UKRAINE: UKRZALIZNYTSIA (UZ) MODERNIZATION STRATEGY Policy Note 2: Managing loss-making mainline passenger services June 24, 2019 Ukraine Railways Policy Note 2 June 24, 2019 Acknowledgments The Policy Note was prepared by a World Bank team led by Baher El-Hifnawi (Program Leader for Infrastructure and Sustainable Development) and Paul Amos (Senior Railway Advisor) and a team consisting of Yevhen Bulakh (Transport Specialist) and Anton Hagen. The primary author of this Note is Richard Bullock (Senior Railway Advisor). The Note benefitted from comments of the Peer Reviewers: Jean-Francois Marteau (Lead Transport Specialist) and Piers Vickers (Deputy Economic Adviser, European Investment Bank). The team is grateful for guidance received from Satu Kahkonen (Country Director), Karla Gonzalez (Europe Transport Practice Manager), Ross Pavis (Senor Operations Officer) and Fiona Collins (Lead Transport Specialist). The team would also like to extend its appreciation to the representatives of the Government of Ukraine and Ukrzaliznytsia for their strong collaboration and constructive discussions. i Ukraine Railways Policy Note 2 June 24, 2019 Contents Acknowledgments.................................................................................................................................... i Preface ................................................................................................................................................... iii Glossary of terms and acronyms ........................................................................................................... iv Summary of Findings............................................................................................................................... 1 1. The Services.................................................................................................................................... 2 1.1 Service classification ........................................................................................................... 2 1.2 Basic Service Characteristics ............................................................................................... 3 1.3 Sample selected for analysis ............................................................................................... 3 2 Passenger Service Costing Framework........................................................................................... 7 2.1 Introduction ........................................................................................................................ 7 2.2 Passenger Service Cost Structure ....................................................................................... 7 2.3 Cost Components ................................................................................................................ 9 3 UZ Passenger Service Costs .......................................................................................................... 14 3.1 2017 Financial Accounts.................................................................................................... 14 3.2 Form 10 ............................................................................................................................. 15 3.3 Analysis of Form 10 costs to generate aggregate costs by activity .................................. 17 3.4 Derivation of unit costs ..................................................................................................... 18 3.5 Application of unit costs to selected services ................................................................... 19 4 Analysis of results ......................................................................................................................... 21 4.1 Introduction ...................................................................................................................... 21 4.2 Overall Service Viability..................................................................................................... 21 4.3 Financial analysis by class ................................................................................................ 22 4.4 Impact of variations in train composition, speed and traction type ................................ 24 5 Possible strategies ........................................................................................................................ 27 ii Ukraine Railways Policy Note 2 June 24, 2019 Preface This Policy Note is one of a series of individual papers originating in a request from Ukraine’s Ministry of Infrastructure (MoI) and JSC Ukrzaliznytsia (UZ) to the World Bank to address specific topics concerning Ukraine’s railway sector. The Policy Notes address the following topics. 1. Railway market opening for cargo services: progress in the meeting rail commitments in the EU-Ukraine Association Agreement, reorganization of UZ, Ukraine’s readiness for and implications of market opening, pre-requisites to avoid leaving UZ in an unfavorable situation. 2. Loss-making long-distance passenger services: service costing, institutional and financial options for providing sustainable transport passenger services for long distance travel. 3. Selected Freight Business Issues: specific matters on which Bank advice has been sought including cargo tariffs, customer service and perceptions, and operating efficiency. 4. Debt management: options for UZ to restructure its debt and reach a financially stable situation. 5. Infrastructure asset management and prioritization of investment. Asset management strategy and life-cycle costing in the renewal and reconstruction of UZ’s railway infrastructure network. This Policy Note is for Topic 2: Managing Loss-Making Long-Distance Passenger Services. Ukraine runs a wide range of passenger services. They form two main groups, regional (called primiiski or prigorodne) and long-distance. The World Bank developed a methodology in 2015 for costing the services which could be used as the basis for a Public Service Obligation (PSO) payment from the local or central government and used this as the basis for analyzing the regional services within Lviv oblast. This work is now being further developed by consultants working under EIB. This Policy Note is concerned with long-distance services. Within this group there are both conventional loco-hauled services and express InterCity services using multiple-units. The express services are considered by UZ to cover their attributable costs and this note concentrates on the conventional loco-hauled services. Some of these recover their costs from fare revenue and some do not, with one class of fare (Platzkart) regulated by the Government (GoU). Currently losses are covered by internal cross-subsidy within UZ, but this will become progressively more difficult with market opening. This note focusses on testing whether, as suggested by UZ, compensation for the regulated Platzkart fares is a valid and practical approach to defining PSOs in the long-distance market. iii Ukraine Railways Policy Note 2 June 24, 2019 Glossary of terms and acronyms CURRENCY EQUIVALENTS Exchange Rate (May 2019) Currency Unit: Ukrainian Hryvnia (UAH) USD 1 = UAH 26.4 ACRONYMS FSU Former Soviet Union GoU Government of Ukraine gtkm Gross tonne-km MoF Ministry of Finance MoI Ministry of Infrastructure pkm Passenger-km PSO Public Service Obligation TAC Track access charges (charges levied for use of railway infrastructure) ntkm Net tonne-km TOC Train operating company (‘Carrier’ in Ukrainian Law) UAH Ukrainian Hryvnia USD United States Dollar UZ Joint Stock Company, JSC Ukrzaliznytsia (registered October 2015) iv Summary of Findings Background 1. Ukraine runs a wide range of main-line passenger services. Some recover their costs from fare revenue and some do not, with fares for one class of travel (Platzkart) regulated by the Government of Ukraine (GoU). Currently losses are covered by internal cross-subsidy within UZ, but this will become progressively more difficult with market opening. This Policy Note examines whether, as suggested by UZ, compensation for the regulated Platzkart fares is a valid and practical approach to defining PSOs in the long-distance market. Key findings 2. The conventional mainline passenger services carry two main classes of passenger service: Kupe and Platzkart. Kupe fares are more than double those for Platzkart and generally recover their ‘above-rail’ operating and capital costs i.e. the cost of operating the service and of renewing the rollingstock when it falls due. Platzkart services generally do not. Policy options for Platzkart services 3. The normal response