Costing System for Cfr

DEVELOPMENT OF POINT-TO-POINT

COSTING SYSTEM FOR CFR

FINAL REPORT

Prepared for:

Compania Nationala de Cai Ferate “CFR”-SA

Victor AlaloufOctober 2002

Table of Contents

1.INTRODUCTION

1.1The Current Status

1.2Recommendations

1.3Response to CFR Comments on Draft Final Report

2.INFRASTRUCTURE COSTS

2.1Current Access Charge Calculations

2.2Access Charge Issues

2.2.1Track Maintenance

2.2.2Train and Shunting Control

2.2.3Signals

2.2.4Recapitulation

3.APPLICATIONS – MARFA

3.1.The Batch System

3.2.Economics of a Line

4.APPLICATIONS - CALATORI

5.DEVELOPMENT OF INPUTS

5.1 Accounting Data

5.2 Operating Statistics

5.3 The Unit Costs

5.4 The Network Data

5.5 The Train Information

5.6 Shunting Data

5.7 Car Utilization

1. INTRODUCTION

This report marks the completion of the tasks carried out by the consultant under the Terms of Reference of the contract entered into in February 2002 between the Consultant and Compania Nationala de Car Ferate – “CFR”-SA.

The following tasks, listed in of the Terms of Reference, have been completed:

a) Install the Cartage costing model on a railway computer;

b) Develop network tables for CFR;

c) Develop other model input;

d) Customize the model;

e) Test model on sample passenger traffic;

f) Test model on sample freight traffic

g) Training of local staff of CFR-SA, Calatori, Marfa and SMF

h) Assist in installing the costing model at users’ computers;

i) Design future improvements.

The project required two visits to Bucharest: they took place in February 25 to March 25, 2002, and May 13 to May 31, 2002. In the latter part of the first visit, a CFR counterpart was designated. The costing model was installed on a CFR-SA computer. Along with the model itself, the consultant installed input data to allow CFR staff to acquire some experience in running the model. The input initially installed consisted of the network data of another railway and fictitious cost and operating data for that railway. At the same time CFR staff was given a costing manual corresponding to the version of the model available at the time.

The network input was completed in the consultant’s home office shortly after the end of the first visit. The network defined for input to the model consists of 406 links, 337 nodes and 1518 stations. During the same period, significant progress was made in the development of other input data.

Concurrently, a number of minor modifications to the model were made, the most significant of which related to the input format. The earlier version of the program could read the fixed input only in Text format. The enhanced version can now read either Text or Excel files. Similarly, in batch mode, the model can now read MS Access files as well as Text.

Contact persons in SMF, CFR-SA, Marfa and Calatori were designated early in the project, and supplied a large amount of useful data. However, the person designated as counterpart near the end of the first visit was not released from his regular duties and had very little time available for the project; during much of the consultant’s second visit, he was on a course unrelated to the project.

1.1 The Current Status

The deliverables of the project are:

The software;
The Costing Manual
The Interim report
The fixed input tables;
The Final Report.

Although a working version of the model is installed on four CFR computers and fully operational, a new version incorporating a few suggested modifications will be transmitted electronically concurrently with this report. The Interim Report was submitted in March 2002. The final fixed input tables will follow the submission of this report.

The purpose and applications of the costing system, its structure and its main characteristics are presented in the Interim Report, and will not be discussed here. The main purpose of this paper is to describe the development of the input data.

The development of the input tables has proven difficult because the data sources are disparate and not particularly well suited to the development of a costing system. In addition, some large gaps had to be filled with estimates that undermine the quality of the final results. Nevertheless, we believe that:

The system as it stands is a vast improvement over anything available now, and,
It serves as a foundation for further improvements in the future. Once there is a structure in place, and a path has been drawn from raw data to the final model inputs, enhancements can be made incrementally

There is no doubt that the upcoming systems such as Iris and Oracle Financial will greatly facilitate the development of input data, and improve the end product. However, one must be somewhat suspicious of such grand schemes as a fully automated path from the recording of events and transactions in the field to the cost output in the Marketing Department. One reason is that such a system requires a degree of systems integration rarely achieved. A more compelling reason is that the input to the costing system is not merely a matter of recording facts. The input to a predictive model, which should be one of the aspects of the costing system, requires the intervention of analytical work and some judgment calls at various stages of its development. No IT system can provide such analyses.

As stated above, this paper deals primarily with the development of the input to the model. One relevant input relates to access charges. These charges present a special problem, and are discussed in Chapter 2. Before proceeding with a description of the other input, we illustrate the costing process through a couple of actual examples, to be discussed in Chapters 3 (Marfa) and 4 (Calatori). Finally, in Chapter 5 we deal with the development of the inputs (other than infrastructure).

