Docstoc

Cost benchmarking NATS relative to selected ANSPs

Document Sample
Cost benchmarking NATS relative to selected ANSPs Powered By Docstoc
					Civil Aviation Authority
Cost benchmarking NATS relative to selected ANSPs

Final Report on benchmarking with existing data


20 May 2005
                            NATS cost benchmarking: Final report on benchmarking with existing data



Table of contents
1    Introduction                                                                                1
     Background to the work                                                                      1
     The impact of adding 2003 data                                                              2
     Scope of the work                                                                           3
2    Performance comparison using existing data                                                  6
     Introduction                                                                                6
     Information Disclosure and ACE                                                              6
     The ACE benchmarking analysis                                                               8
     Information Disclosure and the NATS businesses                                             12
     Our approach                                                                               14
     Adjustments to ACE data for facilities shared with the military                            17
     Comparison of cost-effectiveness: gate-to-gate and en-route                                18
     Gate-to-gate                                                                               19
     En-route                                                                                   38
     Sensitivity of the comparison to exchange rates                                            41
     Purchasing Power Parity                                                                    43
Glossary                                                                                        45




                                              i
NATS cost benchmarking: Final report on benchmarking with existing data




                  ii
                              NATS cost benchmarking: Final report on benchmarking with existing data



1        Introduction
Background to the work
1.1      The Civil Aviation Authority (CAA) is undertaking a quinquennial price control
review for the regulated services of National Air Traffic Services (NATS). This will set the
price control for these services for the period 2006-2010. The CAA set out its proposed
approach and timetable for the review in an Initial Consultation Document in March 2004 and
published its Initial Proposals in November 2004.

1.2      The price control currently covers two elements of regulated services, both
operated by the NATS subsidiary NATS (En route) plc (NERL):

         •    en-route air traffic control services provided in the United Kingdom (UK)
              Flight Information Regions (FIRs), charging for which is covered by the
              Eurocontrol Route Charging System (referred to as the “Eurocontrol
              business”); and
         •    services provided in the North Atlantic FIRs, responsibility for which has
              been assigned by the International Civil Aviation Organisation (ICAO) jointly
              to the UK and Ireland; the UK responsibilities are assigned to NERL under
              its licence (referred to as the “Oceanic business”).

In the Initial Proposals, it was proposed to extend the coverage of price control to services
provided as NERL’s London Terminal Approach business (charging for which is not
covered by the Eurocontrol Route Charging System).

1.3        Of the two price-controlled non-contestable businesses, the Eurocontrol business
is by far the larger, with revenues some 20 times greater. The CAA sets a price control as a
result of NERL having, through its licence, a legal monopoly on providing the price-
controlled services. Its objectives in setting price controls are set out by the Transport Act
2000 and, subject to a paramount duty to maintain a high standard of safety, comprise the
furtherance of the interests of users by means including the promotion of efficiency and
economy in the provision of the services by NERL. These are to be achieved while
ensuring that NERL does not find it unduly difficult to finance its activities, and while
taking account of the UK’s international obligations and relevant aspects of the
Government’s environmental objectives.

1.4      In the initial setting of the parameters for the price control, and in their
adjustment following the 2001 traffic downturn and the refinancing of NATS enabled
through the “Composite Solution”, the price control parameters were set largely by
reference to NATS’s legacy internal cost structures. In the Initial Consultation Document, it
was suggested that at this price control review some element of external benchmarking –
comparison with other providers of similar services – could also help inform the process.
Such benchmarking would also assist NATS in identifying areas for improvement and
movement towards best practice.




                                                1
                               NATS cost benchmarking: Final report on benchmarking with existing data


1.5      The CAA appointed a group comprising the Solar Alliance and Steer Davies
Gleave to carry out an exercise benchmarking NATS’s costs against a group of comparator
Air Navigation Service Providers (ANSPs). This exercise was to be carried out in parallel
with three other studies:

         •      a benchmarking of the general administrative costs of NATS - finance,
                administrative IT and premises management - against practices in other UK
                industries, being undertaken by KPMG;
         •      a benchmarking of human resources management by Saratoga; and
         •      a review of the NATS capital investment programme by Logica CMG.

1.6      The object of the benchmarking against other ANSPs was both to provide
information to support the Price Control Review, and to provide a starting point for future
NATS work in benchmarking to identify best practice. NATS recognises the value of the
work and has therefore cooperated fully with the exercise.

1.7      The work comprised three elements:

         i)      identification and analysis of data collected by Eurocontrol, particularly that
                 collected by the Performance Review Unit (PRU) and analysed as part of
                 their ATM Cost-Effectiveness (ACE) programme;

         ii)     econometric analysis of relative efficiency; and

         iii)    original, primary benchmarking, involving new data collection, against the
                 performance of a small sample of other ANSPs.

1.8       An Interim Report covering work on the first two elements, using data available
up to October 2004, was published in November 2004. This report presents a reworking
of the first element, incorporating data made available from the ACE programme for 2003.

1.9      The Interim Report also contained some analysis of variables relating to traffic
complexity, and some statistical analysis of ANSP performance derived from 2001 and
2002 ACE data. These have not been updated with the 2003 ACE data as the major
differences from 2002 can be inferred from the first piece of work.

The impact of adding 2003 data
1.10      In our Interim Report we observed that NATS in 2001 and 2002 was a higher-
cost service provider than the average of twelve comparable ANSPs. In broad terms,
NATS has higher non-ATCO costs, which are compensated for by lower ATCO
employment costs and higher ATCO productivity. The big differences in costs arise from
the exceptional items that arise from one-off experiences, such as the financial
restructuring, and from depreciation and finance costs. There had been a substantial
improvement between 2001 and 2002.




                                                 2
                              NATS cost benchmarking: Final report on benchmarking with existing data


1.11     The position in 2003 was broadly similar. The fall in non-exceptional costs
continued, but there was a big increase in these exceptional items. NATS non-ATCO
operating costs, after these exceptional items, were now comparable to those of the others
(although it is possible that some costs in other ANSPs might reasonably be considered
“exceptional”).

1.12      The “rankings” of NATS according to various performance indicators are
particularly sensitive to exchange rate changes. In our analysis we have used 2002
exchange rates, so that it is comparable with our Interim Report. We note, however, that if
we had used 2003 exchange rates throughout the analysis, we would have obtained a more
favourable ranking for NATS in all three years examined, because of the fall of the pound
between 2002 and 2003.

Scope of the work
1.13      Given the existence of other parallel studies, the main focus of the work has been
on operating costs and, within operating costs, on those elements that are not covered by
the benchmarking of general administrative costs, although other costs will not be excluded
if relevant and interesting. In particular, there is a potential trade-off between operating
and capital costs so that some awareness of ANSPs’ relative capital intensity would be
helpful. The study is focused on identifying areas of cost difference where there is the
possibility of action to move NATS towards best practice. However, the analysis will
necessarily be constrained by the way in which comparator ANSPs describe their business
and divide their costs.

1.14     The Economic Regulation function of the CAA covers only certain areas of
NATS’s activities, and not all of the regulated activities are currently under price control.
For maximum usefulness for the Price Control Review, then, the focus of benchmarking
should be on those regulated parts of the NATS business, divided up according to the
differences in price control. However, it must be recognised in an international
benchmarking exercise that other ANSPs do not divide the costs and revenues of their
businesses up in the same way.

1.15   In UK airspace, air navigation services for civil aircraft are divided into four
components:

         •    en-route services – regulated and price-controlled
         •    terminal approach services in the London Terminal Area – regulated but not
              currently price controlled, although extension of the price control to these
              services was proposed in the Initial Proposals
         •    services to North Sea helicopters – regulated but not price controlled
         •    aerodrome control services – regulated by conduct only, since there is
              deemed to be competition for the market.

1.16     The relevant divisions, together with the corresponding operational divisions of
NATS, are illustrated in the figure below. In the diagram, the vertical axis broadly
represents flight levels, and the horizontal axis broadly a northwest-southeast cross-section



                                                3
                                         NATS cost benchmarking: Final report on benchmarking with existing data


across UK airspace. The bold boundaries represent NATS operational units. The
different colours represent different price control and regulatory regimes. In the diagram
ACC is used for Area Control Centre, and TCC for (London) Terminal Control Centre at
West Drayton. TMA is used for Terminal Manoeuvring Area (an area around an airport or
set of airports in which aircraft are controlled separately from the area control centre).
TWR is a conventional abbreviation for the air traffic control operation covering final
approach, located generally in the tower at an airport.

1.17      The operational boundaries do not always coincide with the regulated businesses.
In particular, the operation of the London TCC provides both services that are classified as
“en-route” and services that are classified as “London Terminal Approach”. In addition,
there are minor anomalies: final approach at London City airport (LCY) is the
responsibility of London TCC; that at Manchester airport is the responsibility of
Manchester ACC; and the TMA units at Cardiff and Belfast provide some en-route services
to overflying aircraft that pass through the TMAs.

           Figure 1.1            Regulatory and operational boundaries in NATS



                         Scottish ACC                                 London ACC (Swanwick)
                          (Prestwick)

   Oceanic                       Other                                  Southampton
    ACC                        Scotland                                    TMA
 (Prestwick)                      and                 Manchester
                   Belfast                                            Cardiff
                               Northern                 ACC                                London TCC
                   TMA                                                 TMA
                                Ireland                                         Other     (West Drayton)
                                TWRs                  Manchester                E&W
                                                        TWR                     TWRs

                                                                                        Other London
                                                                                           TWRs         LCY

 Key to regulation and pricing regimes
               NERL en-route and oceanic: under price control
  NERL London Terminal Approach: regulated, no price control                                NATS
                                                                                          operational
       NERL North Sea Helicopters: regulated, no price control                               units
              NSL Aerodrome Control: conduct regulation only
        NERL services to military (not regulated) are not shown


1.18      Most (if not all) other service providers use only a two-way split, between en-
route and terminal services, for pricing and cost allocation purposes – they do not
distinguish between terminal approach services and aerodrome control services.
Furthermore, the boundary between en-route and terminal services is not consistently
specified between providers (although the Single European Sky legislation will require this
to change in future). This potential inconsistency, and how it can be dealt with, needs to
be borne in mind throughout our analysis.