to improving the financial recovery of Platzkart services is simply to increase the fares. However, Platzcart fares in Ukraine are heavily regulated by the government for social reasons and that policy seems unlikely to change in the short and medium term. The main options available to UZ to improve the financial performance of the Platzkart service form three groups: • Reduce the unit cost of the inputs provided for these trains (e.g. locomotive maintenance cost, vehicle maintenance and so on). However, while some reduction is attainable, the Bank team’s conclusion is that any reduction from this source is likely to be modest in the long-term. • Reduce the volume of inputs required. UZ is already reviewing the on-board train crew requirements which can reduce the total cost by 15-20 percent. Another option is to review the train composition to better align capacity with demand; this is particularly relevant where there are sections of heavy demand while other sections are relatively lightly loaded. • Obtain financial compensation from the government for providing Platzkart accommodation at the current tariff (assuming fare increases are not available). If this were to be done, it should be implemented on a simple contract basis, with the GoU receiving all Platzkart revenue and UZ being paid a standard rate per Platzkart vehicle-km provided, plus a commission for collecting fares on GoU behalf (a gross cost contract). Alternatively, a net-cost contract could be considered whereby the Government is paying the shortfall in cost. 1 Ukraine Railways Policy Note 2 June 24, 2019 1. The Services 1.1 Service classification Passenger services in Ukraine, as in all ex-FSU railways, form two groups – long-distance services and primiiiski1 (prigorodne in Russian) which are local and regional services. Few primiiski services have any prospect of ever being commercial (recovering the costs of their operation) in their own right and under the draft new Railway Law of Ukraine they are likely to eventually be tendered by the relevant local governments as happens in many countries in Europe. They were analysed by the World Bank in 2015 and are now being studied in more detail by EIB consultants. The long-distance services form eight groups (Table 1.1), operated by two separate units within UZ. The High-Speed Company services are generally considered by UZ to be covering their attributable costs and this report, at the request of UZ, is concerned with those operated by the Passenger Company. Table 1.1: UZ passenger service classification Train type Characteristics Passenger company: Night Express Services with a mixture of accommodation: • SW (two-berth sleeping car) • Kupe (typically four-berth compartments) Night Firm • Platzkart (open berths, two high) Most of these services operate over long distances and transit times thus include (despite their name) Night substantial periods of day travel. They are nearly all locomotive-hauled with many carriages International trains Cross-border services which operate under a range of agreements, some of which carry domestic Ukraine passengers. Includes both long-distance services (e.g. Kiev – Moscow) and short-distance (e.g. Ukraine – Slovakia) Regional trains Medium-distance service with sitting accommodation Day trains Also called Intercity, sometimes with carriages that convert from berths to sitting accommodation during the day High-speed Company 1 These are normally translated as ‘suburban’ but they are essentially regional trains, stopping at all stations, with low fares and basic accommodation. They are not confined to the immediate vicinity of large cities and some go considerable distances of 200 km or more. 2 Ukraine Railways Policy Note 2 June 24, 2019 InterCity (IC) Express (90 km/hr). Offers two classes: First (unregulated fares) and second (regulated) InterCity Plus (IC+) Express (110-120 km/hr). Offers two classes: First (unregulated fares) and second (regulated) The three Passenger Company ‘night services’ will shortly be rebadged as Comfort, Standard and Economy, with more consistent rollingstock standards. All three types will continue to carry Kupe and Platzkart carriages but sleeping cars will only be available on Comfort and Standard services. The Platzkart and second-class fares are regulated in all the types of train in Table 1.1. The fares on the Express services are around double those of the Passenger Company and their revenue is comfortably recovering their operating costs. However, many of the Passenger Company services have difficulty covering their costs. For some services, average revenue does not even cover short-run marginal costs and the more passengers they carry, the greater is the loss. The paper analyses the current cost structure and identifies areas where economies could be made (and are being already made by UZ). It also discusses the possibility of restructuring some of the services and possible ways of implementing any payment from GoU to compensate for the regulated Platzkart fares. 1.2 Basic Service Characteristics Table 1.2 summarises the basic characteristics of the long-distance services operated by the Passenger Company (i.e. excluding the Regional Trains and Day trains in Table 1.1). Table 1.2 Summary characteristics of long-distance Passenger Services 2018 Sleeping Kupe Platzkart(1) Total Number of services - - - 110 Train-km p.a. (million) - - - 53 Carriage-km (million) 45 365 419 829 Passengers (million) 0.9 12.2 21.4 34.5 Passenger-km (million) 608 8061 11453 20123 Revenue (UAH million) 730 5723 3436 9889 Average distance (km) 647 659 536 582 Yield (UAH/pkm) 1.20 0.71 0.30 0.49 (1) includes a small number (0.1 million) of sitting passengers 1.3 Sample selected for analysis A sample of 94 trains (out of the total of 110 operated) was selected for analysis. These cover a wide range of routes and range from 41 km (on the UZ network) to 1657 km. UZ has a wide range of detailed passenger service data available (e.g static and dynamic occupancies by carriage type by service) and the data used in this report has drawn on datasets covering passengers, revenue and operating statistics in routine use within UZ. 3 Ukraine Railways Policy Note 2 June 24, 2019 Table 1.3 Sample long-distance services selected for analysis Pass-km Revenue Dist- Time Speed Cars From To (mill) (UAH mill) ance (mins) (km/hr) K(1) PK(1) K(1) PK(1) (km) K KP 1/2 Ів Франківськ Костянт 131 0 94 0 1453 1230 71 10 0 3/4 Запоріжжя Ужгород 96 0 69 0 1440 1240 70 10 0 6/5 Київ Москва 65 54 46 16 354 292 73 12 5 8/7 Одеса Харків 79 0 56 0 762 555 82 10 0 7/8 Київ Івано-Франківськ 112 0 80 0 712 681 63 8 0 7/8 Харків Одеса 79 0 56 0 762 555 82 10 0 12/11 Львів Одеса 150 0 107 0 745 608 74 10 0 12/11* Київ Новоолексіївка 263 97 187 29 861 737 70 15 3 13/14 Київ Солотвино 152 197 108 59 1017 1205 51 9 6 15/16 Харків Рахів 116 14 82 4 1336 1155 69 9 1 17/18 Харків Ужгород 114 15 81 4 1342 1160 69 9 1 19/20 Харків Москва 2 1 2 0 41 40 61 4 1 20/19 Київ Лисичанськ 64 0 46 0 737 725 61 10 0 24/23 Одеса Москва 89 174 63 53 1008 930 65 7 6 26/25 Одеса Яремче (Рахів) 139 158 99 48 1024 1170 53 7 5 29/30 Київ Ужгород 158 0 113 0 853 732 70 10 0 31/32 Київ Рига 9 0 6 0 240 142 101 10 0 36/35 Одеса Пшемисль 34 0 24 0 828 985 50 4 0 37/38 Київ Запоріжжя 114 0 81 0 657 1185 33 8 0 41/42 Дніпро Трускавець 263 353 187 107 1198 1245 58 10 9 43/44 Ів Франківськ Київ 158 220 113 66 712 1210 35 11 9 45/46 Ужгород Лисичанськ 222 348 158 105 1657 1880 53 7 7 49/50 Київ Трускавець 222 78 158 24 756 704 64 15 3 54/53 Київ Санкт-Пет. 9 5 6 1 293 229 77 7 1 55/56 Хмельницький Москва 38 106 27 32 720 1195 36 7 6 57/58 Ковель Одеса 96 170 68 51 889 977 55 8 8 59/60 Одеса Харків 109 246 78 74 762 780 59 7 10 62/61 Миколаїв Москва 14 72 10 22 740 495 90 3 7 63/64 Харків Київ 200 0 143 0 485 418 70 18 0 63/64 Дніпро Одеса 111 126 79 38 631 664 57 8 6 70/69 Маріуполь Львов 210 373 149 113 1585 1720 55 7 8 72/71 Запоріжжя Київ 72 89 51 27 633 699 54 10 8 74/73 Львів Москва 35 104 25 31 976 1400 42 3 4 74/73 Кривий Ріг Москва 34 52 25 16 552 696 48 5 3 79/80 Дніпро Київ 221 19 158 6 532 447 71 18 1 81/82 Київ Ужгород 333 0 238 0 910 950 57 18 0 81/82 Харків Новоалексіївка 49 105 35 32 530 468 68 6 7 83/84 Київ Маріуполь 151 287 108 87 1032 1073 58 7 10 86/85 Львів Новоолексіївка 78 117 56 35 1415 1440 59 5 5 87/88 Ковель Новоолексіївка 55 91 39 27 1342 1342 60 5 5 91/92 Костянтинівка Одеса 42 160 30 48 940 990 57 4 11 92/91 Львів Київ 205 0 146 0 572 990 35 16 0 92/91 Одеса Константинівка 102 309 72 93 940 990 57 5 11 95/96 Харків Маріуполь 13 20 9 6 702 765 55 2 2 96/95 Маріуполь Бахмут 38 51 27 15 942 1070 53 5 4 97/98 Ковель Київ 87 138 62 42 552 600 55 8 8 101/102 Київ Херсон 174 137 124 41 642 720 54 12 6 4 Ukraine Railways Policy Note 2 June 24, 2019 Pass-km Revenue Dist- Time Speed Cars From To (mill) (UAH mill) ance (mins) (km/hr) K(1) PK(1) K(1) PK(1) (km) K KP 104/103 Жмеринка Маріуполь 186 497 132 150 1299 1280 61 7 13 106/105 Одеса Київ 258 0 184 0 654 525 75 17 0 106/105 Дніпро Москва 12 30 8 9 350 350 60 4 3 108/107 Одеса Ужгород 158 246 113 74 1032 1150 54 8 8 110/109 Львів Херсон 276 190 197 57 1135 1170 58 12 5 111/112 Харків Львів 262 146 187 44 1110 1200 56 12 4 116/136- Київ Покровськ 102 199 73 60 880 1008 52 6 8 115/135 117/118 Чернівці Київ 115 123 82 37 573 1206 29 10 6 120/119 Запоріжжя Львів 124 185 88 56 1243 1482 50 5 5 122/121 Миколаїв Ровно 138 366 98 110 973 985 59 7 12 123/124 Київ Костінтинівка 181 239 129 72 800 885 54 11 9 125/126 Київ Костінтинівка 102 232 73 70 680 715 57 7 10 127/128 Ковель Харків 63 107 45 32 1143 1328 52 8 8 129/130, Суми Зернове 1 5 1 2 262 314 50 1 1 127/128 134/133 Миколаїв Івано- Франківськ 115 320 82 97 1339 1490 54 5 9 135/136 Чернівці Дністр 0 405 0 122 1096 1230 53 0 12 137/138 Хмельницький Лисичанськ 119 369 85 111 1104 2390 28 6 12 139/140 Київ Кам.