1.2 Recommendations

Although we believe the system is usable in its present state, it is far short of what it could become at the price of moderate effort. A full-time analyst at Marfa and a full-time analyst at Calatori are required to maintain the system, for at least a year.

A very crucial resource has been lost as a result of the absence of a full-time counterpart. Under normal circumstances, a counterpart would have been closely involved in the project, would have actively participated in the development of the data, and would by now be familiar with all aspects of the system. To make up for the lost opportunity, Marfa and Calatori will need to be particularly careful in the choice of analyst.

It will become apparent in Chapter 5 that the costing system will place heavy demands on the company’s information systems. These demands will be identified as we proceed with our description of the input, and are summarized in Annex B. The persons assigned to the project should either have considerable facility with computers or have access to a person who has.

Finally, Calatori and Marfa managements should show some interest in using the system.

1.3 Response to CFR Comments on Draft Final Report

In a letter dated September, 2002, Mr. Mihai Necolaiciuc, General Director, SMA, summarized as follows CFR’s response to the consultant’sDraft Final Report:

“Taking into account the a.m. and the way the contract was performed, we consider that following measures are required:

a) Orientation of financial evidence of CFR SA, CFR Passengers and CFR Freight to the calculation of unit costing elements, involved in railway transport, that are missing from the actual system;

b) Obtaining of input for the CARTAGE railway costing model with the help of future IT systems, such as IRIS and Financial Oracle, in order to improve the accuracy and the final value of the product. The conceiving of input data sets for the forecast part of CARTAGE demands human, financial resources and a lot of analysis work and rationale in their different development levels, which an IT system is not able to provide;

c) Continuation of activity for improvement and increasing of customization of CARTAGE railway costing model for each of the two end users, by a close co-operation among the IT specialists of CFR Freight, CFR Passengers and the Consultant, by e-mail exchange.

As a conclusion, we agree with the present report, with the a.m. remarks.”

As stated earlier in this section, the system can be used (albeit very cautiously and only as a guide to further study) in its present state. The system can be much improved with a modest effort. Oracle and Iris will be very useful to the process, but many feasible and necessary improvements need not await the installation of these systems.

Unfortunately, since no full-time counterpart was assigned to the consultant during the project, the absence of any trained CFR staff will be a major hurdle. The obstacle can be overcome by immediately putting one or more qualified persons in charge of improving and maintaining the system. The consultant will be available to cooperate with such a person through email communications.

In addition to the required improvements to the model inputs, the consultant will undertake immediately to effect improvements in the model itself as already specified by Calatori. The consultant remains open to any other suggested improvements.

In addition to the general comments quoted above, Mr. Necolaiciuc’s letter contained a number of more specific comments. Those which are not implicitly covered by the above comment are taken into account through modifications and additions to the wording at the relevant sections of the report.

2. INFRASTRUCTURE COSTS

From the vantage point of the operating companies, the only relevant infrastructure costs are their payments to CFR-SA for the use of the infrastructure. The level and modalities of such payments are prescribed annually by ministerial order. To determine the infrastructure cost of a traffic movement, the operating companies need only calculate the increment in access charges caused by the movement. Since the access charges are fixed rates per train-km, the process is simple: the fixed rates per train-mile need only be entered as inputs to the costing model.

However, we need not assume that the current methodology is a permanent feature of the payments for use on infrastructure. CFR management has advised that they expect a new method for the calculation of access charges will be issued by EU and imposed on all operators. We have therefore considered the possibility that other methods will be considered in the future. In this section we:

Review the basis of the current access charge calculations,
Discuss a number of theoretical issues pertaining to access charges in general,
Evaluate the current access charge calculations in terms of the issues raised;
Discuss the implications of access charge methods on the decision-making process of the operating companies,
Evaluate the financial consequences of that decision process both on CFR-SA and on the operating companies

Although, in the final analysis, we agree with most aspects of the current method, we propose a slightly modified alternative.

2.1 Current Access Charge Calculations

The Minister’s Order No. 457 prescribes the tariff structure for charges to the operating companies. Tables 1 to 4 in the Ministerial order are reproduced in Annex A. Table 1 of Annex A provides a list of 14 activities and establishes the percentage of the cost of each of these activities to be covered by access charges (TUI). For example, 96% of the cost of track maintenance is to be recovered through access charges. The remaining 4% is to be recovered through rental payments for specific services or through commercial activities.