                                                                  4
                              NATS cost benchmarking: Final report on benchmarking with existing data


1.19      A further potential inconsistency in cost allocation arises because certain costs –
to a large extent not controllable by ANSPs – are included in the costs reported by a
number of European service providers, but not in those of NATS. These costs include
variously in different ANSPs: meteorological costs (MET); the costs of regulation;
Eurocontrol costs; and certain payments made to national governments. We have
excluded these costs from the analysis. This confines our examination to NATS’s own
costs, and is also consistent with the work done by the Eurocontrol Performance Review
Unit (PRU) in comparing performance across European ANSPs, where most of their
analysis is confined to the costs of Air Traffic Management (ATM) and Communications,
Navigation and Surveillance (CNS).




                                                5
                                     NATS cost benchmarking: Final report on benchmarking with existing data




2          Performance comparison using existing data
Introduction
2.1       This chapter describes the work that we have done in extending the analysis of
the data collected by the Eurocontrol Performance Review Unit (PRU) in connection with
their ATM Cost-Effectiveness (ACE) benchmarking exercises. It makes some necessary
adjustments to the ACE data, and extends the work of ACE to focus on the performance
of NATS in comparison with a set of more closely comparable ANSPs. Because of the
regulatory framework within which NATS operates, it extends the analysis of ACE to the
subset of en-route costs as well as gate-to-gate.

Information Disclosure and ACE
2.2       The Performance Review Commission (PRC) of Eurocontrol, supported by the
PRU, has instituted a programme of mandatory Information Disclosure to facilitate
performance review of air navigation service (ANS) provision in the Eurocontrol member
states. Their analysis of the data disclosed has resulted in two ATM Cost-Effectiveness
(ACE) benchmarking reports: ACE 2001, covering 29 ANSPs, published in September
2003; and ACE 2002, covering 32 ANSPs, published in August 2004. An earlier pilot
exercise, ACE 2000, covered 17 ANSPs participating voluntarily and the results were not
made publicly available. NATS was a participant in all three exercises. As the ACE 2000
data set has not been subject to the same degree of checking and validation, it was not used
as part of the present study.

2.3      A further exercise for 2003, covering 34 ANSPs, has been undertaken, and an
ACE 2003 report is currently in draft. The present report extends the work in Chapter 2 of
our Interim Report to include results from the ACE 2003 exercise.

2.4      The PRU’s information disclosure mandate relates to the responsibilities of
Eurocontrol. It therefore explicitly excludes Oceanic ANS. Within the European Flight
Information Regions (FIRs) its remit covers the provision of air navigation services to all
phases of flight, “gate-to-gate”.

2.5       The PRU has made substantial efforts over a number of years to obtain a
consistent and reliable data set. They have worked with an active group of ANSPs to
achieve harmonised definitions and useful performance indicators. They have paid detailed
attention to the quality of submissions, and have reconciled each submission where
possible to Annual Reports (where available), CRCO submissions1, and CFMU2 data on
traffic.




1 The Central Route Charges Office (CRCO) of Eurocontrol collects information on km flown by each airline in the
airspace of each member state, assesses the appropriate charges to be made, bills the airlines, collects the cash, and




                                                          6
                                     NATS cost benchmarking: Final report on benchmarking with existing data


2.6       Despite these efforts, certain data issues have inevitably emerged subsequent to
publication of the ACE reports. The PRU has endeavoured, when such issues are
resolved, to correct retrospectively previous years’ data to reflect the improved
understanding. For this reason, the data used in this analysis is not in all respects the same
as has been analysed in the published ACE 2001 and 2002 reports. Neither is it in all
respects the same as that analysed in our Interim Report. Some of the retrospective
adjustments made by the PRU have an appreciable impact on the 2001 and 2002
performance ratios, in particular those made specifically for NATS, described below.

2.7     Our analysis differs from that of the PRU in some further respects. Two of these
concern NATS in particular, and two are generic. The reasons relating specifically to
NATS concern:

           •     the exclusion of flights through off-airways airspace, which nevertheless
                 require services from NATS; and
           •     the treatment of the costs of the infrastructure that NATS shares with the
                 military.

2.8        In early discussions with the PRU, they gave us an estimate of the size of the first
effect. We reviewed its implication and decided that for the purposes of this benchmarking
exercise it was appropriate to include the off-airways flights in NATS’s output. A data set
adjusted by the PRU for this effect was used in the analysis. Subsequently, however, the
PRU has reviewed their position and decided to continue not to include the off-airways
flights in NATS’s output. Both for comparability with the analysis in our Interim Report,
and because these flights require services from NATS, we have continued to include them
in our analysis. We have made a provisional estimate of their value for 2003.

2.9       The second difference results from the treatment of the costs of shared
infrastructure. In the ACE 2001 and ACE 2002 reports, no adjustment was made to the
costs of en-route service provision for the sharing of infrastructure with the military, and
the revenue obtained from the military for this. Subsequently, however, NATS and the
PRU have agreed a changed position. NATS’s en-route costs are, in the ACE 2003 report,
reduced by the revenue obtained from the military. This is different from the treatment in
our Interim Report, and for consistency we have retained the treatment used at that stage.
Our treatment of these costs is discussed in more detail in paragraphs 2.54 to 2.58.

2.10    The generic reasons for difference between our analysis and that in ACE 2003
concerns:

           •     the exchange rates; and



distributes it to member states who in turn distribute it to ANSPs. Each member state makes returns of its costs and
proposals for pricing to the CRCO at least annually.
2 The Central Flow Management Unit (CFMU) of Eurocontrol undertakes the coordination of flow management across
member states and beyond. The data on flight plans that it uses can be analysed to assist in benchmarking work of this
kind




                                                          7
                             NATS cost benchmarking: Final report on benchmarking with existing data


         •   the definition of output.

2.11     Our Interim Report was based on conversion of sterling to euros at 2002
exchange rates. In the subsequent year, the pound fell substantially (around 8%) against
the euro. For consistency with our previous work, we have, however, retained 2002
exchange rates. This assumption has a substantial effect on the ranking of NATS by
performance indicators. As a consequence, the ranking of NATS in 2003 by the financial
indicators will not be the same as appears in the 2003 ACE report, which is, of course,
based on 2003 exchange rates – the apparent performance of NATS in ACE 2003 will be
better by around 8% than in this analysis.

2.12     The PRU defines a composite measure of output, taking into account both the
IFR flight-hours, which are considered to be the most appropriate measure of output for
en-route services, and the number of terminal movements at airports controlled by the
ANSP, which is considered to be the most appropriate measure of output for terminal
navigation services. For ACE 2003, they changed the parameter used in that definition to
weight the two components. For the purpose of year-on-year comparison, we wish to
retain consistency in the definition of the output measure between years. For ready
comparability with the analysis in our Interim Report, we wish to retain the parameter
value used in the ACE 2001 and ACE 2002 analysis. In practice, as described in 2.67, our
conclusions are not sensitive to this change.

2.13     For these reasons, our analysis is not wholly consistent with that in ACE 2003.

The ACE benchmarking analysis
         Scope of the ACE reports

2.14      The ACE reports examine and compare a wide range of indicators across all
participating ANSPs. Furthermore, they have developed a consistent analytical framework
for the analysis of cost-effectiveness, which helps identify and understand the reasons for
differences in performance at an aggregate level, and to understand the trade-offs between
different components of cost-effectiveness.

2.15      ACE 2001 covered 29 ANSPs, comprising the ANSPs of all current Eurocontrol
member states but three (Croatia, Cyprus, and Greece), plus three Baltic States. ACE 2002
covered 32 states, including a new member state (Albania) and two of the three states not
complying in 2001 (Croatia and Cyprus). ACE 2003 has extended the analysis to two more
states, Greece and Ukraine, a new member of Eurocontrol.

2.16      The analysis in the ACE reports focused on the element of ANS costs that the
PRU labelled “ATM/CNS costs”. This does not strictly correspond to the ICAO
definition of ATM/CNS, since it includes in addition the costs of Aeronautical
Information Services (AIS) and Search and Rescue (SAR). It excludes, however, a number
of elements of costs that are substantially outside the control of ANSPs – MET costs,
Eurocontrol costs, and payments to governments for regulatory and other services. The




                                               8
                                        NATS cost benchmarking: Final report on benchmarking with existing data


PRU’s ATM/CNS costs are particularly suitable for a comparison of NATS with other
ANSPs since the costs excluded are largely external to NATS.

              Measuring cost-effectiveness

2.17    The PRU measured cost-effectiveness by dividing a measure of output by a
measure of input. Their measures of output were:

              •     for en route ANS, the en-route flight-hours controlled;
              •     for terminal ANS, the terminal movements controlled.

2.18       Given the potential inconsistency in the way that boundaries are drawn between
en-route and terminal ANS in different ANSPs, and the PRU’s gate-to-gate remit, there
was little separate analysis of en-route and terminal ANS performance. Instead, the PRU
focused on gate-to-gate performance. They therefore needed a composite measure of
output. They defined that measure so that the relative importance of en-route and terminal
output in the composite measure was the same as the relative importance of en-route and
terminal costs across the system. This resulted in a definition of the composite output
measure:

          Composite gate-to-gate flight-hours = en-route flight-hours + 0.24 x IFR airport movements

2.19     This was the measure of output that was used in ACE 2001 and ACE 2002. In
ACE 2003 the definition has changed3. However, for our purposes it is important to retain
comparability across different years, and we have therefore carried out the analysis using
the definition from ACE 2001 and ACE 2002. In practice variation in this parameter
makes very little difference, as described in paragraph 2.67.

2.20          The PRU used two measures of input:

              •     the ATM/CNS provision costs borne by ANSPs – for the users these
                    correspond broadly, under cost recovery, to the costs they pay for the
                    services used;
              •     an estimate of the costs of ATFM delays faced by airspace users – this is the
                    most important element of the external costs faced by users arising from the
                    quality of service provided.

2.21     They were therefore able to define two measures of cost-effectiveness. The
financial cost-effectiveness is the ratio of output to ATM/CNS provision costs. The
economic cost-effectiveness includes, in addition, the costs of ATFM delays in the
denominator.