-Под. 24 54 17 16 458 778 35 6 8 139/140 Дніпро Лисичанськ 45 114 32 34 657 777 51 7 8 141/142 Одеса Маріуполь 26 58 18 17 558 1755 19 6 9 142/141 Львів Бахмут 247 245 176 74 1467 1250 70 11 5 143/144 Харків Санкт-Пітер 5 8 3 2 578 631 55 2 1 143/144 Запоріжжя Івано-Франківськ 81 104 58 31 852 940 54 10 8 145/146 Київ Ізмаїл 94 244 67 74 934 1005 56 5 9 145/146 Львів Ужгород 17 15 12 4 287 1010 17 4 2 148/147 Одеса Київ 100 185 72 56 669 800 50 8 8 149/150 Кременчук Ів.Фран, Ворохта 73 129 52 39 1190 1340 53 8 8 317/318/ Запоріжжя Одеса 35 38 25 12 639 906 42 6 4 617/618 357/358 Київ Рахів 85 131 60 40 835 1125 45 6 6 375/376 Харків Херсон 136 132 97 40 763 1080 42 11 6 601/602 Львів Солотвино 36 48 26 15 447 738 36 5 4 606/605 Львів Рахів 8 41 6 13 279 480 35 2 6 668/667 Ковель Чернівці 17 75 12 23 525 851 37 2 6 737/738 Київ Запоріжжя 88 0 62 0 657 489 81 8 0 749/750 Київ Івано-Франківськ 111 0 79 0 712 489 87 8 0 757/758- Київ Мог.Под.Кам.Под. 8 171 6 52 478 418 69 1 13 769/770 76/75 Кривий Ріг Київ 80 114 57 35 491 501 59 9 8 77/78 Ковель Москва 10 80 7 24 553 603 55 1 5 771/772 Хмельницький Київ 3 78 2 24 366 317 69 1 7 775/776, Харків Херсон 34 262 24 79 1127 1040 65 1 6 766/765 779/780 Суми Київ 6 100 4 30 352 465 45 1 9 782/781/ Черкаси Шостка 6 61 4 18 550 459 72 1 5 774/773 785/786 Київ Шостка 8 71 6 21 303 238 76 1 7 5 Ukraine Railways Policy Note 2 June 24, 2019 Pass-km Revenue Dist- Time Speed Cars From (mill) To (UAH mill) ance (mins) (km/hr) K(1) PK(1) K(1) PK(1) (km) K KP 789/790 Київ Кропиквницький 8 47 6 14 362 303 72 1 3 791/792 Кременчук Київ 7 49 5 15 401 302 80 1 3 797/798 Харків Запоріжжя 0 60 0 18 327 267 73 0 6 798/797 Одеса Вінниця 8 91 5 27 434 495 53 1 8 Average 95 116 68 35 784 810 58 7 5 (1) K = Kupe. For the purposes of this analysis they include sleeping cars, as the average yield per car-km is similar ; PK = Platzkart (2) Trains with numbers above 700 carry first and second class rather than Kupe and Platzkart. An average train in the sample of 94 consisted of seven Kupe carriages and five Platzcart carriages, although there are a significant number which are effectively all one or the other. The average distance is 784 km, taking over 13 hours. The average commercial (end-to-end) speed is relatively low at 58 km/hr, although this is often not an important consideration for passengers travelling overnight. In addition, many services, especially on low-frequency routes, make many intermediate stops. Figure 1.1 shows the sections of the network covered by these services. Figure 1.1 Network coverage of sampled services TRAFFIC DENSITY 2019 TT - MAIN LINE PASSENGER CARS (EXCL INTERCITY) KOVEL KYIV KHARKIV LVIV KOZIATYN POLTAVA VINNYTSYA KREMENCHUK IVANO-FRANKIVSK CHOP DNIPRO KRYVYY RIH ZAPORIZHIZHIA CARS/WEEK (EACH DIRECTION) 250 50 MARIUPOL 200 25 MYKOLAIV BERDIANSK 150 PLATZKART ODESSA 100 KUPE ISMAIL Because of the electronic ticketing system installed by UZ, demand (passenger numbers) and revenue can be directly attributed to individual services. However, as is the case in almost all large railways, it is impractical to record more than a few cost items directly to each individual service in the general ledger (financial accounting) system. These costs have therefore to be estimated by management accounting methods as described in the following two chapters. 6 Ukraine Railways Policy Note 2 June 24, 2019 2 Passenger Service Costing Framework 2.1 Introduction Few railways automatically record the cost of individual train services through their general ledger system. However, most railways record the cost, to a greater or lesser degree, of each of the main technical functions that create services, and then use broadly similar approaches to deriving the cost of the passenger sector as a whole and of individual services. This process consists of estimating a set of unit costs (i.e. cost per unit of output) for a set of outputs associated with passenger services (e.g. locomotive-km, car-kilometre and so on) and then applying them to the resources consumed in creating a particular service or set of services. This chapter discusses the cost structure generally employed and the factors (or ‘cost drivers’) which are the principal influences on the various cost components. 2.2 Passenger Service Cost Structure Table 2.1 shows a functional classification of passenger service costs that is in general use in railways worldwide. The majority of these cost types can be categorised in one of two ways: • whether they are directly variable with the volume of train operations or are fixed (at least until changes are made overall organisational scale or capacity); • whether they are associated with train operations (sometimes called ‘above-rail’ costs) or with infrastructure maintenance and operations (sometimes called ‘below-rail’ costs). Each of the costs includes an administrative component (depot managers, crew supervisors etc). There is also generally a central administration function covering such areas as finance, personnel, legal and corporate management. Finally, many railways have capital charges associated with the initial construction of facilities which are unlikely to be renewed in the foreseeable future; these costs have been included for completeness but are independent of either current or future traffic volumes. Table 2.1 also shows the most common basis adopted for the allocation of costs to particular segments or services. ‘Above-rail’ costs and variable ‘below-rail’ costs can generally either be specifically identified or else allocated using standard operating statistics, such as locomotive-km or gross tonne- km. However, where the cost is ‘fixed’ no such relationships exist and instead costs are allocated using a variety of principles. In recent years, the development of rail track access charges in railways offering third-party access, has seen these principles formalised, although with wide variations of systems and detail. In most railways, the ‘below-rail’ costs have been allocated to passenger services using the bases given in Table 2.1; however, for the fixed costs, these are essentially an allocative mechanism rather than a causal relationship. 