assumption was that if railway management knew precisely the costs of hauling different commodities and providing different services ZZ/ The Region has intosed the Audit that this target has recently been achieved. ;1/ For a brief treatment of the reasons for the generally poor performance with the action plans see Section VII - Findings and leases. 29/ For a summary of the principal civil works undc -aken see Annex 5 of SAR Railway V, pp. 89- 95. To get some feeling of the enthusiasm and scope of the initiative see the article rRail Renaissance in Mexico* in the April 13, 1981 Railway Aie, and *Big Pesos for Mexican Railem in the September 1984 edition of the same journal. Jg/ SAR Railway V, pp. 28-29. 13 (principally passenger services) that raising tariffs to cover costs (or explicit government subsidies) would follow. Getting the prices right would be a major step towards financial rationalization. 40. The other major effort in the financial area was to do something about the long term debt of the railroads. Interest charges were consuming 46 percent of gross revenues by 1983, surpassing losses on passenger services and underpriced freight services as the chief reason for financial difficulties. This was building up because of the railroad's need to borrow to cover operating cash shortfalls. Since this was the government's fault for not allowing tariff increases, the proposal was for the government to take over this element (about 40 percent) of the railroad's long-term debt. 41. Little was said in the 1985 Railway V SAR about cost reductions, work rules, the size of the labor force or the need to disinvest in some areas. As long as traffic was assumed to grow faster than the labor force, labor productivity would continue to improve. Given the emphasis on a railway renaissance and the decision to expand capacity in numerous areas with the SCT construction projects, it is not difficult to understand in retrospect why the contradictions raised by a disinvestment strategy were not welcome.311 42. By 1987 the picture had begun to change. A Transport Sector Strategy Paper was prepared to take into account the national financial crisis.l2/ The report was comprehensive and examined the impact of the vastly reduced publi-: resources on the various subsectors within the transport sector. The most outstanding impact was on the national highway system which carried 65 percent of intercity ton-km. The funds were not available even to maintain the existing road system. Yet, SCT was going ahead with an $ 80 million purchase of first class passenger coaches and electrification of the Mexico-Queretero line - both marginal or submarginal investments. Investments in road maintenance with enormous economic returns were being foregone because of the momentum built into the SCT commitment to the expanded role of the railway system in the transport sector. 43. By 1989 the Bank was becoming more explicit a'obut the overinvestment in railroads relative to highways.3/ The updated sector strategy report raised the possibility that the decline in the railway's performance could not be solved by more investments: The root cause of the declining productivity and resulting low levels of customer satisfaction appears to be that Mexican Railways still does not see itself as a truly commercial organization and does not reward efficiency, or penalize lack of it. (p. 29) 31/ Projects staff has Informed the Audit that a good deal more analysis and refom effort was going on at this time thean the written record reflects. *See parse. 1-5 of Annez III for these recollectimus. / nexico - Transport Sector Strategy Paper, Report No. 6552-1E, February 20, 1987, The World Bank, Washington, D.C. g/ Mexico - Transport Sector Strateim Update, (Yellow Cover Draft), June 30, 1989, The World Bank, Washington, D.C. 14 On locomotive utilization and availability: There is little doubt that this critical department (mechanical engineering) has lacked both the Incentive and the analytical tools necessary to achieve efficiency improvements. (p. 29) On operations: The Operating Department has lacked the necessary technical expertise and motivation to undertake the underlying analysis of operations which is essential if efficient utilization of locomotives and wagons is to be achieved. (p. 29) ...the railway has not yet developed an effective wagon tracking system. (p.30) On work rules: Another major difficulty concerning operations relates to the labor work rules which tend to be restrictive and do not allow flexibility. ...Management needs to be fully supported in these efforts (to reform work rules) which would increase the railway's ability to take advantage of technological developments and compete for traffic.(p. 30) 44. Of all of this sectot work, only the 1979 report on railway bottlenecks was relevant for the preparation of Railway IV. The planning environment at that time was almost exclusively one of growth and expansion and the design of the project reflects that thinking. IV. Project Objectives. Definition and Design 45. As the action plan for Railway IV in Table I specifies, the idea of the lending operation was to (i) make NdeM more efficient in its operations, (ii) approach financial viability, and (iii) provide sufficient capacity to handle the requirements of a 7.3 percent annual growth in freight traffic over the period 1981 to 1983. The support for a major expansion of the role of railway in the transport sector was being put into place. 46. The physical hardware to support the above objectives was mostly locomotives and rolling stock, accounting for 75 percent of the appraisal project cost estimate. This was followed by track and structures (17%), and signalling and telecommunications (4%). The non hardware components were: (i) cost studies for uneconomic passenger and freight services as well as low density branch lines, (ii) a comprehensive management information system, (iii) computer simulation models for railway line and yard capacity planning, and (iv) a manpower plan. 15 V. Imolementation Experience 47. The 7.3 percent annual growth in freight traffic over the period 1981-1983 did not transpire. Freight traffic for 1982 was 9 percent below 1981 levels. For 1983 traffic was back to the 1981 level. The idea then was to stretch out the period covered by Bank support from 1981-83 to 1981-85. If the 9 percent decline in 1982 is ignored, freight traffic between 1981 and 1985 eventually grew at an annual rate of about 1 percent. From 1985 to 1991, freight traffic has subsequently declined at an annual rate of 5 percent.4/ The projection in the SAR called for a long term annual growth rate of 4.6 percent through the year 2000.25/ The stagnation and then secular decline in freight traffic for the period 1981-91 presumably removed some of the urgency for provision of expanded capacity.36I 48. Despite the substantial difference between actual traffic and projected traffic, the hardware component of the project was implemented much as specified except for a five-year, rather than a three-year, implementation period. For Iocomotives, the actual procurement was 317 new units rather than the originally planned 242 units, an inexplicable development given the traffic trends and the target for improved availability.2l Freight wagons, however, were 8,300 new units rather than the planned 9,300. Track rehabilitation targets were also exceeded with 938 km. of new railway rather than the planned 780 km. Bridge strengthenina was applied to 857 bridges rather than the planned 510. Only 612 km of C'[ investments were made rather than the planned 960 km. 49. The software components of the project were not implemented despite the assertion in the PCR that substantial "in house" work was done by NdeM staff .21/ The component of the loan for foreign consultants to assist in introducing modern computerized planning and operational methods was not used because of "reticence" on the part of NdeM operational management. The assertion in the PCR that much of this work was "effectively done" by NdeM staff was not verified by the Audit field visit.39/ 24/ Serieq gatediticas, FIK, 1990, Mexico D.F., p. 29. 3l SAR Railway IV, p. s0. 26/ The Goverment In its coments (Annex II) points out that the period 1981-85 was not strictly a period of stagation (pare. 5 and 6). It also points out that while the not tons offered was declinng the average lead was increasing. This would happen if the shorter haul bulks were being diverted to road transport. For the mare dramatic decline In traffic over the period 1984-90 (pares. 14-17), the Goverment accepts the explanation of the general decline In economic activity, tariff increase and the deregulation of trucking. In addition it points out that the refo=ed public sector uses of the railroad (steal, cement, fertiliser, and fuel uil) were improving their raw materials management. An Interesting internal factor cited was the low quality of service offered by the railway and the lack of specialised equipment such a special purpose flat cars for contaners. R/ PCR Railway IV, p. 21. II PCR Railway IV, pp. 11 and 29. "I In the late 1970s, WdeN acquired the TOPS wagon management system from the Southern Pacific. After two years it was no longer operational. 16 50. The project objective of approaching financial viability, as measured by the Operating Ratio, was achieved despite the poor traffic picture. During a period of strong inflation, substantial real tariff increases were put into place while real wages declinad.o/ The objective of more efficient railway oerationg was not achieved. Locomotive availability declined substantially as did wagon productivity. VI. Proiect Economics 51. For the locomotive investments, the SAR specified the purchase of 242 units at a cost of US$ 221 million in 1980 prices. This was 21 percent of the project cost in the SAR. The locomotive requirement analysis in the SAR actually called for the purchase of 222 units in the period 1981-83 and a total of 342 units over the 1981-85 period.41I The plan was predicated on freight traffic volumes increasing from 36.2 billion ton-lan in 1981 to 48.3 billion ton-km in 1985. The 1985 figures actually reached only 38.1 billion ton-km (and has declined steadily ever since). If we accept the proposed scrapping program in the SAR analysis and rework the locomotive requirements with the actual traffic volumes, the required number of new locomotives over the period 1981-85 is reduced from 342 to 268 - a reduction of 74 units. Since 317 units were actually purchased, a surplus of 49 units accrued by 1985. This represented a premature investment of roughly US$ 45 million (about US$ 67 million in 1991 prices). The incremental returns on these premature acquisitions would be less than the average return of 18 percent estimated in the SAR.42/ 52. For the wagon investments, the SAR specified the purchase of 9350 freight wagons, over the three year period 1981-83, at a cost of US$ 546 millions in 1980 prices./ This was 52 percent of the project cost in the SAR. The wagon requirements analysis in the SAR actually calls for 8,800 units in 1981-83 and a total of 15,800 units in the period 1981-85.441 Since there was no scrapping program in the wagon requirement analysis, the increase in local and foreign wagons required to carry the 1985 as opposed to 1981 traffic would be (assuming no improvements in wagon productivity) proportional to the actual traffic increase, approximately 4 percent (instead of the 34 percent projected). This would require an increase of only 1600 units over the 1981 level instead of 40/ The nature of this extraordinary and short-lived phenomenon is illustrated well in Table I of Annex 2, SAR Railway V, p. 70. 41/ SAR Railway IV, Table 3.6, p. 58. The sort of locomotive analysis undertaken by the Bank in its SAR is not entirely adequate for Mexican conditions. The simple Horse Power analysis employed ignored the Impact of the excessive operating districts (twenty-five) and the extremely hilly terrain of the Mexico City-port traffic. A segment analysis of the basic network is both desirable and possible. 42/ SAR Railway IV, Annex 6, Table 1, p. 91. 43/ This yields a unit cost per wagon of US$ 58,400 which is about 30 percent above comparable wagon prices in the US at the same time. The arrangement for domestic manufacture of wagons appears to have been a costly one. A/ SAR Railway IV, Table 3.7, p. 59. 17 the 13,400 in the SAR analysis.5/ Approximately 8,400 units were actually purchased in the 1981-85 period, a suostantial reduction from the 15,800 planned but still about 6,800 more than actually required over the period 1982-85. Since the foreign wagons in use on the railroad had been assumed in the analysis to have reached a feasible and desirable level by 1981, about 6,800 surplus wagons entered the system for a premature investment of about US$ 400 millions (about US$ 600 million in 1991 prices)."