2.22     For clarity, they inverted these ratios for presentation purposes, so that high cost-
effectiveness is presented as a low cost per unit output.



3   The weighting in ACE 2003 is be 0.26 composite flight-hours per IFR airport movement, rather than 0.24




                                                            9
                               NATS cost benchmarking: Final report on benchmarking with existing data


         The performance framework

2.23     The PRU structured their further analysis of the ACE data around their
“performance framework”. This decomposed the overall cost-effectiveness ratio (service
provision costs per unit output) into a number of component ratios that, multiplied
together, gave the overall ratio. This performance framework is illustrated in Figure 2.1.

         Figure 2.1      The ACE performance framework

                    Gate-to-gate
                     composite
                    flight-hours
                                            ATCO productivity



                ATCO-hours on duty

                                                                         Financial
                                             Employment costs
                                                                     cost-effectiveness
                                              per ATCO-hour
                                                                          indicator

               Employment costs for
                 ATCOs in OPS


                                             Support cost ratio



             ATM/CNS provision costs


2.24     The purpose of this framework is to identify at a more detailed level how
apparent differences in performance arise. The main decomposition used by the PRU in
the ACE analysis is a three-way split of the overall financial cost-effectiveness ratio into:

         •    ATCO productivity (flight-hours per ATCO-hour);
         •    ATCO employment costs (€ per ATCO-hour); and
         •    the “support cost ratio” – the ratio of total service provision costs to ATCO
              employment costs.

2.25     The “support costs” implied in the third ratio include a number of disparate
elements, which can in principle be broken down further using the data:

         •    employment costs for staff other than ATCOs on operational duty;
         •    operating costs other than employment costs;
         •    capital-related costs, comprising:
              − depreciation; and




                                                10
                                      NATS cost benchmarking: Final report on benchmarking with existing data


                 − finance costs, as used in the calculation of the cost base used for en-route
                   charging4.

           ATCO productivity by ACC

2.26      The ACE reports also measured ATCO productivity (flight-hours per ATCO-
hour) at the ACC level. In 2002 and 2003 London ACC and Scottish ACC had ATCO
productivity of around 0.9 and 1.2 flight-hours per ATCO-hour respectively, higher than
the European average of around 0.7 flight-hours per ATCO-hour. London TCC and
Manchester ACC were around the average. In all these cases, the productivity appeared to
have fallen from its 2001 levels.

2.27      ACE attempted to link ATCO productivity with a number of variables intended
to reflect traffic complexity and variability:

           •     “adjusted density” – traffic density adjusted to reflect the concentration of
                 traffic in particular parts of the airspace;
           •     the intensity of vertical movements, measured as vertical movement (in feet)
                 per km flown;
           •     the ratio of traffic in the peak week to that in the average week (seasonal
                 variability).

2.28      On the face of it, it might be expected that more dense airspace might result in
increased controller workload, and hence reduced controller productivity. However, this
relationship between adjusted density and productivity was not observed. This might be
because these diseconomies of complexity might be balanced by economies of scale of
operation, and in particular by increased flexibility to match controller resources to
fluctuating demand. It is notable that the NATS ACCs, with the exception of Scottish, are
among the densest on this measure, with London TCC being almost twice as dense as the
second most dense (Manchester), and even London ACC being among the top 15.

2.29      There was an expectation that a high intensity of vertical movements would
reduce productivity, and this was indeed found to be true, but only for the densest
airspaces. Manchester and London TCC are among the ACCs with the highest intensity of
vertical movement.

2.30     There was an expectation that high seasonal variability would reduce productivity.
This was not demonstrated in the ACE reports. The NATS ACCs have relatively low
seasonal variability, with peak week/average week ratios of around 1.1.




4 This finance cost was intended to represent the cost of capital. ANSPs may take either of two approaches to assessing
this cost; they can either used the actual net cost of financing the ANSP (the approach adopted by NATS), or use a
notional figure comprising net assets multiplied by a cost of capital. ANSPs were asked to apply the approach they used
in calculating their en-route cost base to terminal costs as well, for the purposes of Information Disclosure.




                                                          11
                                      NATS cost benchmarking: Final report on benchmarking with existing data


Information Disclosure and the NATS businesses
2.31      As noted in Chapter 1, the categories of business used in Information Disclosure
are not those that are used in the NATS licence and hence used to regulate NATS. The
remit of the PRC and PRU is explicitly gate-to-gate and their requirements for information
disclosure therefore cover all phases of flight.

2.32     Nevertheless within the gate-to-gate umbrella, Information Disclosure divides Air
Navigation Services businesses into three categories:

            •    en-route navigation services
            •    terminal navigation services
            •    other services.

2.33     The en-route business is defined by the Specification for Information Disclosure
as the business for which Eurocontrol en-route charges are collected – essentially,
therefore, it is that business for which costs are declared in the CRCO submissions. For
NATS, this corresponds with the scope of business currently covered by the RPI-X
regulated en-route charge.

2.34      The terminal navigation services business is not explicitly defined in the
Specification. ANSPs are allowed to use their own definition of what activities fall into this
business. Implicitly (although this is not explicitly stated) en-route plus terminal air
navigation services comprise the gate-to-gate air navigation services over which
Eurocontrol’s remit falls. ANSPs are required in the Specification to state their definition
of terminal ANS and to describe their cost allocation methods. We understand, however,
that the information provided by some ANSPs is not complete in this respect.

2.35      Mapping this definition of services to NATS, the Specification for Information
Disclosure’s terminal navigation services business is an aggregation of NERL’s London
Terminal Approach business, which while regulated, is not currently under price control,
and NSL’s unregulated Aerodrome Control business. The mapping of the boundaries in
the ACE data set to those in NATS is shown in Figure 2.2. The red dotted line encloses
the areas described in ACE as relating to terminal ANS, and the green dotted line those
described as other ANS. All other services are included in en-route ANS5.




5 In 2002 the costs of service to North Sea helicopters were included in en-route. NATS has agreed that this should more

properly be included in “other ANS” and in the analysis we have corrected for this.




                                                          12
                                        NATS cost benchmarking: Final report on benchmarking with existing data


          Figure 2.2             NATS businesses and ACE data



                        Scottish ACC                                      London ACC (Swanwick)
                         (Prestwick)

    Oceanic                     Other                                      Southampton
     ACC                      Scotland                                        TMA              London TCC
  (Prestwick)                    and                 Manchester
                  Belfast                                                                     (West Drayton)
                              Northern                 ACC                Cardiff
                  TMA                                                      TMA
                               Ireland                                              Other
                               TWRs                  Manchester                     E&W
                                                       TWR                          TWRs

                                                                                            Other London
                                                                                               TWRs        LCY

 Key to regulation and pricing regimes
                          NERL en-route and oceanic                     NATS             Information       Information
                  NERL London Terminal Approach                       operational         Disclosure        Disclosure
                                                                         units         “terminal ANS”      “other ANS”
                        NERL North Sea Helicopters
                             NSL Aerodrome Control
       NERL services to military (not regulated) are not shown


2.36      This gives rise to some issues of comparability. The first concerns the en-route
business. The rules in the Route Charging System for how costs are allocated to en-route
are very loose, and in any case are not universally applied. Thus if we benchmark NATS’s
en-route business against other ANSP’s en-route business, some of the observed cost
differences may result from these underlying variations in the approach to cost allocation.
Because of this, the PRU have focused more on gate-to-gate costs and output measures, as
noted above.

2.37      If, however, we follow the PRU approach and benchmark the gate-to-gate
business, we are looking, at the NATS end, at a combination of businesses that are treated
differently for UK regulatory purposes, and inferring relative performance for NATS’s
separate businesses may not be easy. In any case, it is clear that ACE data will offer us no
help in benchmarking the NERL London Approach business separately from the NSL
aerodrome business.

2.38      In future, this inconsistency will be mitigated, as the Route Charging System will
be replaced by a Single Sky uniform charging system, covering all phases of flight. It is
likely that the implementation rules for this system will require more uniform cost
allocation between en-route and terminal costs, and more transparency about how it is
done.

2.39      For the present, it may be possible to elucidate the likely scale of any distortions
to the en-route data set that might arise from lack of uniformity in cost allocation by asking
qualitative questions of ANSPs.




                                                                 13
                                       NATS cost benchmarking: Final report on benchmarking with existing data


2.40      We have for our analysis looked at both levels. We have compared en-route
ANS, using the ANSPs’ own definitions. We have also compared gate-to-gate ANS, where
issues of cost allocation are less important, but where the business being benchmarked
does not map perfectly onto the NATS regulated businesses. We have not benchmarked
the terminal ANS business separately, as the issues of cost allocation will be magnified on
this smaller business, and in any case this is not the focus of the CAA’s regulatory price
control.

Our approach
2.41     In the following paragraphs we present our approach for analysis of the ACE
data. This chapter gives a quantitative description of the comparison of performance
indicators of NATS and other ANSPs. In it we examine the overall cost-effectiveness
indicators for NATS and a group of comparable ANSPs, and examine how differences in
apparent performance arise, and how they are changing over time.

            The analytical framework

2.42      We have used in our quantitative comparison an analytical framework close to
that used by the PRU and shown in Figure 2.1, in which an overall performance ratio (cost
per unit output) is measured and compared, then broken down into a series of ratios. We
break down elements in the chain where the inputs are additive.

            Measures of output and input

2.43        For output measures we used those defined by the PRU. These comprise:

            •     en-route flight-hours for the en-route business;
            •     terminal IFR movements for the terminal business; and
            •     a composite output, defined as the weighted sum of the above, for the
                  combined gate-to-gate business.

2.44      The weight for the latter was defined in the ACE reports to reflect the relative
costs of the two businesses. This, of course, changes from year to year, making analysis of
year-on-year changes difficult. Where making use of the composite indicator we have
therefore made all comparisons using the ACE 2002 weights6.

2.45      A criticism of the PRU output definition was that it did not fully reflect the
differential work that was required to provide one flight-hour of services in different parts
of airspace with different complexity.




6If the appropriate weighting is calculated using the sample of thirteen ANSPs that we have retained for analysis (see
paragraphs 2.50 to 2.53), the weighting is the same for both years, and is the same as the value used for the whole sample
by the PRU in ACE 2002, and in our present analysis.