7 Ukraine Railways Policy Note 2 June 24, 2019 Table 2.1 Functional Cost Categories and Basis of Allocation to Services Cost category Basis of cost allocation Variable costs Above-rail: Fuel and energy Gross tonne-km Train crew (driver and assistant) Train-hours On-train crew (conductors, attendants) Crew-hours (or weighted train-hours) Passenger handling and station operations Passenger Catering Passenger-km Rollingstock maintenance and servicing Vehicle-km (with vehicle-hour component) Rollingstock renewal capital Vehicle-hour Below-rail: Traffic-related track maintenance Gross tonne-km Train planning Train-km Infrastructure renewal capital (traffic- Gross tonne-km related) Fixed costs Below-rail: Traffic-unrelated track maintenance Gross tonne-km Structures maintenance Gross tonne-km Overhead line equipment maintenance Gross tonne-km (electric-hauled) Signals and communications maintenance Train-km Signalling operations, train despatching and Train-km control Infrastructure renewal capital (time-related) Gross tonne-km Administrative overheads Percent mark-up Non-renewable infrastructure capital Omitted The functional areas included under the heading ‘variable’ are largely the ‘above rail’ operating costs. Although in the very short term most of these costs (other than fuel) are fixed, all these costs can be varied in the medium term2. However, the speed with which individual cost areas respond to changes in the level of activity naturally depends on management’s ability to modify resource supply, which 2 Short-, medium- and long- term and their cousins short-run and long-run can be elusive concepts which mean different things to different people. In this paper, the long-run/long-term means that period of time during which both operations and capacity have adjusted to changes in demand, be it an increase or decrease. A long-run cost thus includes the capital cost or saving from the marginal change in capacity. Short-term, on the other hand, means the period of time before any such adjustments have been made; during this period, cost savings will generally only be those costs which are consumable (such as fuel) or the subject of short-term contracts. In railways, increases in infrastructure are often slow to eventuate and it is convenient to consider a third concept, the medium-term, during which adjustments have been made to operational capacity (crew, maintenance, and rollingstock capacity) but changes in the capacity of the infrastructure have not yet taken place. From a practical point of view, the short term for a railway is probably up to 18 months, while the long-term is rarely less than five years. 8 Ukraine Railways Policy Note 2 June 24, 2019 depends as much on institutional arrangements as on technical factors. So the cost of train crews, who can be redeployed to alternative services, generally responds faster than rollingstock maintenance costs, which have a substantial workshop component for overhauls which are only carried out at intervals of four years or more. A large part of the costs of routine infrastructure maintenance are independent of the level of traffic and can be considered fixed, until traffic levels change dramatically, or additions or subtractions are made to the network. Relative scale is always an important factor when considering the ‘fixity’ of any cost as the addition of a new traffic of sufficient volume requires additional infrastructure with its associated costs. Similarly, removing all traffic on a particular line, even if it is a comparatively small volume, would make all costs on that line avoidable if it were to facilitate closure. The third group of costs are administration-related3. They are fixed in the sense that they are largely insensitive to the removal or addition of an individual train service but are more related to the overall scale of the railway. Even then, they generally react only sluggishly. A typical pattern is for these costs to remain more or less fixed (or slowly but steadily increasing) for several years punctuated by periodic management ‘restructurings’ during which a chunk of the overhead is taken out. As long as operations continue at around an existing general level of activity, central administrative costs can generally be treated as independent of traffic volume, although there would be the ability to make a step reduction if passenger services were more dramatically curtailed or withdrawn. The treatment of capital costs also depends on the decisions to be made. The cost of operating an on- going service should include an allowance for renewable capital items such as rollingstock and track excluding only the non-renewable component of infrastructure capital costs4. However, an accounting income statement would include depreciation and interest charges for all assets, usually based on historical costs (and often of doubtful use). In general, the avoidable capital cost is based on either the replacement cost of the asset, for an on-going service, or is zero where the service is unlikely to be replaced when the rollingstock becomes obsolete. 2.3 Cost Components The raw financial data available in most railways with conventional financial accounting systems is broadly consistent with the categorisation of Table 2.. In some cases, labour on-costs, such as pensions, are included as general administration and need to be extracted and, in the absence of specific causal information, allocated to functional areas in proportion to wages and salaries. 5 Functional overheads can then be distributed to the relevant functional area. For example, administrative costs relating to track maintenance (supervisors, district engineers etc) are included in track related direct expenditure and are then allocated on the same basis as the direct expenditure 3 These administration costs do not include all costs incurred at Divisions and HQ. The administrative costs of branches such as Engineering and Electrical are generally treated as overheads to the relevant direct costs, leaving only the accounting and general administration branches such as General Management and Finance. 4 This assumes that the rollingstock can be redeployed elsewhere and that the renewable component of the infrastructure can be salvaged and re-used. The costs of the initial construction of the line, such as earthworks and some structures and buildings, are unlikely to be recovered and the associated capital charges generally continue to be incurred indefinitely until they are written off. 5 Many modern accounting systems allocate on-costs directly. 9 Ukraine Railways Policy Note 2 June 24, 2019 they are supporting. However, general administration overheads, such as general and financial management, are not directly related to a particular function and remain separate. The augmented functional costs (including the labour on-costs and supervision costs) for those components in Table 2.1 which are ‘variable’ can then be allocated to services based on the appropriate causal relationship. 2.3.1 Operating costs Fuel and Traction Electricity Fuel and energy costs are normally booked to system-wide accounts but may sometimes then be allocated to passenger and freight based on estimated consumption rates. Very few railways have on- train meters. The simplest basis to distribute fuel and traction electricity costs to services is in proportion to gross tonne-kilometres; however, more accurate estimates are possible for individual routes if the average grade of the route, the speed of the train and the train size are taken into account. Train Crews On some railways, drivers are booked under separate accounts for freight and passenger work, and sometimes also for multiple-units and locomotives; on other railways all train crew is booked to a common account. Allocation to services is best done in proportion to train-hours; on railways where this statistic is not available, train-km is a second-best alternative. Adjustments need to be made for the number of train-crew used – some railways use single-manning for some types of locomotive and multiple-unit trains. Guards are no longer used on all railways but where this is so, they are normally allocated on the basis of train-hours. On-train crew The cost of on-train crew varies sharply from railway to railway as well as by type of train. At one extreme are sleeper cars with two attendants per carriage on some railways; at the other is a single conductor/inspector – or sometimes none. Some railways also have on-board security staff (or railway police), typically two or three per train. Where no detailed data is available, the only practical basis for allocation is to use weighted train-hours, with the train-hours weighted by the estimated number of on-board crew for each train type. Passenger handling and station operations These are often difficult costs to identify for the typical integrated railway. The major costs are labour- related and therefore can only be identified by analysing (or using a management assessment) of the time spent by staff on passenger-related activities. Staff employed at stations for train despatching and train-working should be included under signalling operations but are often included in general station costs. There are variations in ticket selling and terminal costs depending on the fare/ticketing systems and facilities (some railways have very basic passenger facilities or provide a low level of service for ticket- selling, with long queues) as well as management practices. Some railways have out-sourced ticket sales on a commission basis (and others have set up special internal cost centres which effectively ‘in- source’ this activity). 