/ 53. Thus, of the US$ 794 million allocated for motive Dower and rolling stoc , about US$ 445 million, or 56 percent of the total, was clearly premature in view of the actual traffic carried by NdeM. Given the scrapping requirements for locomotives beyond the 1985 period, the US$ 45 million of premature investment for this subcomponent is not so serious. However, given the long life and generally good condition of the wagon fleet, the US$400 million premature investment in this subcomponent represents a serious misallocation of resources. The steady decline in traffic from 1985 to the present makes the situation even worse. The most immediate manifestation of the overinvestment in wagons is a decrease in the fleet productivity. As the action plan for Railway V indicates, the 1984 wagon productivity was 1,700 ton-km per serviceable car-day, down from the 2,100 figure of 1981. The objective in Railway V is to get that figure back up to 1,900 ton-km per car-day, still far below the 1981 starting point. For locomotives, the surplus of units has been reflected in a decline in availability rather than productivity. Locomotive availability dropped from 77 percent in 1981 to 68 percent in 1984. The objective of the Railway IV action plan had been to raise locomotive availability from 77 percent in 1981 to 81 percent in 1985. The hope in Railway V was to get back to 80 percent by 1988. The overinvestment in these two subcomponents made it impossible to achieve the action plan efficiency targets.4.Z1 54. For track and structures, the SAR specified 780 km of track rehabilitation using new railway and 560 km of rehabilitation with second hand railway. It also called for strengthening or replacement of 510 bridges. All of this was to cost US$ 137 million or 13 percent of project cost and be effected over the period 1981-83. The economic justification for the track rehabilitation was based on (i) reduction in track maintenance costs, and (ii) avoidance of speed restrictions leading to capacity reduction, saturation and ultimately, diversion of traffic to road. Numerous plausible assumptions were 45/ The wagon requirement analysis in the SAR assumes no Iaprovement in the average car load over the 1980-85 period and no Improvement in the car-Ve. per day. In the US during the same period there was a 9 percent increase in average loads and a 6 percent increase in car distances per day. The SAR wagon analysis did not seem to put much credence in the action plan objectives. 46/ The analysis in Table 3.7 of the SAR has the foreign fleet at about 4,500 unite from 1981 to 1983. The action plan for Railway IV assume. there were 10,000 foreign care on line in 1981 to be reduced to 7,000 by 1985. This is another example of the disparity between the action plan and the wagon analysis. As can be seen from the Railway V action plan in Table I of the PPAR, the actual number of foreign cars on line in 1984 was down to 3,500. 4/ For the Government a view of the locomotive and wagon procurement activity during this period see pars. S through 10 of Annex II. 18 made to effect the economic analysis of this element of the project. Unfortunately, there was little empirical evidence to support the assumptions selected. 48/ 55. The Audit subsequently learned that the criteria actually used by FNM for railway replacement is extraordinarily conservative, inconsistent and is not based on the pro forma analysis done in the SAR.49I On the FNM main line, the criteria for railway replacement is said to be 30 internal defects per 100 km of railway per year. Use of the AAR Railway Planning Model indicates that, on financial grounds, a rate of about 240 defects per 100 km of railway per year is a more appropriate criteria for the highest density lines in Mexico (17 MGT). For the lower density lines a defect rate of about 360 per 100 km of line per year would be the replacement threshold.2_I 56. Research in the AAR has also shown that for railway subjected to very heavy axle loads (100 ton cars), railway surface maintenance, primarily grinding, is critical to prolonging railway life (from 50 to 200 percent). Yet, FNM has yet to do any railway grinding.1_I Instead, the emphasis has been on the use of new railway to replace surface damage, the replacement of wood with reinforced concrete ties and the pursuit of the extremely high track standards required for a railroad primarily interested in passenger operations.L2/ 57. A critical analysis of the investments actually made under Railway IV indicates that a great deal of the program was not based on adequate empirical L/ SAR Railway IV, Annex 6, pp. 87-89. To give the Bank analysts credit this is one of the most difficult areas for perfoEming an economic analysis. A special committee of the AAR has spent years attempting to document a Owith and without" situation that has an empirically verifiable basis. However, the saturation-diversion assumption adopted in the SAR is highly unlikely since a single railway line can carry up to 100 million gross tone per year without much difficulty while the maximum traffic on a single line in Mexico is currently only 17 million gross tons per year. 49/ =9as track rehabilitation criteria are set forth In a special annex to Annex II. It concedes that the number of Internal defects is not the main parameter in deciding on track renewal and that the physical external state of the railway is the main reason for replacement. Such an approach is obsolete given the advent of railway grinding for controlling the external state of the railway. So/ The Audit arranged to have the AAR Railway Planning Model run for Mexican conditions. The Audit's overall impression of the track renewal process in Mexico is that it has evolved from a rule-of-thumb approach and that neither financial nor economic considerations have yet been explicitly integrated into the process. See the special annex on Track Rehabilitation Criteria in the Government9s Annex II. 21l It is In the process of purchasing its first railway grinding train. (Para 34, Ansex II) 52/ both the Region and the Goverment (Para. 35, Annex II) object to this statement in connection with the replacement of wooden ties with concrete, stating that concrete ties and fastenings cost only 17 percent mom than wood, have a useful life two and a halt time wood and allow greater track carrying capacity. The data gathered by the Audit In Mexico and the US Indicate that the concrete ties and fastenings are roughly twice the price of the wood and that hardwood (not pie) ties have a suffteently long useful life to reader the Impact of extended life of little moment In a present worth analysis. There is so evidence to suggest a greater track carrying capacity with concrete ties excapt on traffic saturated Ites where track time for maintenance becomes a factor. This is not a factor in Mexico where freight traffic density Is low. 19 evidence and/or was clearly premature.HI The Audit does not have the detailed information to make a comprehensive retrospective economic analysis, but the sensitivity analysis in the SAR indicates that a 50 percent reduction in the assumed traffic growth rate (down to 3.3 percent per year) would yield an ERR for motive power and wagons of about 12 percent.HI/ Since the actual traffic growth rate for the last ten years has been stagnant or declining, it is reasonable to assume that the retrospective ERR for this major component of the project, using the SAR's assumptions, would now well below 12 percent and not much above zero. Similarly, a 25 percent reduction in traffic growth rates in the sensitivity analysis reduced the expected ERR of track renewal investments in the SAR from 17 to 14 percent. A zero to negative growth rate for ten years would lower the ERR to well below 10 percent even under the favorable (and unrealistic) assumptions of the SAR. Given that these components accounted for about 96 percent of the total project, the Audit concludes, unlike the PCR, that the Railway IV investment program was not an economic success.51 VII. Findings and Issues The Performance of the Economy 58. The poor performance of the Mexican economy between 1981 and 1991 has had a serious impact on the demand for railway freight transport. This factor alone largely explains the generally successful use of investment resources under Railway II and Railway III (1972-79) and the generally poor yield of investments under Railway IV and Railway V (1981-88). No analyst could be expected to have prepared a forecast of ten years of stagnant/declining traffic as the basis of an FNM investment plan for 1981-91 since no macroeconomic projections to support such a view were ever provided. The question then needs to be asked as to why, in the face of declining traffic, the procurement of surplus locomotives and wagons continued. For locomotives, part of the answer was long lead times and the previous commitments to buy from foreign manufacturers. For wagons the chief reason was the need to keep the government-owned wagon manufacturing facilities 13/ The Region states that the Audit does not ake use of the economic analysis information in the PC. It is difficult to do this since nowhere In the PCR is the actual traffic in ton-ka compared with those in the SAR. Wor, is there any statemeat concerning the actual traffic growth between 1985 and 1990, information that was available at the tine of PCR preparation in 1990. Nor is there any statmat about the assumption of long tem traffic growth that was used In the retrospective economic analysis. The retrospective very high rates of return in the ICR are s1ply asserted in a table without the underlying assumptions. Sees PCR Railway IV, p. 25. 81/ SAR Railway IV, p. 30 and Table 4.1, p. 62. g/ The Region is of the opinton that this reworking of the economic evaluation based on the sensitivity analyses in the original SA is very crude. The Audit concurs. The Cowerment, In its extensive commants In Anex 11, Maks no s1ation of the fosmal economic analysis and its adequacy. Since W1 staff prepared the economic analysis pres8nted in the PCR they are in the best position to coment on the goodness of the assumptions employed. They have chosen not to do so. 20 operating during a period of extreme domestic economic distress.6/ Had the wagons been manufactured abroad, the procurement would almost certainly have been cut back. 59. The generally poor action plan achievements under Railway IV and Railway V can also be largely attributed to the much less than expected traffic without a commensurate decline in locomotives and rolling stock procurement. Idle capital is reflected in the declining productivity rates. The steady increase in the railway labor force over the same period of economic distress is also understandable given the very high unemployment rates during the recessionary period. As in the case of the wagon manufacturers, the presstre to maintain or expand public employment during an economic crisis is difficult to resist.ZI Revenue Raising 60. The persistent failure of FNM to generate enough revenues to cover its costs for the last twenty years is, however, another matter. Under Railway II and Railway III, the willingness of government to subsidize FNM is manifest. Tariff increases during a time of relative prosperity and traffic growth were clearly feasible but they were not undertaken as the government had (or could borrow) the resources to avoid them. Under Railway IV, when the financial situation of the entire public sector became critical, two major long-sought policy objectives of the Bank were finally achieved: the near rationalization of railway freight rates with relatively greater increases in the most heavily subsidized categories, and tax increases on diesel fuel used by the trucking competitors of the railroads. In a relatively brief period, two long-standing major distortions of government transport policy were rectified. The bad economic times have led to some premature investments but they have also stimulated revenue-generating rationalizations. 61. The brief period of financial improvement under Railway IV has passed. The simple expedient of net-revenue-generating real increases in freight rates is increasingly problematical given the simultaneous deregulation of long haul trucking and the down sizing or privatization of previous government monopolies (steel and fertilizer) that had accounted for a certain amount of guaranteed traffic for the railroads. Railway tariff increases must now be considered in the light of potential traffic diversion or cessation. This raises the much more difficult issue of cost cutting and efficiency measures, most of which have a major impact on labor. 56/ According to the SAR, the government-owaed wagon factories were capable of producing about 3,000 wagon. per year. In order to utilize the installed capacity, 8,800 wagoie were to be produced In the factories during the period 1981-83. SAR Railway IV, para. 3.12, p. 27. The Goverament, in its commnts (para. 21, Annez II) states that the current ephasis Is en rebuilding wagons and that this policy has led to the bankruptcy of the government-oued Constructora Sacional de Carros do terrocarril. 2/ It is also a standard element of recession-fighting fiscal policy providing it can be reversed during times of expansion. 21 Cost Cutting and Efficiency Measures 62. If the operational performance of the financially viable railroads of the world are compared to that of FNN, it is clear that there is considerable room for improvements in efficiency and the cutting of costs.58/ The major routes open to FNM are (i) disinvestment and downsizing, (ii) labor and capital saving methods and technology and (iii) reduction of surplus labor and the changing of work rules. 63. The potential for major cost savings through disinvestment and downsizing is enormous. Over 96 percent of the freight traffic is confined to 13,000 km of the 20,400 km system. This traffic could be handled with 800-900 locomotives operating out of five maintenance shops. The labor force required would be 20-30,00 employees.59/ Instead, FNM employs over 80,000 employees, and operates 1700 locomotives out of about 40 maintenance shops. FJL- must maintain 7,000 km of track that carry virtually no traffic, and hundreds of marshalling yards and freight stations that have had no function for the last twenty years. 