                                                            14
                              NATS cost benchmarking: Final report on benchmarking with existing data


2.46      We understand the logic behind the critics’ position; it is clear that, for a fair
comparison, the specific impact of exogenous factors, loosely grouped as “complexity”,
need to be understood. However, the most transparent way of dealing with this effect in
this kind of quantitative comparison is to assess what makes airspace difficult to handle and
allow explicitly for it (as a mitigating factor).

2.47      For an input measure we have used the ATM/CNS cost measure developed and
calculated by the PRU. This measure screens out elements that are not under the direct
control of the ANSP, such as the meteorological costs (MET), Eurocontrol costs, tax
payable to central government, and regulatory costs. Such elements differ widely between
ANSPs according to national practices and statutes. These elements of cost are in almost
all cases not borne by NATS, so use of this cost definition aids comparison.

2.48      We have also retained the concepts introduced by the PRU of financial and
economic cost-effectiveness. The financial cost-effectiveness is measured as the ratio of
outputs to the financial costs of ATM/CNS, as defined in the previous paragraph. It
therefore captures (part of) the costs that are incurred by the ANSP in providing the
service. It does not reflect any costs that are imposed on the user that are linked to the
quality of services provided by the ANSP, and in particular does not recognise the costs of
delays imposed on the users. The PRU therefore define economic cost-effectiveness which
includes in addition the costs of air traffic flow management (ATFM) delay. In a similar
way to the use of ATM/CNS costs in the financial cost-effectiveness analysis, omitting
costs that are largely outside the ANSP’s control, the economic cost should include only
that portion of delay that is directly attributable to the quality of service provided by
individual ANSPs and omit other causes of delay such as delays caused by runway
constraints, and reactionary delay caused by delays to prior flights. ATFM delay is not a
perfect measure, since it includes delays caused by weather, but is the best measure
currently available on a comparable basis across ANSPs.

2.49       We note that this economic cost-effectiveness indicator omits a number of
important elements of the full economic costs of delay and other elements of imperfect
service: the reactionary delays (those delays caused by delays to previous flights of the same
aircraft), and the costs of sub-optimal routings and flight profiles. Furthermore, the figures
used by the PRU for the unit cost of ATFM delay represent the cost to the airlines only,
and exclude the costs to passengers and freight. Nevertheless, we believe, like the PRU,
that examining the costs of ATFM delay can provide a useful supplement to examination
of ANSPs’ own costs.

         The choice of a set of comparators

2.50     We have applied the analytical framework to compare NATS both to individual
comparators, and to a wider group of comparable ANSPs as a whole. Comparison to the
whole of the compliant ANSPs disclosing information gave rise to three problems:

         •    the composition of the group varies from year to year;




                                               15
                                         NATS cost benchmarking: Final report on benchmarking with existing data


              •     the apparent performance of a large number of the smaller ANSPs was
                    driven very largely by the relatively low wage rates prevalent in those
                    countries; and
              •     it was widely felt that managing air traffic in a dense airspace was different in
                    kind from that in a less dense airspace.

2.51      While the latter two effects could be readily incorporated into a statistical
analysis of performance (and indeed would add richness to the variability and explanatory
power of the analysis) they make a quantitative, descriptive analysis difficult to interpret.
We therefore selected a group of ANSPs in countries with broadly similar unit labour costs,
and relatively high airspace density.

2.52     The composition of this wider group was determined on objective criteria –
density and levels of unit labour costs. Figure 2.3 shows the European ANSPs classified by
the adjusted density, as calculated by the PRU, and GDP/head, which is a readily available
proxy for unit labour costs7. For Spain, the Continental ACCs are separated from the
Canarias ACC, which is remote from the European continent and is not subject to full
radar control.

2.53      We selected a set of thirteen ANSPs with relatively high labour costs and adjusted
density, as shown in Figure 2.3. It corresponds to ANSPs in the pre-2004 EU countries,
plus Switzerland, but excluding Finland and Portugal, where densities are very low, and
Greece, where data was not disclosed in 2001 and 2002. The resulting set of thirteen
ANSPs has been the subject of the further analysis described in following sections.




7   The variables used were 2002 values, since the selection was done at the time of producing our Interim Report.




                                                             16
                               NATS cost benchmarking: Final report on benchmarking with existing data


          Figure 2.3      ACE ANSPs classified by adjusted density and GDP/head

         30
                                                               Belgocontrol

         25
                                                                   LVNL, Netherlands
                                                                                           Switzerland
                                                                 Maastricht UAC
         20                                 DFS, Germany
                                                                  UK
 Adjusted
          15
  density                                                      France
                                                      Italy       Austria
         10
                                           Continental Spain
                                                                                 Denmark
                                                          Sweden              Ireland
          5                                Canarias
                                                                  Finland                            Norway
                                     Continental Portugal

          0
               0   5000    10000   15000     20000       25000      30000     35000    40000      45000   50000
                                            GDP per head 2002 (€/year)



Adjustments to ACE data for facilities shared with the military
2.54     In our Interim Report, we made a substantial adjustment arising from the fact that
NATS shares facilities with the military, and obtains a revenue from the Ministry of
Defence for this. There is no precise analogue in other ANSPs. In Germany, Sweden and
Switzerland, the ANSP provides navigation services for military Operational Air Traffic
(OAT) – not just shared facilities - and derives a revenue from this. In their submissions,
an estimate is made of the cost of providing such a service, and this cost is allocated to
“other ANS” rather than en-route. In the DFS submission, this cost is exactly equal to the
revenue that DFS obtains from the military.

2.55     It was clear, therefore, that the treatment in the published ACE 2001 and 2002
reports was inconsistent and tended to underestimate NATS’s cost-effectiveness. Fair
comparison required an adjustment to be made. At one extreme, it could be argued that
NATS was able to obtain a revenue from sharing the facilities, and this revenue could
therefore be netted off the en-route costs. At the other, it could be argued that, in contrast
to the German position, the shared facilities were required by NATS to carry out their
operations, and that what should be netted off was the avoidable cost of sharing facilities
with the military. It was recognised, however, that this would be very difficult to assess.
An intermediate approach would be to allocate the costs of the shared facilities between
the two parties – NATS en-route business and the military. This would appear to be the
approach adopted by the Swiss. We also understand from DFS that the revenue obtained
from the military is, by agreement, the cost of providing the service; on this understanding,
the Germans are adopting a similar approach. This approach was also consistent with the
approach adopted in NATS’s regulatory accounts. It was agreed that we should adopt this




                                                 17
                              NATS cost benchmarking: Final report on benchmarking with existing data


middle way, and an estimate was made of the appropriate allocation of the costs of the
shared facilities to the military. This estimate was netted off the en-route costs as disclosed
to the PRU.

2.56     As discussed in paragraph 2.9 this was not the approach that was eventually
agreed between NATS and the PRU for the purposes of ACE 2003. However, we have
continued to follow this approach for consistency with the analysis in our Interim Report.
Use of the approach found in ACE 2003 would reduce NATS’s overall unit costs by
around 1.5%.

2.57     We explored whether there were any analogous inconsistencies in other members
of the “twelve”. In consultation with the PRU, it was determined that as far as we and they
were aware, there were none.

2.58      We discussed the possibility of making other adjustments to the ACE datasets,
particularly in respect of “exceptional items”. It was agreed that no further adjustments
should be made for the present work. We discuss the issue of exceptional items further in
paragraphs 2.110 onwards.

Comparison of cost-effectiveness: gate-to-gate and en-route
2.59     Having made the necessary adjustments, we undertook detailed comparison of
NATS against the selected other ANSPs. In subsequent discussion we use the term
“thirteen” to refer to the set of comparators including NATS, and “twelve” to refer to the
set excluding NATS.

2.60     We carried out the analysis first at a gate-to-gate level. This was because:

         •    it was consistent with the PRU’s approach;
         •    some of the required breakdown of costs is not available on an en-route
              basis; and
         •    it avoids distortions arising from different boundaries in defining the en-route
              business in different ANSPs.

2.61     Nevertheless, we follow this comparison with an investigation of the en-route
business, because that is the business that is subject to price regulation by the CAA.

2.62     For gate-to-gate, we follow the framework used by the PRU, focusing on NATS
and the twelve, and investigating more deeply when interesting issues arise, as far as the
data permits. We first examine overall financial and economic cost-effectiveness. We then
break down the financial cost-effectiveness into individual components: ATCO
productivity, ATCO employment costs, and other costs (“support costs”, in PRU
terminology). We also look in more detail at individual components of these support costs:
non-ATCO operating costs, depreciation, and finance costs, and exceptional items. In
each case we examine the comparative position of NATS against other members of the set,
looking at data from 2001, 2002 and 2003.




                                               18
                              NATS cost benchmarking: Final report on benchmarking with existing data


2.63      For en-route, our analysis is not so deep, since the full set of information is not
available. We compare performance at the level of overall service provision costs,
operating costs, and the components of capital-related costs.

2.64      Finally, we investigate two related issues that emerged in the course of the
analysis:

         •    how sensitive are the results to the exchange rates used?
         •    is there a case for using Purchasing Power Parity as a basis for comparison?

Gate-to-gate
         Financial cost-effectiveness

2.65       Figure 2.4 shows the financial cost per unit output for the thirteen. Note that on
graphs of this kind, showing all thirteen individually, the averages presented are for the
thirteen. On graphs where NATS is compared to the others, it is compared to the twelve,
excluding NATS. As discussed above, gate-to-gate output is a composite measure made up
of both flight-hours controlled and terminal IFR movements controlled, as discussed in
paragraphs 2.43 to 2.44. A relatively high level of this indicator therefore indicates
relatively poor performance in this respect. ANSPs are ranked in this and subsequent
charts with the highest unit costs on the left. NATS’s overall unit costs are well above the
average costs for the thirteen. The only ANSPs in the thirteen with higher costs are LVNL
and Belgocontrol, and in both those cases, costs are overstated, since the ANSP bears the
cost of CNS infrastructure for the whole of the country’s territory, while the upper air
traffic is served by Maastricht. The same overstatement is true, to a lesser extent, for DFS.