10 Ukraine Railways Policy Note 2 June 24, 2019 Station maintenance costs are generally only available in aggregate but can be allocated in proportion to staff numbers. Although the connection between station maintenance and staff numbers is more tenuous than for operations costs, the number of staff does give an indication of size and complexity of stations and therefore the level of maintenance effort. The aggregate passenger-related costs are best allocated to services on a per passenger basis. In practice, the high-fare and long-distance passengers have higher costs than commuters and low-fare passengers, so an acceptable alternative is to distribute costs in proportion to revenue. Catering Catering costs should be directly available from the financial accounts. However, such costs generally only include the cost of food and catering staff; they exclude the costs associated with maintaining and hauling any restaurant and buffet cars. Catering revenue and expenditure should be allocated in the same manner; depending on the system, weighted passenger-kilometres is a reasonable basis, with weights reflecting the relative use of these facilities by the different classes of passengers. Rollingstock Maintenance Maintenance cost of passenger rollingstock is generally available by broad type (locomotive, carriages and multiple-units) but is rarely available by class of carriage and not always by type of locomotive. There are arguments for and against differentiating between classes of locomotive, as the locomotives used on any particular service often reflect management discretion rather than being based on specific technical characteristics; in addition, very often the most powerful locomotives are the newest and hence have the lowest unit maintenance costs. It is therefore often best to use a fleet average maintenance cost for locomotives; analysis has shown this can be allocated on the basis of locomotive- km and locomotive-hours, typically 50:50. Similar results hold true for multiple units, although here it is often worth distinguishing between modern 3 and 4-car sets used for main-line services and the life-expired one or two-car units often used for branch-line services. Carriage maintenance costs should also distinguish between the higher- class carriages, often with air-conditioning and generally with much higher interior standards, and the lower-class carriages, for which interior maintenance is minimal on many railways. A reasonable basis for allocating carriage maintenance costs between services is 70 percent on the basis of vehicle-km and 30 percent on the basis of vehicle-hours; however, if costs by class of carriage are available, vehicle-km alone can be a reasonable basis, as most trains with similar types of carriage operate at similar commercial speeds Cost is also a function of the level of service being provided. For example, vehicles for main-line premium services will generally have a higher standard of interior fittings and maintenance than those used on the cheapest branch-line services; cheaper unit costs in such circumstances therefore do not necessarily reflect greater efficiency. Infrastructure costs Even when a common methodology is adopted, infrastructure costs often vary widely, reflecting not only variations in accounting procedures for periodic maintenance, as discussed above, but also the very different infrastructure standards on the different railways. The absolute cost of infrastructure maintenance increases as volume increases not only because of the costs associated with the incremental volume but also because of the higher standards required to provide the required 11 Ukraine Railways Policy Note 2 June 24, 2019 capacity and because higher traffic volumes restrict the time available each day for access to the track, thus increasing the unit cost of maintenance activities. Treatment of major periodic maintenance, both for rollingstock and infrastructure, can also cause distortions. More financially robust railways carry out such work routinely and include the costs as part of operating expenses. However, many smaller railways with low traffic densities can often only do such work when they receive specific funds through loans and grants. In these cases, the work is almost always capitalised; periodic maintenance can represent 50 percent or more of the long-run maintenance cost and its capitalisation creates significantly deflated unit operating costs. The converse arises when rollingstock items, particularly locomotives, continue to be operated much longer than their normal lives as the railway does not want to (or cannot) purchase replacements; the life-expired locomotives can require extensive routine maintenance to keep them operating, leading to comparatively high unit operating costs. Train despatching and control Train control (signalling) and train-working costs are often very difficult to establish with any precision because station staff involved in typical railways are generally multi-tasking; these costs are also heavily influenced by policy decisions on station manning. Headquarters Administration Analysis of these costs on different railways has shown that they are within a fairly close range (5 percent-12 percent) of total working expenditure (i.e. excluding capital charges). In the long run, therefore, they should vary in line with other expenditure and can be included in service costs as a percentage mark-up. 2.3.2 Capital costs Rollingstock The capital cost of rollingstock and infrastructure is reasonably consistent across all railways. Rollingstock manufactured in Europe is generally more expensive on a vehicle-for-vehicle basis but generally has more features and ancillary equipment e.g. air conditioning, more sophisticated toilets, centrally operated doors etc. Given the unit costs and life of an asset, this can be converted into equivalent annual costs, using an assumed real interest rate. The capital cost per unit of use then depends on the asset utilisation; higher utilisation, whether from the type of service being offered (e.g. long-distance overnight services or daytime medium-distance shuttle services) or from the efficiency of timetabling and operation, generates lower per unit capital costs. The appropriate approach to the measurement and allocation of capital costs depends upon the purpose for which it is being estimated. For the purposes of financial accounting statements, assets are generally valued using accounting standards which assume that the railway is a going concern, values are related to historic costs and normal commercial relationships between risk and return prevail. From an operating railway’s perspective, many assets are impossible to value in the accepted commercial sense. Land for the permanent way may have a significant alternative use value but if it is unavailable for that alternative use, its value is a function of the level of income that it can generate. This Policy Note assumes the on-going operation of the railway so the only capital costs included are replacement costs for renewable items. 12 Ukraine Railways Policy Note 2 June 24, 2019 The capital costs are represented as an annual charge equivalent to the depreciation and interest costs that would be incurred if the replaceable assets were replaced at current new costs with a loan over the life of the asset. The interest rate used (i.e. the return to the investor) is set at 7 percent real (i.e. the return excluding the effects of inflation) and asset lives are 25 years (100,000 hours) for diesel locomotives, 30 years for diesel multiple-units and 40 years for electric locomotives, electric multiple units and hauled carriages. Infrastructure In most FSU countries, the periodic replacement of infrastructure is known as ‘kapital remont’ and is treated as an operating cost. No additional cost therefore needs to be included for this activity. 