64. The surplus infrastructure does much more than add to the FNM maintenance burden. A principal reason for the low locomotive availability figures is the operation of about 40 locomotive maintenance facilities (a legacy of the steam era) where only about five are needed. All are seeking to justify their existence by working on costly diesel locomotives.6o/ Hence, the railway adage that "workshops eat locomotives". The other half of the adage states that "yards eat trains". Given the increase in unit trains and the disappearance of LCL traffic, there is no longer any need for most of the marshalling yards that remain in the FNM system.!1l The most effective initiative for improving average train speeds and train productivity is to disinvest in marshalling yards. Every time a train enters a marshalling yard for any reason a large block of time is lost. Keeping locomotives out of workshops and trains out of yards is the low cost way to rapid gains in productivity. One way to achieve this is to dispense with the surplus facilities and labor working rules that consume capital in an effort to justify the existence of a large amount of redundant labor. 58/ See Annex 1 for international comparisons. 59/ This assumes also the cessation of virtually all the intercity passenger services of all classes. 60/ In retrospect it would have made a great deal of sense to have retained a substantial fleet of steam locomotives to give the surplus labor in surplus workshops something to work on that wae much less expensive than a diesel locomotive. Low capital cost steam locomotives can also make good economic sense on low density lines and in situations where a locomotive Is required on station but used infrequently. The dominant labor-saving benefits of dieselization were only temporarily achieved and the less important energy saving element that was achieved would have justified a much more gradual shift out of steam as has been the case in India and China. 61/ The last time anything like this was mentioned officially within the Bank was in the 1964 First Transport Sector Report in which it was pointed out that by 1959, LCL traffit was disappearing and that of the 2,000 freight stations on the system only 36 accounted for 74 percent of the tonnage shipped and 61 percent of the tonnage received. 22 65. A great deal has already been invested in labor and capital saving methods and technology. Attempts have repeatedly been made to introduce the modern car control, operations planning, and management information systems of the effectively restructured railroads north of the border. Attempts were -made to introduce these low cost innovations under Railway II, III, IV and V with some temporary successes in terms of installation of the systems. However the systems were not wanted or used by the management of the Operating Department, for whom they were intended, and they quickly languished. 66. The Operating Department's veto on the use of the UHF and VHF radio communication system (financed by the Bank under Railway II and Railway III) for train control is another example of the problem. The labor-intensive open wire system for issuing train orders by Morse telegraph continues to be operated, one of the last such systems in the world.62I Operations has explained that they have not had confidence in the safety of the radio dispatching system. But, CTC technology has received their approval. This has led to the phenomenon of the premature jump from the open-wire telegraph technology to the high-tech and costly CTC technology without the intermediate step of radio dispatching on lines of moderate volumes (4 to 6 MGT per year).1/ Premature investment in CTC continues to give the illusion of modernization at great cost.ft/ 67. Probably the most significant cost of the 1956 labor peace was the agreement that all managers of the key Operating Department of NdeM must have served as a member of a train crew. As a result there are only five graduates (all in the planning section) of a department of at least 15,000 employees.i/ Nobody in a management position in the Operating Department of FNK has had any experience in the other departments within the railroad, much less outside the 62/ Open wire systems are costly to maintain. In the US, the radio dispatching systems commonly used were sometimes financed by the sale of the copper wire embodied in the abandoned open- wire technology. 63/ In at least one section of single line CTC financed by the Bank under Railway IV, the anmber of trainL per day was only 12 io each direction. Yet there are passing sidings on the average every 10 he. Abandonment of two thirds of the passing sidings to achieve 30 km spacings with existing 80 halhr train speeds and the use of radio train control would have yielded train operating results very silar to the CTC investment at little or no cost. Thirty percent of the Canadian National main lines are run by radio dispatching. Given the relatively low densities on the Mexican system there is scarcely justification for CTC investments anywhere. /41 The economic rates of return on these CTC investments are very high since the Laprovement in train operations over the old open-line system is very great. However, the economic return on che use of the law cost radio train control system would be nearly infinite since the investment in the system has already been made. See pars. 36 and 37 of Annex II for the Governmeat'e view on the matter of cowmniations. Al A substantial achievement of the early Railway I project with the Pacifico Railroad was to obtain the agreement from labor that twenty key supervisory poets of the railroad no longer need be filled from union ranks but could be filled by outsiders. The labor agrements associated with the Pacifico were in general mach more appropriate for a modern railroad than those obtained by deN. Examples are the use of universal mechanics In Pacifice shope, longer operating districts, and a brakeman for every 30 care rather than every 15 care for EdeX. The unification of the railroads into WW has raised serious quations about the unification of work rules. 23 railroad."/ It is not difficult, under these circumstances, to understand why "modernization" investments are frequently either inappropriate or not used.§gZ/ 68. The cost of a single Mexican railway employee is about US$ 4000 per year.§/ The disappearance of say 40,000 workers would save the railroad about US$ 160 million per year, solving about half the financial problem.!9/ The other half would feasibly come from the down-sizing of the infrastructure - workshops, lines, yards and passenger equipment and facilities. The two sources of savings are not independent, however. A great deal of Mexican railway labor (except for the small amount of absentee workers) only becomes completely redundant when downsizing and disinvestment are effected and work rules are changed. Current operating procedures and work rules are effectively designed to keep a large work force underemployed.71l Effectiveness of Bank Suport 69. The 1987 Transport Sector Strategy Paper approaches the problem: Although the Bank's sector reviews all emphasized problems of institutional structure and pricing and regulatory policies, the Government was not receptive and lending tended to take a passive stance 66/ The inbred nature of the Operating Department, engendered by the labor agreements, leads to constant friction with the Engineering, Traffic, Signals and Telecom, and Locomotives and Rolling Stock Departments. For instance, the Operating Department has decided that the average track possession by Engineering to perform track maintenance is to ,be only four hours. In the US the average is 6.5 hour. and the train densities much higher. The problem with the use of radio train control has already been cited. Given the generally higher level of technical training of the management of the other departments and the correspondingly low level of Operations management it Is difficult to find comon technical ground for discussion of such important matters. R/ The Governmet's cments an the problem of qualified staff In the Operating Department are very interesting. The proposal is to deal with this key constraint by means of permanent edvisory services by technical specialists with experience in foreign railways. See para. 42 of Anne II. 68/ In the US the figure is about $33,000; in India $1,700 and in China about $600. The Region notes that the average annual cost per employee seems underestimated. In 1990, salaries and pensioos amounted to nearly US$ 700 alliae equivalent, or soes US$ 8000 averaged over about 90,000 employees. The Audit notes that the pensions are not avoidable with work force reductions unless taken over by the Governmeat9s social security system. 69/ It is not entirely clear just how large the labor force is. In 1985 it was officially 79,700 for all systems. This increased to 83,900 by 1991. Yet, there is a large force of temporary employees said to number about 12,000. In the San Luis Potosi shop, visited by the Audit, there were 2,600 permanent workers and 1.300 temporary. If this ratio is extended to the rest of the labor force tk.r total temporary labor force could be as large as 40,000. Should this be the case the problem is mach more serious than the official figures indicate. Zo/ One operational rule responsible for tying up locomotives In Nexico is the two hour inspection every time a yard is entered. One work rule that makes things worse is the inability to carry out these inspections while refueling because the refueling activity is under a separate shop craft. Instead of universal mechanics in the workshops eleven separate crafts mst negotiate the splitting up the work. (The Pacifico railroad succeeded in negotiating a labor agreement that allowed universal mechanics in the shops unlike the dominant Wden which specified eleven separate crafts.) Other examples, too numerous to describe, are comon knowledge. 24 on these Issues.... there was no dialogue with SCT which was responsible for the above-mentioned policies as well as the port and railway infrastructure investments which were initiated in the 1979-81 period.L_U 70. - With the notable exception of the very large loan to the del Pacifico in 1954, the Bank has been "actively" supporting new technology and operational changes while "passively" addressing the issue of work rules and downsizing. This was because the Government wished to "modernize" the railroad without addressing the thorny issue of work rules or redundant labor. As a result, the perfectly sound operational changes supported by the Bank have not been adopted and the technology investments have not been fully utilized. 71. The question can then be asked whether or not the Bank should have supported, with four lending operations over twenty years what, in retrospect, was a "second best" approach to railway reform. The case can be made that deeply rooted government policies are as much a part of the railway environment as the physical terrain between Mexico City and the seaports. As long as it was possible to justify railway investments that avoided uneconomic diversion of traffic to highways (Railway II and III), the "second best" argument had some merit. However, once *-he traffic began to decline for reasons other than adequate railway transport capacity (Railway IV and V), the traditionel Bank- supported investments in track renewal and rolling stock lost their justification. The debt incurred under Railway IV and V then became part of the problem rather than the solution. VIII. The Future Role of the Bank 72. Increasingly the Bank is involved in programs of restructuring technically bankrupt and inefficient public enterprises such as FNM. It dominates the scene in operations in the new Eastern European member countries. How fast can the viable enterprise be extracted from the bankrupt one and what sort of ownership structure and policy environment will it operate in. Fortunately for Mexico, the public sector is not nearly as large a share of the economy as is the case in most of the former Communist countries. The distortions have not been so great and the basic future policy environment of competition and free pricing already exists. Also, the railway mode in Mexico is not nearly as important in the total transport sector as is the case in Eastern Europe. FNM is also well-endowed with adequate and modern infrastructure and equipment thanks to ten years of large investment programs and stagnant to declining traffic. This latter factor means that the economically justified investment requirements of the FNM restructurine effort will not be large.721 73. This raises the possibility that in terms of lending operations it may be most appropriate for the Bank to do nothing at this stage of the process. This is because the fundamental pressure for change must come from within the ZI/ Mexico - Transport Sector Strategy Paper, Report No. 6552-ME, February 20, 1987, The World Bank, Washington, D.C., p. 3. 72/ See Annex II for the Government's comente. 