                                               19
                                      NATS cost benchmarking: Final report on benchmarking with existing data


               Figure 2.4         Financial cost per composite flight-hour, 2003

           700


           600                                                            Average for thirteen


           500


 € per     400
 flight
 hour      300


           200


           100


               0

                                                                   e
                    l




                                      na




                                                    AV




                                                                                           A




                                                                                                                     C
                                S




                                             FS




                                                                                                             en
                                                                                                   IR
                                                                                  A
                          L




                                                                             l
                 ro




                                                                           ro
                                                                id
                              AT
                         N




                                                                                                                  UA
                                                                                         IA
                                                                                 N
                                     Ae




                                                                                                            d
               nt




                                            D




                                                                                                   IA
                                                                         nt
                                                              gu
                                                  EN
                        LV




                                                                                 D




                                                                                                         we
                              N
             co




                                                                       Co




                                                                                                                  M
                                                                                                 AV
                                                            y
                                                         Sk




                                                                                                        SS
          lgo




                                                                    o




                                                                                              N

                                                                                                    AN
                                                                 str
          Be




                                                               Au




2.66     Figure 2.5 compares the financial cost per unit output for the years 2001-2003.
Costs are expressed in 2002 prices and converted at 2002 exchange rates throughout. The
position of NATS improved in 2002, with its unit costs falling, but this improvement was
reversed in 2003. The average costs for the thirteen have risen.

2.67      In principle, this indicator is sensitive to the weighting of the en-route and
terminal components in total output, as described in paragraph 2.44. For comparability
with our previous work, we have used the 2002 weighting of 0.24 terminal movements per
flight-hour, while ACE 2003 uses 0.26. However, in practice this makes little difference;
use of 0.26 would increase NATS’s unit costs by less than 1% relative to the average.




                                                         20
                                        NATS cost benchmarking: Final report on benchmarking with existing data


            Figure 2.5              Financial cost per composite flight-hour, 2001 and 2002

                                                                                                          2001
            700
                                                                                                          2002
                                                                                                          2003
            600                                                                                           2001 average
                                                                                                          2002 average
            500                                                                                           2003 average


  € per     400
  flight
  -hour     300


            200


            100


                 0

                                                                            e
                      l




                                                      AV
                                         na




                                                                                      A




                                                                                                                 C
                                               FS
                                  S




                                                                                                      en
                                                                                              IR
                                                                                 A
                            L




                                                                            l
                   ro




                                                                         ro
                                                                        uid
                                AT
                           N




                                                                                                              UA
                                                                                     IA
                                                                                N
                                        Ae




                                                                                                      d
                 nt




                                              D




                                                                                            IA
                                                                      nt
                                                    EN
                          LV




                                                                                D




                                                                                                   we
                                                                     yg
                                N
               co




                                                                   Co




                                                                                                            M
                                                                                          AV
                                                                  Sk




                                                                                                 SS
            lgo




                                                                 o




                                                                                          N

                                                                                              AN
                                                             st r
           Be




                                                           Au




2.68   The increase in cost per unit output for the thirteen (and indeed the fall for
NATS) could have risen through two effects:

           (a)        a rise in total costs; or

           (b)        a fall in output.

2.69      The latter possibility is particularly important given the events of 11 September
2001. We therefore investigated the extent to which these two factors contributed to the
overall rise in unit costs.

2.70      Figure 2.6 shows the cost and volume components of the change in gate-to-gate
financial costs per unit output between 2001 and 2003. The upper graph shows the change
between 2001 and 2002 and the lower that between 2002 and 2003.

2.71      The thirteen as a whole showed a drop in volume of 2.2% in 2002. The fall in
volume for NATS was slightly greater than the average at 3.0%, although it should be
borne in mind that the 2001 figure for NATS included six full months at post-9/11 traffic
levels (October 2001 to March 2002), whereas that for the other eleven only included three
months. This suggests that the overall impact of 9/11 might have been substantially
greater for NATS.

2.72      Many ANSPs showed substantially greater falls in volume; Belgocontrol in
particular was badly hit by the collapse of Sabena. DNA on the other hand showed an
increase in traffic.




                                                           21
                             NATS cost benchmarking: Final report on benchmarking with existing data


2.73    Most ANSPs showed increases in costs, despite the falls in traffic. NATS was
among the exceptions.

2.74      In 2002-3, many ANSPs showed increases in costs, some very substantial. The
total cost for the thirteen rose by over 5%. NATS, however, was again an exception, with
costs static in real terms. Volume, by contrast with the previous year, rose by over 1% for
the thirteen as a whole, although that for NATS showed a small decline.




                                              22
                                   NATS cost benchmarking: Final report on benchmarking with existing data


          Figure 2.6         Components of change in unit costs

         20%

                                 2001-2002                                                    Cost effect
         15%                                                                                  Volume effect



         10%



          5%
Change
2001-2
          0%
              na                 l    l      S                                           TS          ide
                Ae    den ontro ontro      DF
                                                   A
                                                 DN ENA
                                                        V
                                                               IA
                                                                    A        NL     AC        IR
                   we       C       c                                   LV        MU   NA AVIA Skygu
                SS      tro      lgo                                                      N
         -5% AN Aus           Be



         -10%



         -15%



         20%

                                  2002-2003                                                   Cost effect
         15%                                                                                  Volume effect



         10%



          5%
Change
2002-3
          0%
              na                l      l
                Ae    den ontr o ontro     DF
                                             S     A    V           A        NL     AC   TS   IR     ide
                   we             c              DN ENA        IA       LV        MU   NA AVIA Skygu
                S S stro C elgo                                                           N
         -5% AN Au          B



         -10%



         -15%




                                                    23
                                           NATS cost benchmarking: Final report on benchmarking with existing data


                Economic cost-effectiveness

2.75     The PRU defines economic cost-effectiveness in terms of the ratio of economic
costs, which include both the ATM/CNS costs borne by the ANSP and the costs of
ATFM delay8, to outputs. As discussed above, a number of economic costs are omitted
from this indicator, notably the costs of delays to passengers, and the costs of sub-optimal
routing or flight profiles.

2.76      Figure 2.7 below shows the economic cost-effectiveness indicator, the economic
cost, in euro, per unit output.

                Figure 2.7        Economic cost per composite flight-hour, 2003

         700
                                                                                     Airport ATFM delays
                                                                                     En-route ATFM delays
         600                                                                         ATM/CNS costs
                                                                                     Average economic costs (13)
         500


    € per 400
    flight
    hour 300


         200


         100


           0
                                   e
                           l




                                                        AV


                                                                  na




                                                                                                    A
                                                                                            C
                                                 FS
                                         S




                                                                                                                     en
                                                                                                           IR
                 L




                                                                                  A
                                                                             l
                        ro




                                                                           ro
                               uid


                                       AT
                N




                                                                                          UA


                                                                                                  IA
                                                                                 N
                                                              Ae




                                                                                                                      d
                                                D
                      nt




                                                                                                          IA
                                                                         nt
                                                      EN
           LV




                                                                                 D




                                                                                                                   we
                             yg


                                       N
                     co




                                                                       Co




                                                                                        M




                                                                                                        AV
                           Sk




                                                                                                                SS
                  lgo




                                                                        o




                                                                                                       N

                                                                                                            AN
                                                                    st r
                 Be




                                                                  Au




2.77      The graph shows NATS to have substantially higher economic costs per unit
output in 2003 than the average of the thirteen. Its position has improved since 2002
because of the fall in delays. Much of the difference between NATS and the average in
that year arose from high en-route ATFM delays, largely a transient consequence of the
move to Swanwick, which took place in January 2002. We recall that NATS “2002” data in
fact relates to the year from April 2002 to March 2003, whereas for other ANSPs it reflects
the calendar year.




8 Air Traffic Flow Management (ATFM) delay is the delay resulting from users not being able to achieve their desired
flight plans because of restrictions imposed by ANSPs on entries to their airspace, either en-route or that around airports.
This results from capacity constraints which may be temporary (because of weather or technical difficulties) or
permanent.




                                                             24
                                 NATS cost benchmarking: Final report on benchmarking with existing data


         The components of financial cost-effectiveness

2.78      ACE data allow us to break down the overall financial cost-effectiveness into a
series of ratios, which, when multiplied together, give us the overall financial cost per unit
output. As discussed in paragraphs 2.23 to 2.25, this breakdown allows us to examine what
factors contribute to the overall cost-effectiveness.

2.79     The first level that these ratios can be broken down to is to the three ratios shown
in Figure 2.1:

         •     the ATCO productivity;
         •     the unit ATCO employment cost; and
         •     the support cost ratio.

2.80      Note that these ratios have been defined to be intuitively easy to understand.
Some are therefore inverted from what would be obtained by strict decomposition of the
overall ratio. For example, the overall cost-effectiveness is expressed as a cost per unit
output – a higher value implies poorer performance. For two of the components: the unit
ATCO employment costs and the support cost ratio, the same is true. However, the third
ratio has been inverted, so that a higher value implied better performance – this is so that it
accords with the intuitive definition of productivity.

2.81      The ATCO productivity is the ratio of output – in the case of the gate-to-gate
analysis, composite flight-hours controlled – to the hours put in by ATCOs on operational
duty. The latter concept is defined in detail by the PRU in the Specification for
Information Disclosure, and includes all operational time, including supervisory duties, but
not non-operational duties such as planning or training. Because this ratio corresponds to
the normal concept of “productivity”, the PRU express it as output per unit input, as
opposed to the overall ratio which is expressed as inputs (cost) per unit output.

2.82      The unit ATCO employment cost is expressed as the ratio of the employment
costs of the ATCOs on operational duty to the hours spent on duty. Employment costs
include salaries and the associated on-costs such as employment taxes and pension
contributions.

2.83      The support cost ratio is the ratio of total ATM/CNS provision costs (as
defined in paragraph 2.16) to the employment cost of ATCOs on operational duty. This
ratio analyses the relationship between spending on ATCOs and overall spending. A
support cost ratio of 5 means that for every € spent on ATCOs, another €4 is spent on
other things.