13 Ukraine Railways Policy Note 2 June 24, 2019 3 UZ Passenger Service Costs 3.1 2017 Financial Accounts UZ is a vast organization with many wholly and partly-owned subsidiaries. Some of these do significant work for organizations outside UZ but many work only for the railway departments of UZ, for which they receive income. The best starting point for an analysis of the costs of UZ services is the segment analysis provided in Section 8 of the audited Annual Financial Statements (Table 3.1). This gives the cost in 2017 6 of each three main market segments: mainline passenger, suburban passenger and cargo. Table 3.1 Segment results 2017 (UAH million)(1) Long- Suburban Auxiliary All other Cargo Unallocated Eliminations Adjustments Total distance passenger activity segments Sales to third parties 60,991 6,745 589 6,498 9 – (894) – 73,938 Inter-segment sales – – – 7,409 1,777 – (9,186) – – Total revenue 60,991 6,745 589 13908 1,785 – (10,080) – 73,938 Staff costs (17,409) (5,318) (2,585) (4,636) (617) (607) 585 (1,388) (31,974) Depreciation (10,714) (1,633) (870) (1,274) (36) (168) 34 – (14,662) Electricity (5,306) (1,236) (722) (7,952) (56) - 7,428 (3) (7,866) Fuel (4,971) (955) (605) (769) (47) (5) 44 – (7,308) Maintenance (3,758) (837) (528) (367) (1,106) - 1,447 (661) (5,810) Other expenses (4,756) (1,423) (556) (1,363) (11) (1,717) 7,572 4,827 (1,061) Inter-segment – – – 3,398 – – (3,398) – – Expenses Total cost (46,914) (11,420) (5,866) (12,963) (1,874) (2,499) 10,080 2,775 (68,681) Segment result 14,077 (4,675) (5,277) 945 (89) (2,499) – 2,775 5,257 (1) As given in 2018 financial statements A good explanation of what the various headings mean is given in the notes to the 2018 accounts and is reproduced below. For management purposes, the Group is organized into business units based on their services, and has five reportable operating segments: • Cargo segment includes cargo transport services provided by the Group. • Long-distance passenger transport segment comprises cross-regional passenger transport services provided by the Group. • Suburban passenger transport segment includes intraregional rail passenger transport services. • Auxiliary operations segment mainly includes activities of certain structural units of Regional branches and branches of the Company that produce industrial products and provide services for internal consumption within the Company and for sale to third parties outside of the Group. Auxiliary operations segment produces industrial products and provides services related to cargo transport and suburban passenger transport, construction, reconstruction and modernisation of railway tracks and railway infrastructure, repair and maintenance of various railway-related equipment etc. The transfer of products (works, services) between the structural units of one branch and between branches within the Company is considered as internal turnover and is reflected as a reallocation of expenses (“Inter-segment expenses” line) with no revenues recognised (except for certain types of transactions). A segment that receives reallocated expenses includes them in the "Other Costs" category. 6 Much of the analysis reported in this paper was undertaken before the 2018 statements and supporting documentation had been finalised. Updating will be completed when this data becomes available. 14 Ukraine Railways Policy Note 2 June 24, 2019 None of Auxiliary segment operations are individually of sufficient size to be reported as separate segment. None of these operations can be aggregated with reportable operating segments described above due to dissimilar economic characteristics. • All other segments include repair and maintenance of rolling stock and other services provided by the Company’s subsidiaries. The operating segments results do not include the effects of some adjustments that may be considered necessary to reconcile the management accounts to IFRS consolidated financial statements. Segment revenue is revenue that is directly attributable to a segment, whether from sales to external customers or from transactions with other segments. Segment revenue does not include: • Interest income; • Foreign exchange gains; • Gain on disposal, change in fair value and reversal of impairment of financial assets; • Gain on disposal of property, plant and equipment; • Other income. Segment expenses are expenses resulting from the operating activities of a segment that is directly attributable to the segment and the relevant portion of an expense that can be allocated on a reasonable basis to the segment, including expenses relating to sales to external customers and expenses relating to transactions with other segments. The segment result is calculated as the difference between segment revenue and segment expenses. Transfer prices between operating segments are on an arm’s length basis in a manner similar to transactions with third parties, except for operations of electricity transfer. The costs presented in Table 3.1 include both train operating costs (the ‘above-rail’ costs) and the costs of maintaining and operating (train control and related activities) the infrastructure (the ‘below- rail’ costs. As the below-rail costs include a significant element of joint costs, this component of the segment costs includes allocation on a general basis rather than a causal one, as discussed in section 2. 3.2 Form 10 It was impractical for the many rail-related organisations making up the rail sector in the FSU to have a single chart of accounts, given the vast geographical scope and the wide range of activities. Instead, a uniform set of accounts (known as Form 69) was developed into which all organisations converted their data to create a consistent basis for subsequent management accounting. It was used extensively for tariff-setting and for allocating expenditure between passenger and freight traffic. Many FSU railways, including UZ, still use this form (now known as Form 10 in Ukraine), although it has been frequently revised to deal with changing operational circumstances. The Ukraine version now has around 900 accounts, arranged into 13 main groups broadly corresponding to the main functional groups (passenger operations, freight operations, traction, civil engineering etc) for each of which expenditure is divided by type: labour, staff benefits, electricity, diesel, materials, depreciation and other (Table 3.2). ‘Other’ includes what in some charts of accounts is termed ’sundries’ e.g. external services, travel, stationery etc Each operating unit completes its own Form 10 and these are consolidated centrally into one covering the whole of UZ. In many cases units are providing services for other units and in these cases the cost is cross-charged. These are shown in the last two columns of Table 3.2 and included as debits (and corresponding credits) in the ‘other’ costs. The individual costs in Form 10, in aggregate, exactly match the segment costs in the audited financial statement shown in Figure 3.1. Whilst there may be, as in all large organisations, occasional miscodings of costs, detailed analysis of Form 10 by the World Bank 15 Ukraine Railways Policy Note 2 June 24, 2019 did not uncover any significant instances where costs seemed unreasonable and there is no reason to question the reliability of these costs. Within Form 10, each cost is allocated to one of three segments, cargo, long-distance passenger and suburban passenger, corresponding to the three segments in Table 3.1. In most cases, other than infrastructure, the allocation is straightforward because of the structure of the individual accounts, which have developed over time to record the expenditure to these segments at as detailed a level in the organisation as possible. Where costs do have to be allocated to more than one segment,this is done on the basis of a suitable physical measure, similar to the procedures described in the previous chapter. Table 3.2 Typical Form 10 cost records (UAH 000) Including Total Transferred Article Activity Material costs Deprec- Payroll Benefits Other From To Materials Diesel Energy iation Sale of domestic 1001 7703 135,722 30,061 7,952 1 1,008 17 9,620 184,381 1 573 tickets 1003 Domestic baggage 7705 7,240 1,363 162 11 3 465 9,244 49 44 On-board 1008 equipment and 7708 57,014 12,749 6,895 29,316 2,932 4,139 13,274 126,319 1,598 2,303 services Maintenance of 1009 domestic and 7709 1,063,682 242,922 13,852 9,693 56,075 9,485 102,687 1,498,396 45,929 21,877 international cars In any unit, there are a number of cost heads that are of a support nature and cannot be directly related to a specific activity undertaken by that unit. Examples include management costs, employee sick leave and training or the maintenance of buildings in a workshop. These costs (known, for example, as ‘shop overheads’ in workshop costing) are included as a separate item which can then be subsequently distributed over the relevant direct costs. Table 3.3 summarises the costs for the three market segments by the 15 main Form 10 cost groups. Table 3.3 Segment costs by department 2017 (UAH million) Departments Long- Suburban Cargo Other Total distance passenger Mainline passenger 5547 176 5723 Suburban passenger 172 3472 3645 Cargo 2173 2173 Operations 2579 375 281 3234 Locomotive 14906 3231 263 94 18493 Wagon 5460 10 3 5474 Track 14827 746 480 16054 Buildings and structures 805 259 382 43 1490 Signalling and communications 1744 750 541 3036 Electrical 1168 174 205 1547 Supply 159 16 4 180 Recovery trains 1086 81 58 1225 Management 74 74 IT/security 728 31 21 780 16 Ukraine Railways Policy Note 2 June 24, 2019 Central management 1071 1071 Total cost 45635 11393 5886 1285 64200 The first six departments broadly constitute the above-rail costs (UAH 38.7 billion in 2017, about half the total costs). The next four (UAH 22.1 billion) broadly constitute the below-rail costs while the last five (UAH 3.3 billion) are various services which are provided on a system-wide basis. The Passenger and Suburban Passenger departments maintain and crew the self-propelled trains they operate. The costs incurred for passenger services from the Locomotive Department are thus only those for loco-hauled services (the majority for main line passenger services but very few for suburban passenger services). Wagon costs are almost exclusively for freight. Locomotive-hauled passenger vehicles are maintained by the Passenger department itself. As the track costs in Form 10 are currently distributed on the basis of gross tonne-km, the majority are allocated to freight, while suburban passenger attracts comparatively little. 