25 Government - not the Bank. And, given the experience in other countries, the origin of the pressure for reform is alya the intolerable financial cost of an unreformed railway. Action plans implying structural change simply do not get pursued seriously if the political will is not there to take extremely unpleasant initiatives. These initiatives cannot be taken without a great deal of preparation in which the political, human, technical and financial elements of reform are developed in considerable detail. Things have not yet reached this stage in Mexico.Z3I 74. Considerable progress in ports, aviation and trucking by the current Government has been made.74/ A start is being made with railways with the proposed 1991 Structural Change Program from the management of FNM. It indicates that the costs of doing business an usual have become nearly intolerable.5/ The need for a drastic reduction in the railway labor force is now recognized explicitly with a proposal to cut 28,100 jobs from the 1991 level of 83,900. This would leave a labor force of 55,800 by 1995. By the year 2000 the numbers would be as low as 45,000 - a 46 percent reduction in the total railway labor force. Nothing like this has been mentioned officially within the Bank since the 1971 Transport Sector Report when a 17 percent overmanning problem was mentioned. It is now acceptable within Mexico and within the Bank to discuss this most crucial element of solving the railway problem. This, above all, is the major potentially productive departure from previous attempts at action plans. However, the important related problems of work rule changes and management of the Operating Department have not yet received the -t :ention they must have if movement is to take place.76I 75. The Structural Change Program also recognizes explicitly the need to disinvest and downsize in the locomotive repair shops. The Audit mission was informed that the current plan was to reduce the numbel: of workshops from the present 35-40 down to 14 and eventually 10. This is still at least twice as many as needed but is a clear recognition of the need to make drastic changes if locomotive availability is ever to approach the 85-90 percent level it should be at.ZZ/ The need to disinvest in yards is not treated in the Program so this remains an important missing element. 73/ According to the Govertmat0 commnte Io Annex II, things have reached this stage. Z4/ Mexico Road Transport and Telecomunications Sector Adiustment Project, Report No. P-5254-HE, May 3, 1990, The World Bank, Washington, D.C. 75/ The Audit tae provided copies of: Programa do Cambio Estructural 1991-1994 - Resumen Elecutivo, FN, (no date), Mexico, D.F. The document is printed and bound but is not paginated. 76/ According to the Govervmnt0s caments (para. 42, Annex II) the recognized constraint in the Operations Department has generated a potentially feasible approach to dealing with it. For work-rule changes there is no evidence of similar progress In thinking. W/ This alne will not solve the problem of locomotive availability. Revision of the extraordinary system of 26 operating districts and the inefficient train operating practices needs to bed considered in conjunction with the decisions about which workshops are to be retained. 26 76. The other explicit positive changes proposed in the Structural Change Program are more like those discussed in the past, ie. suppression of losing passenger services and the closing down of uneconomic lines. Numerous studies to satisfy the Bank's desire for reform in these areas have been done in the past but nothing significant ever came from them. There is nothing in the Structural Change Program to suggest that thinking on these issues has progressed beyond the stage of general discussion.ZaI 77. What is particularly worrisome at this stage is that, despite an explicit recognition of the poor operational and financial performance of FNm and the need for doing things very differently in the future, the Strutural Change Program continues to include very large sums for investments for the sort of general modernization and improvements that were included under the previous four investment programs supported by the Bank.71 As pointed out in Section VI of this report, the track and bridge and CTC components are highly suspect investments. The large component for completing ambitious and costly SCT- initiated railway construction projects needs to be scrutinized very carefully. It isn't enough to continue to put money into new construction efforts simply because so much has already been invested. Painful as it may be, an analysis assuming sunk costs may show that it is better to defer further investment indefinitely.Io/ The clear message of this component of the Plan is that an all- important consensus on the need for major railway reforms in Mexico has probably not yet been reached. A substantial contingent of the railway renaissance element remains in SCT and FNM and any future upturn in traffic and revenues could easily put this element back at the center of national railway policy.81I 78. Flowing from the foregoing (paras. 74-77) this audit offers the following recommendations on future Bank landing for railways in Mexico. I1 Paragraph 19 Io Annex I is representative of these laudable general Initiatives. Paragraph 40 In Annex II i san indication that some very wiplicit action has recently been taken. 79/ For the period 1991-94 the program proposes a total public investment of US$ 1.8 billion. The primary elements would be track and bridge reconstruction (41 2), completion of previously started SCT construction projects (22 2), locomotives and rolling stock (18 Z), CTC investments (6 2) and others (132). In the Goverment0s coments (para. 27, Annex II), these numbers have been revised downard so that the current estimate is now US$ 1.5 billion over the four years 1992-95. This wold be about US$ 375 million annually which is sti1l, In the Audit's opinion, a very large number. Both the Region and the Government dismiss this view on the grounds that even a fao growth" railway generates a large need for replacement and rehabilitation. Yet, when compared to the replacement and rehabilitation requirements of US railroads it appears excessive. According to Annex II, over a 12 year period the Mexican railway system rwed 50 percent of its locaotive fleet. During the sm period the US Clase I system renewed only 30 percent of its locomotive fleet. The extraordinarily conservative approach to railway and bridge replacement I Mexico L also noted. 8g/ This is particularly true for the proposed completion of the electrification of the Mexico- Queretero line. If the overhead catenary is put In place as designed it would preclude the use of this important line by double-stack container trains because of inadequate overhead clearance. Sea Paragraphs 11 and 12 of Annex II for Government coments on this point. 81/ As recently as 1986, NdeN analysts had proposed a doubling of passenger traffic over a 15 year period with the acquisition of 869 passenger coaches. The losses were to be financed by cross subsidies from the freight service. 27 79. First, any new Bank loan for the Mexican railways should be contingent upon (i) implementation, prior to Board presentation, of a number of subtantial steps and measures to restructure FMN which should have been discussed and agreed with the Bank, and (ii) full agreement, between the Bank and the Government, prior to Board presentation, on a detailed and comprehensive program and timetable for the completion of the restructuring of FNM during project implementation. 79. Second, in addition, the size of any new Bank loan could (should) be considerably less than that of previous railway projects given the emphasis to be placed on downsizing the railroad (disinvestment, staff reduction) and key operational and policy changesl2/. 80. Third, and finally, the audit feels that any new Bank railway project should make adequate (generous) allowance for Bank staff time to enable careful and frequent monitoring of the implementation progress of the restructuring program (including a comprehensive annual review) during project execution. JI/ This is not to say there would not be taportant investments to be made in a down-sizing railroad, on the contrary same amall investets in domestic Intermodal facilities and selected pieces of tract maintenance and workshop equipment would be essential to success. It is to sy that nothing like the resources previously allocated for rolling stock, track rehabilitation and CTC investments 'would be required if anything like the potential productivity gains of serious restructuring were to be realized. 29 Annex I Page 1 of 8 ANNEX I: CURRENT PRODUCTIVITY AND INTERNATIONAL COMPARISONS The few comparisons made between Mexican railroads and other railroads in the past has involved references to the railroads of Europe. This may have been appropriate at one time, but the evolution of Mexican railway to long hauls and unit trains, with equipment compatible with that of the US and Canada, makes it equally (and perhaps more) appropriate to examine the productivity of the Mexican operation in the context of other high-volume, long-haul freight operations as well../ As the action plan exercise in Table 1 of the text indicates, yesterday' inefficiencies may become today's standards. Some notion of what is good and desirable needs to be developed from world wide experience with railway transportation. For purposes of this exercise, comparisons are made with the European systems of France, Germany and Spain, the Conrail system of the US, and the systems of India and China. (The data and sources are set forth in Tables la and lb.) Conrail is similar to the Mexican freight system in extent and technology (except for intermodal) and has undergone considerable restructuring in the last ten years. India's is a larger, higher density system with a dominant passenger operation and labor problems (if not wage rates) more comparable to those of Mexico.j/ China is included simply because it gives some notion of what maximum physical productivity levels are achievable in a capital resource constrained environment. Geographic Coverage In 1991, Mexico has a 20,400 km railway system for a national area of 1.97 million sq km., approximately one km. of railway for every 97 sq. km of area.!/ The European systems are much more dense from 9 sq km per km of railway in Germany to 40 sq km in Spain. This can be compared to the Chinese railway system which serves 181 sq. km. with each km. of railway and the Indian system serving 51 sq. km. per km. of railway. The US Class I railroads serve 46 sq km per km of railway. In terms of average geographic coverage, FNM is only about half as dense as the US and Indian systems and about one quarter the density of the European systems. / In 1988, approximately 88 percent of FNM total traffic units were freight unite. At the ease time, the average freight haul in Mexico vas 718 km. In Europe in 1988, passenger and freight traffic were of roughly equal importance and the average freight hauls varied from 215 km in Spain to 360 km in France. For US Conrail, freight was 100 percent of output and average hauls were 729 km. In China, freight was 75 percent of output and average hauls were 701 km. Average railvay labor costs for Mexico are presently about US$ 4000 per worker. For the US the figure is $ 33,000. For Iridia and China the figures are $1,700 and $ 600 respectively. 3/ For paved roads, the approximately 70,000 km network yields one km of red for every 28 sq. km. of area. 30 Annex I Page 2 of 8 Network Productivity If the total traffic units in 1988 are divided by the length of the Mexican railway system it yields a network productivity figure of 2.3 million traffic units per route-km. This may be compared with 2.1 million in Spain and 3.7 million in Germany. For US Conrail the figure jumps to 7.1 million while that for India is 8.0 million. China is in a class by itself with an incredible 24.7 million traffic units per route-km. This implies that the Mexican network, despite its relatively low density, is lightly trafficked relative to the US and Asian railways and only slightly less trafficked than the European systems. Reduction of the Mexican system to the 13,000 core kilometers that account for over 96 percent of the traffic would yield a network productivity figure of 3.7 million - much closer to international standards. Wagon Productivity If the total capacity of the Mexican freight wagon fleet in 1990 is estimated as the total number of wagons (60,930)ii times the average load of a loaded wagon (58 tons) and this is divided into the ton-km of freight traffic for 1990 it yields 11,335 ton-km per ton of wagon capacity as a wagon productivity measure. This compares with 30,000 for Indian Railways and an extraordinary 53,600 for China Railways.J/ For US Class I railways the figure is only 12,000 ton-km per ton of wagon capacity.!/ Mexico has access to a modern wagon fleet but its utilization of that fleet is low relative to China and India and only slightly lower than the Class I railroads of the US. Locomotive Productivity For locomotive productivity the most accessible (though not the most appropriate) measure is to divide the number of annual traffic units (ton-km plus pass-km.) by the number of diesel and electric locomotives. For Mexico the figure was 27 million traffic units per locomotive in 1988. In Europe the figures varied between 18 million for Germany and 27 million for France.1I For The total fleet is estimated to be comprised of 46,600 railway-owned wagons and 14,330 wagons owned by private Mexican firms and foreign railways. 5/ The figures were not available for the European railroads. 