2.84     Mathematically, the relationships shown in Figure 2.1 can be expressed as:

        Cost per composite flight-hour     =        ATCO unit employment costs

                                           ÷        ATCO productivity




                                                  25
                                              NATS cost benchmarking: Final report on benchmarking with existing data


                                                        ×         Support cost ratio

2.85           We first break overall cost per unit output down into these three ratios.

               ATCO productivity

2.86       In Figure 2.8 we show the ATCO productivity in 2001-2003. Productivity across
the thirteen fell in 20029. A proportion of this could be attributable to the fall in traffic
following the events of 11 September 2001; ATCO utilisation will diminish as a drop in
traffic is not matched by a corresponding drop in ATCO numbers. This is corroborated
by the small rise in 2003.

2.87     NATS’s ATCO productivity is well above the average for the thirteen in all the
years examined.

               Figure 2.8            ATCO productivity 2001-2003

                    1.8
                                                                                         2001
                    1.6                                                                  2002
                                                                                         2003
                    1.4
                                                                                         2001 average
                                                                                         2002 average
                    1.2
                                                                                         2003 average
    Flight-hours    1.0
         per
    ATCO-hour       0.8

                    0.6

                    0.4

                    0.2

                    0.0
                                                                                                               e
                           l




                                                       AV
                               na




                                                 A




                                                                                                                      C
                                                                              FS




                                                                                              S
                                         en




                                                                                      IR
                                                                         A




                                                                                                     L
                                                                          l
                        ro




                                                                        ro




                                                                                                            id
                                                                                            AT


                                                                                                    N




                                                                                                                   UA
                                                IA




                                                                       N
                               Ae


                                        d
                      nt




                                                                              D

                                                                                    IA
                                                                      nt




                                                                                                          gu
                                                     EN




                                                                                                  LV
                                                                     D
                                     we




                                                                                           N
                      co




                                                                    Co




                                                                                                                   M
                                                                                  AV




                                                                                                           y
                                                                                                        Sk
                                    SS
                   lgo




                                                                  o




                                                                                  N
                                AN




                                                              st r
               Be




                                                            Au




               ATCO unit employment costs

2.88      Figure 2.9 shows the ATCO unit employment costs for 2001-2003. For the
thirteen as a whole, unit employment costs rose slightly in 2002 (expressed in constant
2002 prices) and massively (by nearly 15%) in 2003. These changes are driven by some
very large changes in individual figures. Those for Belgocontrol rose in 2002 by nearly




9The case of ANS Sweden is anomalous; they used different reporting assumptions in ACE 2001 and ACE 2002. We
have not yet been able to correct for this effect.




                                                                26
                                        NATS cost benchmarking: Final report on benchmarking with existing data


40%, but fell again in 200310. There were also large rises in NAVIAIR and the IAA in
2002, whereas in Skyguide and NATS there were substantial falls. In 2003 two major
ANSPs (DFS and Aena) showed very large increases in ATCO employment costs, which
were in large part the cause of the overall increase. We have not yet been able to
investigate whether this effect is real or a reporting inconsistency. We understand that the
DFS rise may be caused by a “top-up” required in their pension fund, a factor that might
well be regarded as “exceptional”.

2.89      NATS ATCO employment costs per hour worked were close to the average of
the twelve. A significant drop in average ATCO costs between 2001 and 2002 (10%),
resulting from a small fall in total employment costs, arising from a number of senior
ATCOs leaving NATS and being replaced by less costly junior staff, and, much more
importantly, a substantial rise in the hours worked.

2.90    In 2003 the massive increases in Aena and DFS meant that NATS, in spite of a
14% increase in ATCO employment costs per hour, ended up with a unit cost appreciably
below the average of the thirteen.

            Figure 2.9           ATCO unit employment costs 2001-2003

             140                                                                             2001
                                                                                             2002
             120                                                                             2003
                                                                                             2001 average
                                                                                             2002 average
             100
                                                                                             2003 average

              80
   € per
 ATCO-hour
              60


              40


              20


               0
                             e




                                                            l




                                                                              AV
                           na




                                                                                                               A
                                    C

                                         FS




                                                                  S




                                                                                                       en
                                                                                       IR


                                                                                               A
                                                                        L
                                                           l

                                                         ro
                                                         ro
                         uid




                                                                AT


                                                                       N
                                 UA




                                                                                                             IA
                                                                                              N
               Ae




                                                                                                       d
                                                       nt
                                        D




                                                                                     IA
                                                      nt




                                                                            EN
                                                                      LV




                                                                                             D

                                                                                                    we
                      yg




                                                                N
                                                    co
                                                   Co
                                 M




                                                                                   AV
                   Sk




                                                                                                  SS
                                                 lgo
                                               o




                                                                                 N




                                                                                               AN
                                           st r


                                              Be
                                         Au




10 The PRU believes this to arise from differences in reporting from year to year, which they have been unable to resolve

with Belgocontrol.




                                                           27
                                  NATS cost benchmarking: Final report on benchmarking with existing data


         Support cost ratio

2.91     Figure 2.10 shows the support cost ratios across the twelve in 2001-2003. NATS
has among the highest support cost ratios in the sample (5.4). This implies that for every €
spent on ATCOs, €4.4 is spent on other things. The only higher support cost ratio is that
of LVNL. NATS’s support cost ratio, like the average for the thirteen, appears to be
declining over time.

         Figure 2.10         Support cost ratio, 2001-2003

         10
                                                                               2001
             9
                                                                               2002
             8                                                                 2003
                                                                               2001 average
             7
                                                                               2002 average
             6                                                                 2003 average

             5

             4

             3

             2

             1

             0
                                                            e
                                         AV




                                                                                        l




                                                                                               na
                                   A




                                                                   C
                                                FS
                        S




                                                                                      IR




                                                                                       n
                             A
                  L




                                                                                        l

                                                                                     ro
                                                                                     ro
                                                         id
                      AT




                                                                                    de
                 N




                                                                UA
                                  IA
                            N




                                                                                              Ae
                                               D




                                                                                   nt
                                                                                  IA


                                                                                  nt
                                                       gu
                                       EN
             LV




                            D




                                                                                 we
                      N




                                                                                co
                                                                                Co
                                                                M

                                                                               AV
                                                        y
                                                     Sk




                                                                              SS
                                                                             lgo
                                                                              o
                                                                    N




                                                                        AN
                                                                         st r


                                                                         Be
                                                                       Au




2.92     These “support costs” cover a wide range of items, and it is therefore worth
looking at the individual components in more detail. We do this in the following
paragraphs.

         Further analysis of support costs

2.93      Given their importance in the overall performance gap between NATS and the
twelve, it is worth examining the components of support costs in more detail. Support
costs, as defined in ACE, comprise:

         •        all operating costs other than the employment costs of ATCOs;
         •        exceptional items; and
         •        capital-related costs, comprising:
                  − depreciation; and
                  − the finance cost used in calculating the cost base for user charges.




                                                     28
                                         NATS cost benchmarking: Final report on benchmarking with existing data


2.94     ACE data permits this breakdown. It also permits the further breakdown of the
operating costs into “staff costs” – that is, employment costs – and other operating costs.
We have not chosen to explore this division further in our benchmarking as this ratio
depends on the extent to which different processes are outsourced and is therefore to some
extent arbitrary, rather than a measure of efficiency.

2.95     For further analysis of support costs, we do not use further multiplicative ratios,
because the individual components of support costs are additive. Instead we analyse and
compare the ratio of each individual component of support cost to the output.

               Non-ATCO operating costs

2.96   Figure 2.11 shows the non-ATCO operating costs (all operating costs – including
employment costs and exceptional items, but excluding depreciation - other than the
employment costs of ATCOs on operational duty) per unit output.

               Figure 2.11         Non-ATCO operating costs per unit output, 2001-2003

                 450

                 400                                                                       2001
                                                                                           2002
                 350                                                                       2003
                                                                                           2001 average
                 300                                                                       2002 average
                                                                                           2003 average
                 250
     € per
 flight-hour
                 200

                 150

                 100

                  50

                    0
                                                       e
                         l




                                           AV




                                                                                   na
                                                                             A




                                                                                                  C
                                                                FS
                                     S




                                                                                                          en
                                                                                          IR
                                                                      A
                               L




                                                                 l
                      ro




                                                               ro
                                                    id
                                   AT
                              N




                                                                                               UA
                                                                           IA
                                                                     N




                                                                                 Ae




                                                                                                           d
                    nt




                                                           D




                                                                                        IA
                                                  gu




                                                             nt
                                         EN
                             LV




                                                                     D




                                                                                                        we
                                   N
                   co




                                                           Co




                                                                                               M
                                                                                      AV
                                                   y
                                                Sk




                                                                                                    SS
                lgo




                                                                 o




                                                                                      N




                                                                                                   AN
                                                             st r
               Be




                                                           Au




2.97     In all the years considered NATS showed a substantially higher non-ATCO
operating cost per unit output than the average for the thirteen, with figures only exceeded
by Belgocontrol and LVNL. However, in 2002 and 2003 this difference was substantially
reduced.

2.98     Some of this difference, and the changes over the years, may concern the way that
the PRU collects data and assigns it to “support”. NATS uses Air Traffic Service
Assistants (ATSAs) in an operational role; but ACE counts their costs as “support”. This
could be responsible both for NATS’s higher than average ATCO productivity and for an




                                                               29
                                       NATS cost benchmarking: Final report on benchmarking with existing data


element of the higher support costs. The use of ATSAs was reduced with the move to
Swanwick. This issue is discussed further in paragraph 2.122 onwards.

2.99      There may also have been in 2001 and 2002 additional costs relating to the
transition to Swanwick, in the form of training, and an overburden of ATSAs and
maintenance staff, as well as the exceptional costs of relocation discussed below.

2.100 NATS non-ATCO operating costs showed a substantial reduction in 2002 and
2003. Part of this arises from a fall in exceptional items, but there are substantial falls even
after taking this into account. These falls are associated with a programme of cost
reduction instituted in 2002, and may also result from the operating regime at Swanwick
requiring fewer ATSAs.