3.3 Analysis of Form 10 costs to generate aggregate costs by activity For the purposes of passenger train costing, each of the Form 10 accounts was reclassified into one of 13 activities corresponding to the cost structure given in Table 2.1 and shown in Figure 3.1: Table 3.4 Segment costs by activity 2017 (UAH million) Mainline passenger Suburban passenger Diesel fuel 762 476 Traction energy 957 589 Loco maintenance 524 18 Car maintenance 1912 91 EMU maintenance 43 689 DMU/RB maintenance 34 326 Shunting locos maintenance 38 8 Station 975 522 Train crew 840 1029 On-train crew 2226 - Shunting crew 105 Infrastructure 1629 1323 General admin 110 78 Depreciation 1238 728 Total 11393 5885 17 Ukraine Railways Policy Note 2 June 24, 2019 Figure 3.1 Mainline passenger costs as summarised in Form 10 2017 3.4 Derivation of unit costs 3.4.1 Unit operating costs The aggregate costs in Table 3.4 can be related to the relevant cost drivers to derive a set of unit costs (table 3.5). Most railways keep records of the operating statistics which are the key cost drivers (train- km, locomotive-km and so on) by broad category, in the case of UZ for cargo, mainline passenger and suburban passenger. Table 3.5 Derivation of unit operating costs for mainline passenger services (UAH 2017) Cost (UAH Cost driver Activity 2017 Unit cost mill) Diesel fuel 762 Diesel gtkm (000) 5764 132 Traction energy 957 Electric gtkm (000) 33361 29 Loco maintenance 524 Loco-km 54469 10 Car maintenance 1912 Car-km 625532 3 EMU maintenance 43 EMU km 22698 2 DMU/RB maintenance 34 DMU km 5829 6 Station/shunting 1118 Passenger 49000000 22 Train crew 840 Train hour 967000 868 On-train crew 2226 Car-hour 11362000 197 Few railways maintain detailed records of shunting activities and these costs (fortunately relatively small) have been combined with station costs to provide a composite passenger handling cost. 18 Ukraine Railways Policy Note 2 June 24, 2019 UZ do not keep separate maintenance costs for electric and diesel locomotives, nor for the different types of passenger carriage, and in particular not for kupe and platzkart. The maintenance cost of electric locomotives is typically about two-thirds that of corresponding diesel locomotives and this was used to develop costs for each type consistent with the overall average. The same approach was adopted for the platzkart and kupe carriages; kupe carriages are maintained to a higher standard and were given a maintenance cost 40 per cent higher than the platzkart cars after discussion with UZ staff. 3.4.2 Unit capital costs The capital cost of rollingstock and infrastructure is reasonably consistent across all railways, although it naturally depends on the features provided e.g. air conditioning, more sophisticated toilets, centrally-operated doors etc. Capital cost can be converted into equivalent annual costs, using an assumed asset life and real interest rate. The capital cost per unit of use then depends on the asset utilisation (Table 3.6). Higher utilisation, whether from the type of service being offered (e.g. long- distance overnight services or daytime medium-distance shuttle services) or from the efficiency of timetabling and operation, generates lower per unit capital costs. Table 3.6 Derivation of unit capital costs for mainline passenger services (UAH 2017) Cost Life Depr. ROI/Int Total Util. P.a.. Cost per unit ($) USD ($000) YRS $000 $000 $000 Depr ROI Locos (1 section) 2500 25 100 115 215 4000 25.00 28.75 Hour Carriage (K) 1000 40 25 50 75 4000 6.25 12.50 Hour Carriage (PK) 800 40 20 40 60 4000 5.00 10.00 Hour 3.5 Application of unit costs to selected services It would be possible to extract these for individual services from the operational database but UZ does not yet do this. Instead, the relevant statistics for individual services are normally calculated on a service-by-service basis, using standard operating information on train times and compositions. Table 3.5 shows a sample of the information used in the service costing in this report. Table 3.5 Service operating data Distance Days/ Time Carriages From To Electric Diesel week (mins) Kupe Platzkart 1/2 Ів Франківськ Костянт 1312 141 3.5 1230 10 0 3/4 Запоріжжя Ужгород 1440 0 3.5 1240 10 0 6/5 Київ Москва 354 0 7 292 12 5 8/7 Одеса Харків 762 0 3.5 555 10 0 7/8 Київ Івано-Франк 572 140 7 681 8 0 7/8 Харків Одеса 762 0 3.5 555 10 0 12/11 Львів Одеса 745 0 7 608 10 0 12/11* Київ Новоолексіївка 861 0 7 737 15 3 13/14 Київ Солотвино 862 155 7 1205 9 6 This data enables all the cost drivers to be defined, given the weights of the locomotive and rollingstock. The cost of each service can then be derived by combining them with the relevant unit costs. 19 Ukraine Railways Policy Note 2 June 24, 2019 20 Ukraine Railways Policy Note 2 June 24, 2019 4 Analysis of results 4.1 Introduction There are two important financial thresholds that help establish the financial viability of rail passenger services: • Is the revenue earned by the service greater than the above-rail operating costs (excluding capital)? If so, it is worth continuing to operate the service as it has a positive cash flow, as long as there is not a better use for the rollingstock used in the service • Is the revenue earned by the service greater than the above-rail operating costs (including capital)? If so, it is worth reinvesting in the rollingstock for that service when it becomes life- expired. Both thresholds assume that there are other users of the infrastructure over which the service operates (a reasonable assumption in the case of mainline passenger services in UZ). If not, then either the service needs to generate sufficient financial surplus to cover the cost of the infrastructure or there needs to be external financial support to achieve this. The next section presents the results of the analysis for 2017. 4.2 Overall Service Viability Figure 4.1 addresses the first threshold above and shows the distribution of the above-rail operating cost ratio as a function of the proportion of kupe carriages. In both Figure 4.1 and 4.2, the size of the bubbles indicates the passenger-km on the service. Figure 4.1 Above-rail operating cost ratio and proportion of kupe carriages in a train 21 Ukraine Railways Policy Note 2 June 24, 2019 Of the 94 services analysed, 57 had a ratio greater than 1 and were thus earning more on each trip than it was costing to operate them. Almost all services for which kupe cars were over 70 per cent of the train composition were over 100 percent, but none of those with under 20 percent kupe cars were. When above-rail capital costs (locomotives and carriages) are included, only 20 out of the 94 services had a ratio greater than 1 and were thus earning enough to fund replacement rollingstock as it became due (Figure 4.2). Only services with 60 percent or more of kupe carriages are passing this test and even then it is by no means guaranteed. Figure 4.2 Above-rail operating + capital cost ratio and proportion of kupe carriages in a train An even smaller number of the sample would be able to contribute significantly to the cost of infrastructure. If the track access charge for mainline passenger were set at UAH 30 per 000 gross tonne-km, only 9 out of the 94 services analysed would be able to pay. Given the revenue per passenger-km for a kupe passenger is more than twice what it is from a platzkart passenger, the mix of classes is clearly a critical element in determining the financial viability of a service. This is discussed in more detail in the next section. 