Items 1 and 2 of Table 2 provide that expenses chargeable on a train-km basis are initially split equally between Marfa and Calatori. Items 3 and 4 determine that, of the total attributed to freight rail transport, 94% is related to public infrastructure, and is to be recovered through TUI. The remaining 6% are related to assets designated as private are to be recovered according to use through rental charges.

Items 5 and 6 provide analogous figures for passenger traffic. In this case, 98% of the charges relate to public assets.

Table 3 of Annex A provides the final results of the TUI calculation:

For freight traffic, a tariff of 160,000 lei per train-km (equivalent to 5.8 Euro per train-km on the date of approval of the government decision);
For passenger traffic, , a tariff 36,500 lei per train-km, (equivalent to 1.3 Euro per train on the date of approval of the decision).

Finally, Table 4 provides tariffs for services not included in the TUI calculations, such as telecommunication services, IT services, use of space in buildings, etc.

To develop these charges, CFR proceeded as follows:

The activities of the infrastructure company are classified into 14 categories. For each category, a basis of allocation is selected: kilometres of line, train-kms, megawatt of energy consumption, etc. The values of these variables are developed for each of the following users:

Common to Marfa and Calatori
CFR’s own use
Marfa: TUI
Marfa: other charges
Calatori: TUI
Calatori: other charges
SAAF
SMF
Other

The values that are common to Marfa and Calatori are then split between the two according to a variable referred to as “train-kilometres weighted for tonnage and speed”. The information required for these calculations is funnelled to headquarters from the eight regions

Unfortunately, the consultant has had no access to the calculations that supported the presently established rates of 163000 lei per train-kilometres for Marfa and 36500 lei per train-kilometre for Calatori. Instead, we obtained experimental and now obsolete working files that have nevertheless the merit of displaying their methodology in some detail. We understand the method used in the working files was the same as the one used to establish the present rates. We used the working files to replicate the methodology with different data, and to measure the effects of selected modifications to the methodology.

We have replicated the calculation of the working files with 2001 accounting data. Tables 2-1 and 2-2 show the calculation. The total 2001 infrastructure expenses of 10666.1 billion lei were used as a base. This amount was broken down into the 19 groups based on the percentages obtained in the working files. The expenses allocated to each of the 19 groups was then further split among the users, again according to the percentages provided in the working files. Table 2-5 shows the percentages used in these allocations. Table 2-2 shows the results.

The allocation criteria used for most of the 19 items are quite reasonable, for example, the use of telephones is allocated according to the number of subscriptions, the cost of buildings according to the space used, the cost of emergencies according to the number of interventions, etc. Nevertheless, we have tested the effects of slight modifications to three of the largest of the expense items: track (item 1), signals (item 2), and train and shunting control (item 8). The first two of these modifications (items 1) and 2)) were done in response to the issues raised in sub-section 2.2, and are described in that sub-section. They involve the use of a different set of allocation variables. The modification of Item 3 (Train and shunting control) is merely a recalculation using the same method and the same variable (train-kms), but substituting the year 2001 values of train-kms for the forecast values used in the working files.

2.2 Access Charge Issues

Access charges must meet at least two conditions. First, they must, on the average, be high enough to cover the infrastructure company’s total costs. Secondly, they must be low enough to allow the operating companies to attract traffic that yields a significant contribution to the combined marginal costs of the operating and infrastructure companies. The following issues are implicit in these two conditions:

a) Allocation between operating companies: the tariff structure should avoid cross-subsidization; it must be such that the total payment charged to each operating company corresponds to the infrastructure costs it generates. If this condition is not met, one operating company is likely to accept traffic that does not pay its way, and the other to reject traffic that does.

b) Allocation within operating companies: the tariff structure should avoid cross-subsidization within as well as between operating companies. In other words, each segment of the company’s business should be charged its own costs. Currently, there is a single rate per train-km for each company. But if, within the same company, some train-kms generate significantly more cost than others, then again some traffic may be incorrectly rejected or accepted.

c) Allocation of fixed costs: in the preceding paragraphs we referred to the costs of specific traffic movements as if all costs could be attributed unambiguously to specific movements. In fact, there are large blocks of fixed or common costs that cannot be so attributed, except through an arbitrary numerical device. The incremental costs that would be incurred by taking on new traffic, or avoided by relinquishing it, is called the variable cost. However, a company cannot survive by charging only variable costs; it must, therefore, include an allocation of fixed costs in its tariff structure. But any arbitrary allocation of such costs produces the risk of rejecting traffic that could benefit the railway by paying significantly more than the variable cost but less than the fully allocated cost.