2.101 Figure 2.12 shows the breakdown of non-ATCO operating costs between the
employment element and the non-staff element. This graph illustrates the diversity of
practice. On average, around 64% of these support operating costs are employment cost.
However, particular ANSPs have a tendency to have higher non-staff costs: ENAV, IAA
NATS and Aena all have proportions in the 40%s and 50%s. In ENAV and Aena we
understand that this is largely because of outsourced maintenance. In other cases it might
be because of rented rather than owned premises, other use of non-owned assets, or (in
NATS’s case) because of the substantial exceptional items. Others, by contrast, tend to use
own staff more. The figures for Austro Control and DFS are in the upper 70%s and 80%s.
The figure for Maastricht Upper Area Centre (MUAC) is also high, but this is a special case
because the balance of support activities is quite different and includes no CNS activity.
We have not been able to determine what the negative “exceptional costs” for Skyguide
comprise11.




11 They give rise to a difficulty in presentation in Figure 2.12. In the lower portion of that figure, the non-ATCO

employment costs are the sum of the pale blue bar an the yellow bar – with a negative item, this is the only way to make
the total correct.




                                                           30
                                              NATS cost benchmarking: Final report on benchmarking with existing data


           Figure 2.12                Breakdown of unit non-ATCO operating costs 2002

         450


         400                                                      Non-staff operating costs (less exceps)
                                                2002              Exceptional items
         350                                                      Non-ATCO employment costs

         300


         250


    € per 200
flight-hour
    2002 150


         100


          50


           0




                                                                                                                      IR
                                                                                                 A
                                                                                        A
                                                                 FS
                             L
                   l




                                                                              l




                                                                                                          na




                                                                                                                              AC



                                                                                                                                        en
                                      S


                                            AV



                                                      de
                 ro




                                                                            ro
                                   AT
                            N




                                                                                                IA
                                                                                     N




                                                                                                                    IA
                                                                                                          Ae




                                                                                                                                       ed
                                                      ui


                                                                D
               nt




                                                                         nt
          -50




                                                                                                                            U
                         LV




                                          EN




                                                                                    D
                                                    yg
                                  N




                                                                       Co




                                                                                                                                     Sw
             co




                                                                                                                 AV



                                                                                                                           M
                                                  Sk
          lgo




                                                                     o




                                                                                                                N




                                                                                                                                  S
                                                                                                                                AN
                                                                  str
        Be




                                                                 Au




         -100




                450

                                                                  Non-staff operating costs (less exceps)
                400
                                                                  Exceptional items
                                                                  Non-ATCO employment costs
                350


                300


                250


      € per   200
  flight-hour
      2003    150


                100


                 50


                  0
                                                                                                      A




                                                                                                                         IR
                                                                                            A
                                                                   FS
                          l


                                  L




                                                                                                                                   AC
                                                                                    l




                                                                                                               na
                                          S




                                                                                                                                            en
                                                 AV



                                                           de
                        ro




                                                                                  ro
                                       AT
                                 N




                                                                                                     IA
                                                                                          N




                                                                                                                       IA
                                                                                                               Ae




                                                                                                                                          ed
                                                          ui


                                                                  D
                      nt




                                                                               nt




                -50
                                                                                                                                 U
                              LV




                                               EN




                                                                                         D
                                                        yg
                                      N




                                                                             Co




                                                                                                                                        Sw
                    co




                                                                                                                     AV



                                                                                                                                M
                                                      Sk
                lgo




                                                                              o




                                                                                                                    N




                                                                                                                                       S
                                                                                                                                     AN
                                                                           str
             Be




                                                                      Au




             -100



           Capital-related costs

2.102 Figure 2.13 shows the capital-related costs per unit output for the thirteen in 2001
and 2002. NATS’s capital-related costs are well above the average for the thirteen (and
indeed for the whole of Europe). They showed a very significant rise in 2002, associated
with the start of operation at Swanwick.




                                                                      31
                               NATS cost benchmarking: Final report on benchmarking with existing data


2.103 Capital-related costs to a large extent represent past decisions and therefore are a
reflection of past performance. It would be helpful to examine where NATS’s business
and investment plan places it in future in relation to current performance of the twelve, but
this has not been possible in the time available.

          Figure 2.13     Capital-related costs per unit output, 2001-2003

           160
                                                                                 2001
           140                                                                   2002
                                                                                 2003
           120                                                                   2001 average
                                                                                 2002 average
           100                                                                   2003 average

     € per
             80
 flight-hour

            60

            40

            20

             0
                                               e

                                                     AV



                                                               l

                                                                   na




                                                                                                  A
                                                                                     C
                          FS
                S




                                                                                            en
                                                                            IR
                                 A
                      L




                                               l




                                                            ro
                                            ro

                                           uid
              AT


                     N




                                                                                  UA




                                                                                                 IA
                                N




                                                                   Ae




                                                                                             d
                                                          nt
                          D




                                                                          IA
                                         nt




                                                   EN
                    LV




                               D




                                                                                          we
                                        yg
             N




                                                          co
                                      Co




                                                                                 M
                                                                        AV
                                     Sk




                                                                                       SS
                                                       lgo
                                      o




                                                                        N




                                                                                     AN
                                  st r




                                                     Be
                                Au




2.104 Figure 2.14 shows how capital-related costs break down into the depreciation and
cost of capital (finance) elements. NATS is among the top two in both categories. Clearly
depreciation costs can vary substantially over the investment cycle, particularly if major
assets continue in use beyond their accounting life, and this may be an explanation of some
of the major differences visible. It is harder to understand the very small values for the
cost of capital for some ANSPs. However, the cost of capital for NATS is a market-related
value; this may not be true for all publicly owned ANSPs. NATS might therefore be
expected to have a relatively high cost of capital per unit output.




                                                32
                                NATS cost benchmarking: Final report on benchmarking with existing data


          Figure 2.14     Capital-related costs per unit output in 2003

           160

           140                                                               Depreciation
                                                                             Finance
           120

           100

     € per
             80
 flight-hour

            60

            40

            20

             0
                                                 e

                                              AV


                                                 l

                                                              na




                                                                                            A
                                                                            C
                          FS
                S




                                                                                    en
                                                                    IR
                                               A
                      L




                                                 l




                                              ro
                                              ro

                                             uid
              AT


                     N




                                                                         UA




                                                                                         IA
                                             N




                                                            Ae




                                                                                     d
                                            nt
                          D




                                                                   IA
                                           nt




                                        EN
                    LV




                                           D




                                                                                  we
                                          yg
             N




                                        co
                                        Co




                                                                         M
                                                                 AV
                                       Sk




                                                                              SS
                                     lgo
                                      o




                                                               N




                                                                             AN
                                 st r




                                 Be
                               Au




          Summary of NATS position relative to the twelve, gate-to-gate

2.105 The above analysis shows that NATS in 2001-2003 was a relatively high-cost
ANSP, although its costs were falling over the period towards the average, as those of
others were rising. In this section we summarise these differences comparing NATS to the
twelve others in the sample.

2.106 In 2003 costs per unit output were 17% higher, although this had fallen from a
32% gap in 2001. The picture was more complex, however, when one examined the
individual components in the PRU ACE framework. NATS has relatively high ATCO
productivity, and relatively low ATCO employment costs, but high support costs. Some of
this difference may be related to different practices in the use of Air Traffic Services
Assistants (ATSAs) in an operational role – this is discussed further below in paragraphs
2.122 to 2.126.

2.107 This picture is summarised in Figure 2.15. The left-hand bar represents the
average cost for the “twelve”. The successive elements of the chart, moving to the right,
show how NATS’s performance in each of the three major areas in the framework
contribute to the overall position, at the right of the chart. The chart measures cost, so a
high bar signifies relatively poor performance. In addition, we have coloured areas where
NATS’s performance indicators are worse than the twelve in red, and areas where they are
better in blue.

2.108 The first blue bar illustrates the impact of NATS’s high ATCO productivity,
which serves to reduce unit costs. The second blue bar shows the impact of NATS’s
relatively low ATCO employment costs, which serve further to reduce costs. The red bar




                                                 33
                                 NATS cost benchmarking: Final report on benchmarking with existing data


illustrates the impact of NATS’s very high support cost ratio. Note that these ratios are
multiplicative:

Impact of 18% higher productivity                     →          Cost index 82%

Impact of 6% lower ATCO costs                         →          Cost index 82% × (1-0.06) = 76%

Impact of 53% higher support cost ratio               →          Cost index 76% × 1.53 = 117%



         Figure 2.15       NATS cost per unit output compared with the twelve, 2003

                                                                                           Overall difference
  140%                                                                                     Costs 17% higher
         18% greater ATCO productivity     53% higher support cost ratio
  120%


  100%


  80%


  60%


  40%                      6% lower unit ATCO employment costs


  20%


   0%
             Twelve         ATCO productivity   Unit employment       Support cost ratio             NATS
                                                      costs



2.109 Given the importance of support costs, it is interesting to examine how their
composition has changed between 2001 and 2002. This is illustrated in Figure 2.16. This
figure shows the support costs per composite flight-hour, broken down into three
components:

         •    non-ATCO operating costs – all operating costs other than ATCO
              employment costs;
         •    exceptional items
         •    depreciation; and
         •    finance costs (cost of capital).

2.110 NATS’s costs in the period considered included a large amount of “exceptional
items”. These related principally to:




                                                    34
                                     NATS cost benchmarking: Final report on benchmarking with existing data


           (a)     the costs of restructuring and relocation as a consequence of the move to
                   Swanwick;
           (b)     write-offs of capitalised development costs associated with the planned new
                   Scottish ACC, and the costs of terminating the contract for the
                   development of the systems for that centre12; and
           (c)     the financing costs related to the “Composite Solution” for the refinancing
                   of NATS.

2.111 None of the other ANSPs in the sample showed any material “exceptional items”,
with the exception of Skyguide in 2003, which disclosed negative exceptional costs, the
nature of which has not yet been elucidated. We have noted, however, that some of the
large rise in DFS’s employment costs in 2003 might reasonably be regarded as
“exceptional”, as we understand that it arises from a retrospective “top-up” to their
pension fund.

2.112 The justification used by NATS in disclosing these items as “exceptional” was
that they were so declared in NATS’s accounts. They (and the PRU) attached some
importance to consistency with the accounts.