4.3 Financial analysis by class Figure 4.2 shows that almost all all-kupe train services are able to cover their reinvestment costs, while no all-platzkart train services are. Figures 4.3 and 4.4 show the cost per vehicle-km and per passenger- km. 22 Ukraine Railways Policy Note 2 June 24, 2019 Figure 4.3 Cost per vehicle-km of platzkart and kupe services Figure 4.4 Cost per passenger-km of platzkart and kupe services 23 Ukraine Railways Policy Note 2 June 24, 2019 The cost of platzkart per vehicle-km is only slightly less than for kupe, because of the lower maintenance cost. Per passenger-km, however, the difference is more pronounced because of the higher occupancy rate for platzkart. Even so, however, this is not sufficient to generate comparable earnings per vehicle-km (Table 4.1) and platzkart will normally always fail to cover its full costs because of the lower regulated fares (Figure 4.5). Table 4.1 Average revenue per vehicle-km Kupe Platzkart Average load (pkm/vkm) 26 41 Yield (UAH/pkm) 0.71 0.30 Revenue/pkm 19 12 Figure 4.5 Comparison of kupe and Platzkart revenue and operating cost7 4.4 Impact of variations in train composition, speed and traction type With UZ’s current cost structure, the cost of passenger trains is dominated by the carriage-related costs (on-board attendants, maintenance and capital) rather than those associated with the locomotive. Some of these are variable with time while some primarily vary with distance. 7 FDC is fully distributed costs. 24 Ukraine Railways Policy Note 2 June 24, 2019 Figure 4.6 Train and vehicle-related costs About 75 percent of the total costs are associated with carriages, with costs also dividing approximately equally between passenger and freight. This indicates that, while improving travel times should have a noticeable effect on overall cost, varying the size of the train will have comparatively little effect if average occupancies remain unchanged. Figure 4.7 shows the variation of cost with speed for a 12-car Platzkart train. The blue columns show the distribution of the analysis sample by speed in terms of car-km per week. Doubling the commercial speed reduces the cost per vehicle-km by about 40 percent. 25 Ukraine Railways Policy Note 2 June 24, 2019 Figure 4.7 Cost variation by speed The variation in cost with the size of the train is much less marked (Figure 4.8). Doubling the size of the train reduces the cost per vehicle-km by only about 10 percent. Figure 4.8 Cost variation by size of train Although train-related costs (which includes the locomotive) are relatively small, electric-hauled trains are cheaper than diesel-hauled ones by about 15 percent. 26 Ukraine Railways Policy Note 2 June 24, 2019 5 Possible strategies While the Kupe services seem relatively secure financially, the Platzkart services are only surviving because of (effectively) cross-subsidy from the Kupe and freight services. However, this will become increasingly difficult and an alternative approach needs to be found. The normal response to improving the financial recovery of these services is simply to increase the fares. However, Platzkart fares in Ukraine are heavily regulated by the government for social reasons and that policy seems unlikely to change in the short medium term. The other main options available to UZ to improve the financial performance of the Platzkart service form three groups: • Reduction in the unit costs of the inputs to the services • More efficient use of the inputs • Compensation from the government Whilst unit costs can almost always be reduced by reviewing methods and procedures, in many cases such reductions are not large and reviews sometimes show that costs should be increased rather than decreased in order to ensure maintenance is properly carried out. There is little evidence that the unit costs in UZ are unduly high compared to those in other countries, based on experience of railways in Eastern Europe that the World Bank has worked on in recent years. For example, locomotive maintenance in UZ averages UAH 10 per loco-km – (about US$ 0.40). Typical values elsewhere range from $0.7-$2 per loco-km, depending on the age of the locomotive and whether it is electric or diesel. Similarly, the maintenance cost of loco-hauled carriages, estimated at UAH 3/vehicle-km (US 0.10 ), is similar to a typical maintenance cost for a freight wagon on other railways. Reductions in unit cost will therefore help at the margin but will not be the long-term solution to losses. UZ is already trialing a change in their inputs to these services. The traditional manning for the on- board attendants has traditionally been two people per carriage. A single train might therefore carry upwards of thirty on-board staff. UZ are now testing one person for two carriages, albeit with a 30 percent pay increase. This move represents a significant reduction in one of the largest cost components, reducing on-train crew costs by around two-thirds. 27 Ukraine Railways Policy Note 2 June 24, 2019 Figure 5.1 Above-rail operating + capital cost ratio – modified on-train crew If this trial were extended to all services it would increase the number of services in the analysed sample covering their above-rail operating and capital costs from 20 to 36 and bring many others into an achievable distance of doing so. Another option for UZ is to more closely tailor capacity to demand, especially on some of their very long-distance services. Although the average occupancy (measured as the dynamic load factor: passenger-km/seat-km) for the analysed services is 73 percent, this varies from 100 percent to 40 percent or less. Whilst load factors on services with many intermediate stops are much harder to manage than for services which are largely point-to-point(like airlines), this variability does suggest that capacity is being provided on some services to handle specific heavily-loaded sections which then goes unused for the rest of the journey. Addressing this problem is not easy, as modifying train consists en route can be expensive but there is a prima facie case for reviewing some of the service schedules and compositions. It seems unlikely that the demand for Platzkart services will reduce significantly in the short-medium term. The average journey length is relatively long (536 km) and long-distance bus services are relatively few in Ukraine (e.g. five services daily between Kiev and Lviv), take as long as the train over indifferent roads and are unpopular because of both safety and travelling conditions in winter. Nevertheless, any reorganization and rescheduling of services should be accompanied by market surveys of both rail and competing modes to ensure the resulting services remain attractive to potential passengers. UZ could also apply for financial compensation from the government for continuing to operate these services. As there is unlikely to be a significant problem with an all-Kupe service, this equates to compensation for operating Platzkart services at fares regulated to levels insufficient to cover costs. There are two main ways of doing this: 28 Ukraine Railways Policy Note 2 June 24, 2019 • UZ could keep the revenue from the Platzkart passengers and then be compensated for the difference between this revenue and the cost of operating the Platzkart vehicles • UZ could operates the Platzkart vehicles under a supply contract with the government. The Government would then receive all Platzkart revenue and would pay UZ a fixed amount per Platzkart vehicle-kilometer The first of these will create a large amount of work to cost individual services, many of which may change their composition on a regular basis (e.g. by day of week). In addition to the problem of data complexity, there are also conceptual problems in allocating part of the train-related costs (locomotive and driver) to the Platzkart vehicles. The second option seems far simpler to implement and administer, as well as providing greater certainty to both government and UZ. The only data required is the vehicle-kilometers performed by Platzkart carriages and this can be combined with a single unit cost per vehicle-kilometer applied to all trains to yield a compensation level.. In theory, such a cost would vary by train composition and train speed (see Figures 4.7 and 4.8) but adopting an average cost based on an average service should mean that any deviation from a service- by-service costing would be minimal. In summary, the immediate priority for UZ is to find an approach to PSOs which is acceptable to all parties and compliant with the railway Law and the Accession Agreement. Strategies to improve passenger service and operating efficiency need to be longer-term and, in most cases, will require continuing investment in rollingstock and passenger facilities. 29