61 The low figure for the US Class I railroads as a group is due to the Lrge number of specialized wagon. owned by leasing firms and shippers - not the railroads. These wagons allow the U.S. Railroads to provide services and frequencies that are not offered by the other railroads. The Burlington Northern in its coal operations is able to achieve wagon productivities in excess of 70.0 thousand ton-km per ton of wagon capacity, which probably represents the upper boundary of physically achievable wagon utilization rates. /I The European productivity figures are overstated because a substantial amount of their passenger traffic is carried in multiple unit passenger care that do not require a separate locomotive. In the computation this MUV traffic is added to that carried by the locomotive fleet. 31 Annex I Page 3 of 8 India the figure was 99 million while China was 217 million.V Conrail achieved 55 million traffic units per locomotive. Part of the reason for the low productivity of the Mexican locomotive fleet is that at any one time iaout 40 percent of the fleet is under repair or otherwise unavailable for use.21 The Indian fleet is about 90 percent available as is that of the European and Class I US railroads. The Chinese fleet is about 80-85 percent available. If the Mexican productivity per available locomotive is measured the figure would be 46 million units. Those for Europe would vary between 20 million 30 million traffic units. India and China would achieve 110 and 255 million respectively. Conrail would achieve 61 million units. An improved availability figure (90 percent) would raise Mexican locomotive productivity to levels well above those of Europe and not far below those of Conrail. Labor Productivity To compute the productivity of labor a useful measure is to divide the annual traffic units by the transportation labor force.ioI In Mexico in 1988, the 81,200 man work force accounted for 587,000 traffic units per employee. This is comparable to France (538,000) and well above Spain (435,00) and Germany (398,000) In China it was estimated to be 729,000 units per employee while the figure was 329,000 for India. For Conrail, the figure was over four million traffic units per employee. Paradoxically, the Mexican labor productivity figures are similar to those of Europe where labor is much more costly and India and China where labor is much less costly. Conrail labor is in a class by itself being seven to ten times as productive as the European and Asian systems. The ratio of the wage bill to total revenue sheds some light on this matter. The ratio was 90 percent in France, 126 percent in Germany and 125 percent in Spain. In the latter two cases total revenues are not enough to cover labor costs. This means that the railroads of Europe are heavily subsidized and, given their advanced technologies and low labor productivities, almost certainly heavily overmanned. By contrast, in India, the wage bill was about 51 percent of revenues while that of China was only 23 percent. In Conrail the wage bill was about 40 percent of revenues. Thus, the US and Asian systems form a distinct VI The productivity figures for China and India are also overstated because the traffic units in the numerator includes those pulled by steam locomotives but the denominator includes only the diesel and electric fleet. If the steam locomotives are included in the denominator then the locomotive productivity figures for India (based on 1985-86 figures) drops to 44 million traffic units while that of China (for 1987) drops to 97 million. 91 Based on 1991 data obtained from FON by the Audit. Locomotive availability (and wagon turn- around) statistics are no longer routinely included in the annual statistical bulletin of FN. 10/ Since India and particularly China have large manufacturing and construction activities an adjustment has to be made to delete this element of the labor force. 32 Annex I Page 4 of 8 cluster as do the European systems. The Mexican experience has fluctuated between the two clusters. In 1980, when tariffs were far below costs the ratio was a European-like 85 percent. In 1984, after large real increases in tariffs and lrge (temporary) declines in real wages it reached a US\Asian-like ratio of 45 percent. A rough estimate of the annual cost of a railway worker yields the following figures: $33,000 (US), $4,000 (Mexico), $1,700 (India), and $600 (China). This explains why China and India can have such low labor productivities and still have such a low wage bill compared to revenues. European wage cost of the same order as the US combined with Asian productivities would explain the financial problems of the European systems. The low Mexican labor productivity relative to the US (about one eighth US levels) is almost exactly offset by wages that are about one eighth those of the US. If the 30,000 man reduction in force envisioned in the Mexican 1991 restructuring plan were to be achieved the productivity of the remaining labor would increase to 750,000 units per employee, well above that of China or India. To arrive at levels half those of the US Class I railroads (say 2.0 million traffic units per employee) the Mexican railway labor force would have to be reduced from 83,290 to about 20,000 employees. Unit Revenues and Costs Average revenue per freight ton-km in Mexico in 1990 was about 3.3 US cents. If the annual government transfer of about US$ 300 million is treated as revenue then the figure rises to 4.1 US cents per ton-km. This would be a rough measure of unit freight costs. This compares with about 2.4 cents unit revenue in India. 2.0 cents for US Class I railways and 7-9 cents for European railroads. Thus, the Mexican cost structure for freight operations is much closer to the high density long haul India and US operation than that of the short haul European operations. The Importance of Passenger Traffic Passenger traffic in Mexico in 1988 accounted for about 12 percent of railway traffic units. European and Indian figures are all in the region of 50 percent. China is at 25 percent while Conrail is zero. The railways with the highest labor productivity (Conrail, China and Mexico) have the lowest passenger traffic shares. This supports the fact that passenger operations are inherently more labor intensive than freight operations. It also raises questions of the appropriateness of the one ton-km - one pass km assumption for computing railway output. In the Bank's 1987 Transport Sector Study for Mexico, the point was made that passenger services represented only 8 percent of gross income but accounted for 12 percent of the traffic units and 28 percent of train kilometers.111 In .i/ Mexico Transport Sector Strategy Paper, Report No. 6552-ME, Feb. 20, 1987, The World Bank, Washington, D.C., p. 28. 33 Page 5 of 8 the Bank's 1971 Transport Sector Study for Mexico the passenger service accounted for 14 percent of revenues but was estimated (on a "fully allocated cost" basis) to absorb 47 percent of railway "effort" and accounted for 80 percent of the railway deficit.121 This suggests that government attitudes towards passenger operations is particularly crucial to labor productivity measures. Specialized all-freight railways such as Conrail probably have an inherent advantage in the labor productivity area. A comparison of passenger revenue shares of total revenue and passenger traffic as a share of total output is revealing. The railways of France and Spain have become increasingly passenger railways. In 1987-88 railway passenger traffic in France was 55 percent of railway output and accounted for 59 percent of revenues. In Spain the equivalent figures were 58 percent passenger traffic and 64 percent passenger revenues. Germany at 41 percent passenger traffic and 50 percent passenger revenues was at the borderline. India with 53 percent passenger traffic and only 28 percent passenger revenues is a special case of a railway that is primarily passenger in terms of output and freight in terms of revenues, indicating the presence of a massive cross subsidy from freight to passenger operations. China, with 25 percent passenger traffic and 21 percent passenger revenues remains a freight railway as does Mexico with 12 percent passenger traffic but only 4 percent of revenue from passengers. Government attitudes towards passenger operations is crucial not only in the labor productivity area but also in the area of financial viability. Summary On the negative side, FNM has the lowest network productivity measure of all the systems indicating that it is an overextended system despite its low density geographic coverage. This squares with the fact that 96 percent of the current freight traffic moves on the "basic" 12,800 km of the 20,400 km system. Also, an unusually low locomotive availability figure (around 60 percent compared to 80-90 percent for all other systems) prevents the achievement of otherwise quite respectable locomotive productivity figures. Finally, the passenger service of Mexico has, by far, the lowest revenue contribution figure of all the railroads examined, indicating that the railway passenger subsidy policy is one of the most durable and costly policies of the Mexican government.13/ More neutral findings are that Mexican wagon productivity figures are near those of the US, though far below Indian Railways. Mexican labor productivity exceeds that of all the European railway systems and that of India. It is, however about one eighth that of the US Conrail system. The Mexican railway wage bill as a share of total revenues has fluctuated between the high labor cost European systems and the low labor cost US and Asian systems. 1/ The Transport Soctor in Mexico, (in four volumes), Report No. PTR-88, May 13, 1971, The World Bank, Washington, D.C., Vol. 11, p. 2. 13/ Since intercity railway passenger transport accounts for only 2 percent of the national. total (all modes), this service is largely symbolic - vhich is not to say that symbolism is not important. It is to say that cessation of this service would have little or no national economic impact except on the greatly improved finances of FNH. 34 Annex . Page 6 of 8 On the positive side, Mexican railway has the lowest share of passenger output to total output of all the systems except Conrail, maintaining its emphasis on freight operations as its raison dletre. This is importaat because the other positive finding is that Mexican railway freighit unit revenues (and probably costs) are much lower than European unit freight revenues and closer (though still about 100 percent higher) to US and Indian unit freight revenue figures. Given the very great potential for cost cutting and efficiency in FNM, this means that FNM has the potential for being one of the lowest cost freight operations in the world. 35 Annex I Page 7 of 8 Table 2.1: International Comparisons of Size and Outputjg/ Passenger Passenger Avg. Laboril Railway/ Route Outputl6/ Share Share Freight Force Country Ka. Billion TU Output (%) Revenue (%) Haul (Km) (1000) Mexico 20,400 47.8 12 4 718 81 France 34,600 114.6 55 59 360 213 Germany 27,300 100.0 41 50 316 251 Spain 12,700 26.4 58 64 215 61 Conrail 17,900 127.3 -- -- 729 31 India 62,000 493.8 53 28 698 1500 China 53,000 1311.7 25 21 701 1800 14/ The most important source for this information is the International Railway Data Base maintained by the Transport Division of the Infrastructure and Urban Development Department of the World Bank. Most of the non financial data employed herein (1988) Is extracted from the October 31, 1991 print-out. The financial information (1987) Is largely extracted from the July 2, 1990 print-out. The non financial data for Mexico are extracted, for the most part, from the FN etatistical abstract "Series Estadisticas FM, 1990". 1/ Adjustments had to be made to the labor force statistics of India and China to delete the industrial and construction workers in order to arrive at an estimate of the transportation workers. The Indian figures are for 1985-86 while those for China are for 1984. See: PPAR India: Railway Modernization and Maintenance Project, OED Report No. 7020, November 30, 1987, p. 7. 16/ TU is Traffic Units - the sum of ton-ka and pass-km. 36 Annex I Page 8 of 8 Table lb: International Comparisons of Productivity Railway/ TU (1000)/ TU (1000)/.L/ TU (1000)/ Wages/,L/ Country Route-Km Locomotive Employee Revenues (%) Mexico 2343 27440 387 85-45 France 3312 26863 538 90 Germany 3663 18051 398 126 Spain 2079 19144 435 125 Conrail 7112 54800 4174 40 India 7965 99017 329 51 China 24749 217421 729 23 I/ See footnotee 8 and 9 of Annex I for qualifications. 1/ The 852 figure for Mexico was for 1980; the 45Z figure was for 1984. Sees SAR Railway V, Annex 2, Table 1, p. 70. It is not known to what extent the Chinese figure included income in-kind in the wage bill. This is a substantial element of income for Chinese railway labor. However, the estimate of US$ 600 per year for Chinese railway labor does include an estimate of in-kind as well as cash income. 37 Annex II Page 1 of 19 ANNEX II: COMMENTS BY THE BORROWER BANOBRAS International Financial Agencies Department Mexico City GOFI-064-92 Mexico City, January 21, 1992 Mr. Yves Albouy Chief. Operations Evaluation Department World Bank Washington, D.C. Subject: Loan 1929-ME Fourth Railway Project In response to your letter of December 4, 1991, I am attaching Ferrocarriles Nacionalek de M6xico's comments on the Project Performance Audit Report. Neither the Ministry of Finance and Public Credit nor BANOBRAS have any comments to make on the report. If the Ministry of Communications and Transport has any comments. we will forward them promptly. Sincerely, /s/ Ismael Diaz Aguilera Manager 38 Annex Il Page 2 of 19 FERROCARRILES NACIONALES DE MEXICO COMMENTS ON THE PROJECT PERFORMANCE AUDIT REPORT ON THE FOURTH RAILWAY PROJECT WITH THE WORLD BANK (LOAN 1929-ME) January 1992 39 Annex II Page 3 of 19 COMMENTS ON THE PROJECT PERFORMANCE AUDIT REPORT ON THE FOURTH RAILWAY PROJECT WITH THE WORLD BANK (Loan 1929-ME) SUMMARY It is recognized from the outset that quite apart from the results of the investments and the financial benefits from the financing provided, the Bank's participation was of great value in the different aspects of FNM's operations and its recommendations have, in the majority of cases, been taken into account by the Federal Government. The period 1981-85 covered by the report cannot strictly be considered as one of stagnation in freight traffic since FNM achieved its highest traffic levels during the period, which prompted the procuzement of a significant number of locomotives and wagons. It should be noted that the apparent oversizing of the locomotive and wagon fleets was partly due to a large number of units that were no longer economically repairable not having been taken out of service. Moreover, the procurement of new units was concentrated in the first few years of the period and shoved a distinct downward trend thereafter. Regarding the new infrastructure works started by SCT, although the crisis of the early 1980s prevented their completion, some of the most important ones are now partially in service. With respect to track rehabilitation and maintenance, this work has to be done systematically and although at any given moment a large part of the system is in good condition, it has to be kept in this state, especially the lines forming the basic system; however, the progress made with track and structures during the life of the loan was modest. Certain of the causes that played a part in the decline of rail traffic between 1986 and 1991 are noted, but the growth of intermodal traffic is emphasized. During this period, locomotive and car procurement was kept to a minimum while priority was given to maintenance and rebuilding of equipment and especially track reconstruction and bridge strengthening. Regarding the report's conclusion that Bank support for infra- structure investments in FNM's Structural Change Program should be minimized, the investment plans have been reviewed and adjusted to the realities of the present situation, with the aim of seeking to cover the minimum needed to ensure safe operation combined with a real and sustained improvement in operating efficiency. The investments considered for the rationalization without growth scenario are concentrated in reconstruction of an indispensable minimum of 350 km of track per year in the lines to frontiers and ports; rebuilding of locomotives and procurement of specialized equipment; installation of only 40 Annex II Page 4 of 19 260 km of CTC-type signals; modernization of the telecommunications system and of the operations information and control systems; and completion solely of infrastructure works that are in the most advanced stage and can be brought into service most quickly. Special mention should be made of the progress achieved during 1991 in certain aspects of the Structural Change Program that the report views with reserves, such as the cancellation of 71 nonproductive passenger, mixed and freight services and the trimming of 1.181 positions from the payroll. FNM hopes for significant participation by the Bank in its Structural Change Program and will appreciate possible assistance for financing the compensation payments, and also for obtaining consultancy resources for improving key areas. It would therefore be disappointing if the Bank were not to participate or only did so on a limited basis, since this would increase the cost of the project by making it necessary to seek other more costly sources of financing. FNH proposes to continue the negotiations begun for authorization of a new loan from the Bank. with the final decision as to the amount of this loan and the investment categories to be subject to the progress made in 1992 in the Modernization and Structural Change Program, which will be dependent on the support furnished by the Federal Government to FNM for its medium and long-term development. Finally. FNM reiterates its appreciation to the World Bank for the valued assistance received through the loans granted to the railway subsector. 41 Annex Il Page 5 of 19 COMMENTS ON THE PROJECT PERFORMANCE AUDIT REPORT ON EXECUTION OF THE FOURTH RAILWAY PROJECT WITH THE WORLD BANK (LOAN 1929-ME) I. Introduction 1. Ferrocarriles Nacionales de M6xico (FNM) has carefully reviewed and analyzed the Project Performance Audit Report on Execution of the Fourth Railway Project (Loan 1929-ME), which it considers to be a well-documented study by the Bank staff responsible for its preparation and recognizes its honesty and impartiality. 2. In general terms, the observations and conclusions relate to real situations and take into consideration the interrelationship between the various factors that impacted and affected the plans and programs. The report consists basically of two parts, one of which reviews the results of the investments made during the execution period of Loan 1929-ME while the other considers the role that the World Bank should play in FNM's future development. It can probably be considered a significant shortcoming of the report that it takes the position that the present railway organization has not changed substantially since 1972, whereas its physical conditions. dimensions and the aggravation of certain problems date from 1987, the year when the three Mexican railway companies (Chihuahua-Pacific, Sonora-Baja California and Pacific Railways) which formerly operated under their own managements, the sleepingcar service and the infrastructure works construction area, were consolidated into FNM. II. Role of the World Bank 1970-91 3. Before proceeding to give our opinion of the above-mentioned two parts, we consider it highly interesting to comment briefly on FNM's view of the World Bank's participation over the past 20 years. In actual fact, the loans were granted on homogeneous bases in each case and served as a practically permanent line of credit. However, even disregarding the economic results of the investments, including the financial benefits from loans at a very reasonable interest rate and long grace and amortization periods, it should be emphasized that the successive Bank missions over the period under review provided valuable advisory services on the problems of FNM's various components. Their recommendations were heeded by the Federal Government and in the majority of cases the pertinent steps have been taken to implement them. Specific instances such as the levelling of tariffs, revaluation of assets, and the assumption of liabilities illustrate this statement. 4. In other words, the World Bank has not only contributed funding, but has furnished valuable advice through its missions and has been an important factor in railway planning and evaluation. It is FNM's opinion that these relations are still close. 42 Annex II Page 6 of 19 III. Period 1981-85 5. As regards the analysis of traffic in 1981-85, the execution period of Loan 1929-ME, notwithstanding what the redort states this cannot strictly be considered a period of stagnation of freight traffic since the volumes achieved during those years were the highest in FNM1s history, not counting, of course, the railways incorporated in 1987. 6. The ton-km figure in 1981 was in fact 6.5% higher than that for 1980. In 1982 it dropped almost 12% on account of the economic crisis declared in that year, but freight then rose 13.9% in the following year. Even bearing in mind that the i983 figure includes the Ferrocarriles Unidos del Sureste traffic, the growth in real terms exceeded 10%. In 1984 and 1985, traffic growth was 2.0% and 2.8%, respectively; in other words, increases in traffic were posted in every year except 1982. As a result, the ton-km figure for 1985 was 12% higher than the 1980 figure and net tons were 9.6% higher than in the base year. 7. Moreover, annual average ton-km generated in 1981-85 were 35,575 million, which is 19.6% higher than the annual average for 1976-80. In net tons the difference is less significant (6.9%), since the first figure is also influenced by a considerable increase in average distance, which rose from 546.1 km to 611.0 km. 8. As regards locomotive and wagon procurement, this was partly justified by the fact already noted that freight traffic did not remain static; moreover, as the report itself mentions, the macroeconomic parameters that would have made it possible to foresee the sharp drop in demand after 1985. while a major step that was not taken was scrapping of locomotives and wagons that were no longer economically repairable, resulting in apparent oversizing of the two fleets, lack of shop capacity for proper maintenance and lack of funds for the purpose. 9. It should be added that although the numbers of locomotives and wagons procured were large, they show a clear downward trend over the period, since locomotive procurement by Nacionales de M4xico alone, including new units and kits for assembly in Aguascalientes, totalled 317, of which 237 were purchased between 1981 and 1983 and only 80 between 1984 and 1985; in 1986, 19 locomotives were built with kits purchased in 1985. 10. The position with wagons is similar, since Nacionales de M4xico purchased 8,358 between 1981 and 1985, almost two thirds of which were acquired in the first two years of that period. The as yet unconsolidated railways received 1,576 wagons between 1981 and 1983 but did not procure any more thereafter. The relevant statistics and graphs are presented in Annexes 2 and 3. 11. Regarding new infrastructure works, it must be taken into account to begin with that up till the end of 1988 this activity was a responsibility of the Ministry of Communications and Transport (SCT), so that resource management policy for works was extremely sensitive to the austerity measures 43 Annex II Page 7 of 19 instituted from year to year by the Federal Government, with the result that certain of the projects were suspended for a number of years or else allocated such small sums that progress was very slow. 12. Nevertheless, some of the main works initiated by SCT are now in operation. For example, the straightened 75-km stretch of the Mexico City- Veracruz line has been in use for some years now. The Monterrey line improvement, of 60 km. planned for dvuble track, is now operating over its first 40 km as single track. The 120-km Salinas-Laguna Seca section, forming part of the short Guadalajara-Monterrey route, is almost finished, as is the straightening of 50 km from Ajuno to Caltzontzin on the line to the port of IAzaro CArdenas. The new Mexico City-Quer6taro double-track line has also been operating for some time, independently of the electrification work under way. It is, of course, FNM policy not to start any new works until those in progress have been completed. - 13. Track rehabilitation and maintenance is a systematic investment, and although a large part of the system is considered to be in good shape at any given time, it is vital to continue this work so as to steadily upgrade track quality in terms of rail quality, type of superstructure and bridge capacity, in particular. Since application of track materials is classified as investment, it will depend on the budget authorized and the policies and priorities of the authorities. The annual average system-wide figure for track rehabilitation with new rail in 1981-85 was 312 km, while for second- hand rail it was 143 km/year. As regards bridges and drainage systems, the number per year was 185. For track rehabilitation in particular, the figures were modest. IV. Period 1986-91 14. Assessing the return on the investments made on the basis of traffic is commenting on history, since the railways served a derived demand, that is, there has to be a product to be moved, of either domestic or foreign origin, so the volume of demand served does not depend entirely or even for the greater part on actions by the railway. The period 1986-88 was when the real impact was felt of the crisis that began in 1982, characterized by successive and deep devaluations and high inflation, with difficult situations in certain sectori of the economy. 15. Between 1984 and 1990 freight traffic measured in net tons dropped by 20%, from 64.1 million to 51.0 million. The largest reductions resulted from better raw material management by the steel, cement, fuel oil and fertilizer industries, plus lower bulk agricultural traffic (mainly imports) and mineral traffic in general, all traditional mainstays of rail transportation. 16. During the same period, the railways were successful in increasing quantities of other products hauled; of particular significance today are the increased container traffic, assembly kits and finished vehicles for the auto industry, and movements of oil products (except for fuel oil), paper and board products, and general industrial goods. 44 Annex II Page 8 of 19 17. Internal factors that played a role in the decline in freight were the tariff increases in combination with deregulation of the trucking industry, lower quality in some services and lack of special.'.zed equipment. 18. It is accordingly considered that the reactivation of the economy under the present government will bring about, within two or three years, a strong upsurge in demand for transport services that the railways must be prepared to meet, if they are not to become a major bottleneck in the country's economic development. 19 Because of the railways' need to improve their financial position it has been decided to consolidate the freight service and to rationalize passenger service, with a view to obtaining a better financial return on the latter. The policy is only to retain the passenger services with high demand where costs can be recovered. Unproductive services and those not serving parts of the country totally without other alternatives will be withdrawn. 20. Locomotive and wagon procurement has been held to a minimum since 1987. High priority has been assigned to maintenance and rebuilding of tractive equipment and only 100 new locomotives have been added, under leasing arrangements, the purpose of which is to obviate the need to rent locomotives, which would otherwise have been necessary in 1990 and 1991. 21. Regarding freight equipment, emphasis has been placed on rebuilding wagons, in accordance with the budget funds available for each year, and only just over 600 were purchased between 1987 and 1991. The bankruptcy of Constructora Nacional de Carros de Ferrocarril is an outcome of those actions. Purchase or leasing of equipment by users is currently being encouraged, with the incentive of differential tariff treatment. 22. Unlike the other areas of activity, track maintenance was considerably expanded between 1986 and 1989, with a little over 2.400 km being rehabilitated with new rail and 1,211 km with second-hand rail. Work was also done on a total of 955 bridges during those four years. Budget constraints caused drastic cutbacks in 1990 and 1991. V. The Structural Change Program and the Future Role of the World Bank 23. Regarding the PPAR's most important conclusion, namely that in view of the large investments made in FNM over the past 10 years coupled with the stagnation and decline in traffic, FNM is now well set up with infrastructure and modern equipment, so that the restructuring program's future investment needs cannot be large and the Bank can therefore discontinue lending for railway investments, also bearing in mind that there is extensive scope for trimming operating costs and therefore improving financial and operating performance, we would ask the Bank to consider the following: 24. Although in the immediate past the national rail system's planning has been influenced by the phenomena noted and expressed by the report, the present Structural Change Program, drawn up in 1991 and still being reviewed and adjusted for effective implementation, looks at the entire range of problems to be resolved, and while it is true that reference was initially made 45 Annex II Page 9 of 19 to annual investment figures slightly greater in real terms than the annual averages posted in the recent past, these figures have now been reviewed and downscaled in light of the railways' present situation, with a view to covering the minimum requirements for safe operation and assuring a real and sustained improvement in operating and financial efficiency, such that in the medium term there will be a rail system capable of handling the foreseeable increase in goods traffic between Mexico and the U.S. and Canada. while also meeting the changing domestic demand in an environment of open competition and in a manner that complements other transportation modes. 25. The Structural Change Program is accordingly designed to proceed gradually, based on differing scenarios as regards resource availability, both political and financial from, basically, the Federal Government, and which have also been submitted for consideration by the World Bank, through its representatives. 26. One of these scenarios, termed "Rationalization without Growth." which in the opinion of the Bank's representatives is conceivably the most viable for FNM, represents the equivalent of the alternative proposed in the PPAR, comprising a combination of a conventional project for the World Bank, with lower basic investment requirements, while the specific policy, operating and labor-related aspects are prepared and implemented. 27. In this context, the basic investment program for 1992-95 amounts to US$1,491 million over four years, i.e. US$373 million per year (at the average exchange rate for Mex$3,159.3 to the dollar for 1992, according to the Mexican Government), which assumes assistance in the amount of US$400 million from the World Bank. 28. The main investment items included in this no-growth scenario are: (1) Trac reconstruction with new rail at an essential minimum of 350 km per year, as compared with the annual average of 565 km between 1986 and 1990. (2) Rebuilding of 71 locomotives. ()) In response to the Bank's comments, installation of CTC will only be considered between Monterrey and Nuevo Laredo, so that the system will be in use from Mexico City and to give this route a high haulage capacity. (4) Assignment of priority to modernization of the telecommunications systems and of the operations information and control systems. (5) Completion only of infrastructure works that are well advanced, so that they can be brought quickly into service and generate the greatest benefits. 29. Of the total investment envisaged, 46% relates to procurement of foreign-made components that will have to be paid for in foreign exchange, so 46 Annex II Page 10 of 19 that it will be essential for FNM to be able to continue to rely on loans such as those granted by the World Bank and other international institutions. 30. In this connection it must be borne in mind that more than 3,000 km of the basic track system totalling 13,000 km still consists of 100-lb/yd rail that is over 30 years old and appreciably worn. Moreover, around 1,500 km of the rail of the 13,000-km basic system will reach the end of its useful life in the coming four years, on the basis of the relatively low density of present traffic. In addition, there are in this basic system around 3,100 temporary or low-capacity bridges and culverts that will need strengthening to ensure and standardize train operating conditions. 31. It will therefore be necessary to continue laying heavier rail in our tracks in order to prevent accidents due to rail breaks and to support operation of trains with up to 120-ton cars. 32. The track rehabilitation program for 1992-95 only envisages 350 km of renewed track per year, which is the minimum required to prevent the present backlog in the basic system from growing. As regards bridges, because of their age and low bearing capacity a strengthening program has been under way since the 1970s and although some progress has been made the basic system still includes the large number of temporary and low-capacity bridges already noted, which is an obstacle to the running of heavy trains. The 1992-95 program, which aims to upgrade 98 structures per year to Cooper E-80, is only equivalent to covering 13% of the total needs. 33. It should also be noted that track and bridges chiefly on the lines to frontiers and ports need to be improved in light of the prospective signing of the United States-Canada Free Trade Treaty, which will involve exchange of larger and heavier equipment so adequate high-capacity infrastructure will have to be in place. 34. Among new technologies for lengthening the useful life of rail special attention is being paid to grinding. With a view to using this technique in Mexico we plan to procure a rail-grinding train with funds from Loan 2575-ME and the desirability of obtaining another will also be studied. 35. Concrete ties have been found to be appreciably more economical in Mexico than wooden ones; a concrete tie lasts two and a half times longer than a wooden tie while its cost -- including fastenings -- is only 14% more. Because of their greater rigidity and stability, tracks with concrete ties offer greater carrying capacity. The criteria for track renewal programs are presented in Annex 1. 36. Regarding FNM's telecommunications system, which was installed with Bank assistance, it can be stated that although it has not been fully utilized in some applications, such as train dispatching by radio, it has given valuable service in both operating and administrative activities through the use of selective telephony for train dispatching and of walkie-talkies for communication between train crews and yard staff, thereby helping to expedite operations and increase safety, while utilization of the telephone network enables communication with a large part of the system. 47 Annex II Page 11 of 19 37. In signalling, installation of CTC (Centralized Traffic Control) has been concentrated in the trunk lines carrying the most traffic, such as Mexico City-Guadalajara and Mexico City-Quer6taro, which because of their characteristics need more efficient dispatch systems than those based on telegraph. telephone or radio. However, we agree with the PPAR's recommendation that less costly systems more consistent with the traffic volumes on our lines should be used. 38. Moreover, while it is true that in 1985 there was a small surplus of locomotives and wagons, it must be borne in mind that virtually no loco- motives and only a very small number of wagons have been added to the fleet in the past six years. These have not been sufficient to make up for those lost due to accidents or scrapping during the period. 39. As to the other points mentioned in the PPAR and which require action in order to accomplish a thorough restructuring of the railways, namely labor questions chiefly in the area of operation and government policies for the sector, these matters have been under review since 1991 and negotiations are expected to continue during 1992 with the aim of reaching favorable agreements in the course of the year that will make it possible to implement the needed changes envisaged in the Structural Change Program. 40. In this connection the first definite steps were taken in 1991, with funding authorized by the Federal Government in the amended budget for that year. The rationalization of passenger services and withdrawal of unproductive services is proceeding; to date, 71 services have been withdrawn (passenger, mixed and local freight services), which is equivalent to 142 trains in both directions. The number of positions eliminated amounted to 1,181 with the closing of the Aguascalientes and San Luis Potosi foundry works plus cutbacks in the staff of the Sleepingear Service, the express agencies, general services and other areas in general. 41. It is considered that the negotiations for continuing this process can carried out in the short term, without needing "decades" for the purpose as the PPAR states. 42. In regard to operations, while more qualified staff are needed in this area, it must be recognized that there are very few people outside who can take the place of an experienced railway man possessing the empirical knowledge needed to fill the senior positions. This constraint can be eased by means of permanent advisory services by technical specialists with experience in foreign railways, backed by the Directorate General and senior staff of the Office of the Assistant Director General for Operations, in the implementation of the measures recommended by the consultants. VI. Recommendations 43. As already noted, we have studied the World Bank's PPAR with great interest and consider it to be frank and impartial; however, it is FNM's opinion that it needs to continue investing in capacity expansion and maintenance and in the search for new technologies that will enable the railways to become more competitive by providing better service and thereby 48 Annex I Page 12 of 19 doing their part toward meeting the challenge contained in the reactivation of the economy as envisaged in the National Development Plan. 44. FNM would like the Bank to participate in its Structural Change Program with a loan of US$300-400 million and welcomes the idea put forward in the PPAR to the effect that financing might be provided to fund compensation payments, which are expected to amount to some US$200 million, together with the proposal to deploy more consultancy services to improve key areas. 45. It would therefore be a setback for FNM to receive nothing or only very small-scale financing from the Bank, since that would raise the cost of the investment program as regards the local component and debt-service payments, while the financing would also be concentrated in virtually one single bank. 46. Accordingly, by way of recommendation, it is proposed that the negotiations begun for securing a new IBRD loan be continued and that the final decision on the amount of the loan and the investment categories be made subject to the progress accomplished in the Modernization and Structural Change Program during 1992, which will depend on the support allocated by the Federal Government to the Railways for their medium and long-term development. 47. Finally, FNM reiterates its appreciation to the World Bank for the valued support received through the loans granted to the railway sector, which - as the PPAR notes -- have practically constituted a permanent line of credit for the past twenty years and have been a major factor in making FNM the significant asset it is to the country in that it possesses a system consistent with Mexico's present and foreseeable future needs. 49 Page 13 of 19 ANNEX I TRACK REHABILITATION CRITERIA The Mexican Railway System classifies track into six categories, from 1 to 6, according to annual tonnage and train speed. As regards track route, since Mexico has many stretches of line with considerable curves and gradients, tracks are classified in three groups. For preparing track renewal programs with new rail and also rehabilitation programs using second-hand rail, an estimate is made of the useful life of the rail in the various stretches, in accordance with the track category and the characteristics of the route and based on FNM's experience, according to which the life of the rail can range from 7 years in renewed and 7 years in track with second-hand rail to 44 years in renewed track and 24 years with used rail. Verification of the physical state of the rail and of the other components of the track makes it possible to refine the medium-term programs and finally, for each annual program, all the scheduled stretches are checked one year before the program is started to ascertain that they are indeed those of greatest priority. As regards the number of internal defects, for FNM this is not the main parameter in deciding on track renewal or rehabilitation as it is in U.S. or Canadian railways, since annual tonnages are lower in Mexico. However, the external physical state of rail in terms of wear, external deformations, burns and any external defects in general is the main reason for replacement. In Mexico from 8,000 to 10,000 km of track is checked every year with the Sperry car, to determine the condition and deterioration of the rails and in particular to take immediate steps to replace any with internal defects. Nevertheless, the useful life of rail is determined more by fatigue, wear and plastic deformations in the joins. In Mexico's only Category 1 line, the Mexico City-Quer4taro line, the criterion for maintenance is that there must not be more than 30 internal defects per 100 km of line. However, this does not mean that the entire track will be renewed, but only that some defective rails and thermal welds with defects will be replaced. For Category 6 lines, the tolerance set per 100 km of track is up to 320 internal defects. 50 Annex II Page 14 of 19 LIFE OF SECOND-HAND RAIL ACCORDING TO FNM's EXPERIENCE ROD. ' TRACK DURATION CATEGORY (YEARS) WINDING WITH CURVES EXCEEDING 60 4 7 WINDING WITH CURVES EXCEEDING 60 5 9 WINDING WITH CURVES EXCEEDING 60 6 14 SEMIWINDING WITH CURVES BETWEEN 3.50 AND 60 3 10 SEMIWINDING WITH CURVES BETWEEN 3.50 AND 60 4 12 SEMIWINDING WITH CURVES BETWEEN 3.50 AND 60 5 13 SEMIWINDING WITH CURVES BETWEEN 3.50 AND 60 6 18 WITH CURVES LESS THAN 3.50 3 14 WITH CURVES LESS THAN 3.50 4 15 WITH CURVES LESS THAN 3.50 5 18 WITH CURVES LESS THAN 3.50 6 24 TRACK CLASSIFICATION: CATLGORY 1, 2, 3, 4. 5 OR 6 BY ORDER OF IMPORTANCE IN TERMS OF TRAFFIC DENSITY AND SPEED OF OPERATION 51 Annex II Page 15 of 19 ANNEX 2 LOCOMOTIVE PROCUREMENT _NACIONALES DE MEXICO 1 Other rtw Total Rait system Year Assebted New Total Rebut it Procurement SY 1980 8-;- -- 26 :2 22'1 1981 40 80 120 44 164 1982 39 39 78 12 90 1983 5 34 39 0 39 1984 28 30 58 0 58 1985 3 3 0 3 1986 19 98 27 1981-1986 134 183 3_ 1t6 -0- 1987 0 29 0 0 1988 0 162 0 0 1989 0 25 0 0 1990 48 48 32 0 48 1991 52 52 33 0 52 1987-1991 1Q _ -10 _* TOTAL4192A489 77 I I NB: - LOCOMOTIVES ASSEMBLED IN 1986 BUILT FROM 1985 KITS. - LOCOMOTIVES LEASED IN 1990 AND 1991. - 1991 FIGURE DOES NOT INCLUDE 15 LOCOMOTIVES DENOTED TO FNM. TRACTION EQUIPMENT PROCUREMENT 1980 - 1991 UNITS UNIDADES 250 1 200 150- La 100- 50- 0 W __ 80 81 82 83 84 85 86 87 88 89 90 91 ii 0 OTHER RLWYS N. de M. TOTAL SYSTEM 53 Annex II Page 17 of 19 FREIGHT EQUIPMENT PROCUREMENT r - -~----------------- r-------------------r------------------- -i Y .N. de M. OTHER RLWYS TOTAL RAIL SYSTEM - ------ - --------------- --------------------- --------------------1 1980 2.948 463 3.411 1981 2,805 776 3,581 1982 2,354 509 2,863 1983 1,577 391 1,968 1984 1,022 0 1,022 1985 600 0 600 1981-1985 8.358 1.676 10.034 1986 400 0 400 19871 418 0 418 19881 100 0 100 1989 0 0 0 1990 20 0 20 19911 0 80 801___ 1986-1991 1,018 0 1018 TOTAL . 12. _1392 14.&463 -- ---------9-------------9---2- January 13, 1992 FREIGHT EQUIPMENT PROCUREMENT 1980 - 1991 UNITS UNIDADES 4,000 3,000 Ln 2,000 1,0 - 80 81 82 83 84 85 86 87 88 89 90 91 OTHER RLWYS N. de m. TTAL SYSTM FREIGHT EQUIPMENT PROCUREMENT BY N. de M. GONDOLAS 10% WAGONS - WAGONS 41% 25% GONDOLAS OTHER 33% 9% OTHER HODPERCARS 21% HOPPERCARS 18% 1981 - 1985 9186 - 1991 8,358 UNITS 1,018 UNITS o 56 Annex III Page 1 of 3 ANNEX III: COMMENTS ON EVENTS SUBSEQUENT TO RAILWAY IViI 1. Generally I believe you have raised the principal issues in designing a Bank strategy for future relations with NdeM and you are correct in expressing the problematic "tug-of-war" between the railway enthusiasts of NdeM and some of SCT versus the Ministry of Finance (MOF) and Planning and Budgeting (SPP). Your analysis of the Fourth Project is also correct. I am concerned, however, that despite your discussion of Mexican railway history, you are missing critical material relating to what happened in the mid-1980's with regard to many of the issues you raise and which set the stage for Railways V and the events that occurred soon after that loan. These have implications for assessing the railway's performance and designing an appropriate future lending strategy. 2. Firs: Moves for Restructuring. The Government pressed for a "restructuring" of NdeM in the mid-1980's. The initiative came from the Government, not the Bank, and developed into a broad negotiations of the "contract plan" between DOF, SPP, SCT and NdeM. At this point NdeM was not consolidated yet and the focus was solely on NdeM (excluding Pacifico, etc.). The new Railway V loan was held up until a full agreement had been worked out and signed between the Government and the Railway. Board presentation was further held up until there was agreement with DOF on railway tariff policies and a phased in program for bringing all tariffs in line with estimated variable costs as a minimum. Passenger tariff policy focussed on getting first class rates right but left the lower class tariffs to Government Public Service Obligations. The negotiations over the agreement extended over the course of a year with direct intervention by the various Secretaries and the President. The concept was to place NdeM in a commercial environment which would provide the needed pressure to tackle the internal organizational and technical issues. 3. The Bank's analysis at that time did not suggest that there was sericus overstaffing. I believe our official numbers for NdeM were about 65,000 and while we had heard about temporary laborers, the numbers from accounting did not show significant costs. There was great agreement on the subject of restrictive work rules, mainly for the workshops and the train crews (as well as for middle- management positions having to come from the specific trade limiting cross- moves). Intense labor negotiations ensued focussing on the workshops, trying to reduce the number of job classifications. This rationalization of workshop positions was negotiated and pay incentives arranged accordingly, related to training. The Railway realized that real wages had declined by about 35% and would be raised shortly so why not move early and try to get concessions. 4. It is true that many railway enthusiasts thought that this restructuring was an opportunity for greater investment. In fact, however, the investments were scaled back very severely (or were supposed to be). New locomotive orders from the U.S. were curtailed and a number of Vias Ferreas works were stopped or reduced significantly based on economic analyses and helped by the basic budgetary crunch. Even the classification yards were looked into to determine I/ This annex is based on the written comments of a Bank staff member economist who worked for several years on the Mexican railway lending program. It is aimed primarily at the sector lending operation (Rail V) subsequent to the project under audit and is included here to provide a more balanced view of a project that has yet to be audited but was, of necessity, addressed when considering the total context of the Railway IV lending operation. 57 Annex III Page 2 of 3 how to bypass some and get the main ones working more efficiently. Expertise was brought in on track, capacity analysis and yards to joint Bank missions (mainly from CN before and during the loan). 5. The Government's commitment to the contract was clear. A year after the contract, the President of NdeM was removed and a non-railway man appointed. Changes were made especially in the finance department and reportedly in commercial and traffic department was well. There was to be decentralization with regard to day to day operational matters. Unfortunately, the Government (and the Bank) assumed that the issue was one of management which could be resolved by a good manager and that railway's role in passenger and freight was unlimited with good management. 6. NdeM ConsoLidation: One of the major unexpected events was the consolidation of the various railways soon after Railways V was effective. The Bank had been pushing this for years but felt it would not happen at this point because of legal issues and because of labor agreements on the Pacifico. It came as a surprise for which we were unprepared. Your report does not make much of the consolidation but I am left wondering if it has had a significant impact on many of the efficiency issues which you raise, i.e. labor productivity, workshops, low traffic lines. 7. My main question would be: how much of your productivity analysis would be changed if we just looked at the NdeM original system. To what extent are we seeing the effects of agreements made by government to get the unions to agree to the consolidation? Was it worth the price? 8. Investment Needs: A large part of your argument on appropriate Bank strategy is based on the your assertion that there were no significant investment needs to warrant project lending. I would suggest you look more carefully at this. You refer to NdeM as a relatively minor player in the transport system (although.with some greater potential). Yet it is carrying 15-20% of traffic, not insubstantial, representing almost 50 billion TU's. NdeM carried significantly high volumes on mainlines and sufficient traffic on secondary and tertiary lines to warrant minimal track servicing. Basically I think you play down zhe railway too much. While I agree with your criticism of over optimistic forecasts and the dangerous over investment in .The Vias Ferreas works., locomotives and CTC, there was still significant need for improvements on high traffic sections, track rehab and loco and wagon replacement and rehab. Most investment, under Railway V, were for maintenance and rehabilitation (including railway and equipment), equipment replacement, spare parts for loco rehab. Major investments were pushed out to later years as traffic and macroeconomic uncertainty would have stabilized in one way or another. In any case, there were still substantial investments required just to keep running with the existing traffic. 9. The locomotive problem was difficult. The Bank was hoping that availability would rise under traffic pressure and that the mothballed GE electric locomotives would finally come on stream. As -teither of these situations occurred, locomotive capacity was hurting railway services increasing pressure to buy (probably a cycle that led to the earlier purchases during the period of Railway III). 58 Annex III Page 3 of 3 10. Demand Forecasts: The macroeconomic adjustment had more and immediate impacts on railway traffic flows than anyone could have imagined. First came the growth of domestic grain production taking away the high imports of grain through the ports. This was already clear by the mid-1980's but there was still hope that Ndeft could capture certain traffics such as sorghum, which I presume never occurred. Multiple origins and destinations worked to the trucks' advantage (with or without deregulation). The steel industry restructuring was taken into account but the fertilizer restructuring was a major blow. It also appears to me that with the growth of non-traditional exports, traffic in general in Mexico is not radiating to and from Mexico City but many new traffic directions to and from and between secondary cities is occurring. This probably led to the break- up of regulated trucking since the trucking zones were designed for the radial movements. We missed the boat on this one but did recognize the underlying instability of the traffic situation. 11. Lending Strategy: The main question you raise and which deserves a broad discussion within the Bank and Mexico is whether to look for an adjustment operation or a sector investment operation. I would agree that if the issues are those of labor redundancy, regulatory policy or major institutional restructuring such as privatization, a SECAL would be more appropriate. There are investment needs and technical needs for reforming railway operations, however, that require a lending instrument that keeps our technical people in touch with theirs. This would suggest a hybrid operation to get the best of both worlds. The question you are raising is whether we should stay away from lending until the SECAL type issues are ready to be addressed. I'm not convinced of that. What I would suggest, however, is that prior to any investment operation, concrete actions on technical issues such as the workshops and locomotive availability be taken by NdeM (and not studios or paper agreements). EnsanimANTTUCIO gAll A NNgORAuEfNl Nu.M CEe TX l PCs0 8AC10OALn DEltt -"ANANTONmIP OT A R AI~ P, AE u u ORHR ONIABCALIFORN "'tATEAU CHINVi dAlAL PACIF-C- M~~ RN AllROADSArA Ato. SCaNG TI-dl ZAAEA . iochIZ Tlo:C