2.113 The breakdown of these exceptional items into these three categories is shown in
Table 2.1.

           Table 2.1           Breakdown of NATS’s exceptional items

                                                               2001          2002           2003
Reorganisation and relocation                                 10.97         12.61           3.10
Costs associated with new Scottish ACC                        59.99
Financial restructuring                                                     27.80          52.04
                 Total exceptional items                      70.96         40.41          55.14


2.114 All ANSPs are subject to occasional reorganisations with consequent relocations
and redundancies, as systems were replaced and centres were consolidated. Within the
twelve, Germany was also undergoing substantial reorganisation in the period. It was
therefore reasonable to include costs of type (a) in operating costs. Costs of types (b) and
(c) were exceptional in a wider sense – ANSPs not undergoing NATS’s changes in
corporate ownership would not expect to be subject to such costs.

2.115 We were not fully aware at this stage whether any others of the twelve had items
of cost corresponding to types (a), (b) or (c) that were not, possibly for reasons of domestic
accounting rules, counted as “exceptional”, although we were able to confirm that, for a




12These costs arose because the new owners of NATS under the Public-Private Partnership wished to change the plans
for the Scottish ACC from the previously planned bespoke system based on that at Swanwick to an “off-the-shelf”
system developed jointly with other ANSPs. This required the writing off of capitalised development costs, and a
payment to terminate the contract to develop the bespoke system.




                                                        35
                                    NATS cost benchmarking: Final report on benchmarking with existing data


number of the twelve, including DFS, IAA and ENAV, costs of type (a) were not regarded
as “exceptional”. For that reason, we included in the first instance all NATS’s exceptional
costs in the analysis. Nevertheless, we have exposed the different components in the
analysis below, as shown in Figure 2.16.

                 Figure 2.16    Non-ATCO costs per unit output, NATS and the twelve,
                                2001-2003


          500                                                                   Finance

                                                                                Depreciation

                                                                                Exceptional items (b) and (c)
          400
                                                                                Exceptional items (a)

                                                                                Non-ATCO op costs (less exceps)
          300


 € per
 flight   200
 -hour



          100



            0
                  Twelve-2001            NATS-2001     NATS-2002    NATS-2003                      Twelve-2003


          -100



2.116 The left-hand side of the figure shows the comparison in 2001, the right-hand
side in 2003. In the centre we see how NATS’s support costs have changed.

2.117 The overall picture is as follows. NATS’s support costs per unit output are
substantially greater than those in the twelve. The position improved in 2002. There was,
however, a major change in the balance of these support costs for NATS. NATS support
operating costs per unit output declined substantially in 2002 and 2003. Much of the
difference between NATS and the twelve, however, was a result of the exceptional items.
The residual operating costs, after the removal of exceptional costs of a financial, rather
than an operating nature, types (b) and (c), had fallen to a level similar to that of the twelve.

2.118 The average gate-to-gate operating costs per unit output (excluding ATCO
employment costs) for the twelve in 2003 were €221 per flight-hour. Those for NATS,
including all exceptional items, were €262, or around 18% higher. Removing exceptional
costs of types (b) and (c), however, reduces this to €212, 4% lower than the twelve.

2.119 The fall in unit support operating costs was partially offset by an increase in
depreciation and finance costs. NATS’s depreciation costs per unit output rose from
€57/flight-hour in 2001 to €84 in 2002 and to €88 in 2003. At the same time the average
for the twelve rose to €61. In 2003, then, the NATS depreciation per unit output was 44%




                                                     36
                                            NATS cost benchmarking: Final report on benchmarking with existing data


higher than that of the twelve. The increase in NATS’s depreciation costs per unit output
is associated with the entry into operation of Swanwick in January 2002 (just before the
close of the year labelled “2001”).

2.120 NATS’s finance costs per unit output also rose sharply in 2002, followed by a fall
in 2003. NATS’s finance costs are substantially higher than the average for the twelve -
€48 compared to €17, and much higher even than the next highest, DFS at €35.

2.121 Figure 2.17 extends this comparison to include all costs, adding the ATCO
employment costs. Because of higher ATCO productivity and lower unit employment
costs (as discussed above), NATS has lower ATCO employment costs per unit output.
The graph shows that NATS unit operating costs, including all elements of operating costs
and excluding exceptionals of types (b) and (c), are around 11% lower than the average of
the twelve, but that higher than average capital-related costs raise its total costs well above
the average.



               Figure 2.17          Total costs per unit output, NATS and the twelve, 2001-3

                  Finance
         600
                  Depreciation

                  Exceptional items (b) and (c)
         500      Exceptional items (a)

                  Non-ATCO op costs (less exceps)
                  ATCO costs
         400


         300
€ per
flight
-hour
         200


         100



           0
                Twelve-2001                         NATS-2001    NATS-2002   NATS-2003                Twelve-2003

     -100



               The issue of Air Traffic Service Assistants (ATSAs)

2.122 Two major features of NATS comparative performance might at least in part be
explained by consideration of the operations room practices at the former LATCC and
Swanwick, compared to those at typical European centres.

2.123 NATS’s practice at LATCC and its other centres was to staff operating positions
with a mixture of qualified ATCOs and Air Traffic Service Assistants (ATSAs) qualified to




                                                                37
                                 NATS cost benchmarking: Final report on benchmarking with existing data


a less rigorous standard. This practice is unusual in the European context. Figure 2.18
shows the proportion of ATSAs or equivalent staff (described by the PRU as “ATC
assistants” and defined in the Specification for Information Disclosure) to the number of
ATCOs on operational duty.

2.124 There is clearly a wide disparity in practice. Of other members of the twelve, only
ENAV even comes close to NATS in its use of ATSAs.

2.125 It could be argued that the use of ATSAs in an operational role might increase the
productivity of the ATCOs. Since, however, by the PRU’s definitions, ATSAs and their
equivalents elsewhere are counted as “support”, it would increase the support cost per unit
output. This is consistent with the apparent high productivity of NATS ATCOs,
combined with their high support costs.

2.126 However, we note that the only other member of the twelve with comparable
numbers of ATSAs, ENAV, has neither particularly high ATCO productivity nor
particularly high support costs. Exploring this issue with selected ANSPs suggests that in
practice, the equivalents of ATSAs in a number of organisations (DFS, ENAV, MUAC and
IAA) perform a similar mix of operational and non-operational roles.

         Figure 2.18     ATSAs as a proportion of operational ATCOs, 2001-2003

         90%
                                                                  2001
         80%                                                      2002
                                                                  2003
         70%
                                                                  2001 average
         60%                                                      2002 average
                                                                  2003 average
         50%

         40%

         30%

         20%

         10%

          0%
                                                                                                      e
                             l




                                                AV
                           na




                                                        A




                                                                      C
                                   FS




                                                                                    S
                                                                             en




                                                                                          IR
                                          A




                                                              L
                            l

                          ro
                          ro




                                                                                                   id
                                                                                  AT
                                                             N

                                                                   UA
                                                       IA
                                         N
                       Ae




                                                                              d
                        nt


                                  D




                                                                                        IA
                       nt




                                                                                                 gu
                                              EN




                                                            LV
                                        D




                                                                           we


                                                                                  N
                     co
                     Co




                                                                  M




                                                                                      AV


                                                                                                  y
                                                                                               Sk
                                                                       SS
                  lgo
                   o




                                                                                      N
                                                                      AN
              st r


              Be
            Au




En-route
2.127 The previous section compared NATS’s performance in various components of
gate-to-gate cost-effectiveness with those of twelve other broadly similar European
ANSPs. This analysis is informative because it compares businesses with a similar scope of
activities, and ties in with the analysis already done by the PRU in its ACE reports. More




                                                  38
                               NATS cost benchmarking: Final report on benchmarking with existing data


analysis is possible at the gate-to-gate level, since not every piece of information is split into
en-route and terminal.

2.128 However, a gate-to-gate comparison is less than ideal for the purposes of
informing the Price Control Review, because it relates to a combination of businesses,
some of which are price controlled and some of which are not, as described in paragraphs
1.14 to 1.19. Ideally, we would wish to compare businesses that are similar in scope to
NATS’s en-route business. However, this introduces a new source of non-comparability.
While all Eurocontrol member states declare costs of their “en-route” businesses both to
the CRCO and the PRU, the specification of what should be included in “en-route” is
neither precise in its specification of how costs should be divided and allocated, nor
universally followed. Comparing “en-route” businesses introduces an extra source of
uncertainty – that produced by differences in approaches to allocation across ANSPs.

2.129 Furthermore, the operations of NERL are not all subject to price control. As
discussed in Chapter 1 and illustrated in Figure 1.1, NERL business includes the London
Terminal Approach business, which is not subject to price control. For the purposes of
the ACE analysis, we understand that the costs allocated to the London Terminal
Approach business have been excluded from “en-route”.

2.130 This gives rise to a further difficulty concerning the definition of output. Flight-
hours in the ACE data set have not been split into en-route and terminal; only a total is
given. We have made the assumption that the flight-hours associated with terminal
navigation services form a small and comparable proportion of the total, and that the
comparison will not be distorted by the use of total flight-hours as a measure of en-route
output.

2.131 Bearing these potential difficulties in mind, we have repeated the analysis above
for ANSPs’ en-route businesses. Because much of the story is broadly similar, not all the
analysis is reproduced, and the sections concerning en-route should be read in conjunction
with those on gate-to-gate.

         Financial cost-effectiveness

2.132 Figure 2.19 shows the financial cost per en route flight-hour for 2001-2003. The
overall picture for NATS is similar to that for gate-to-gate. NATS is among the most
costly of the twelve, and the only others with comparable unit costs have their costs
overstated because of bearing the costs of CNS infrastructure shared with Maastricht. The
costs have shown a marked decline over the period. This is a consequence of the cost
reductions in 2002 and 2003, particularly those in exceptional items, being concentrated on
the en-route business.




                                                39
                                         NATS cost benchmarking: Final report on benchmarking with existing data


               Figure 2.19         Financial cost per en-route flight-hour, 2001-2003

                 900
                                                                                             2001
                 800
                                                                                             2002
                 700                                                                         2003
                                                                                             2001 average
                 600                                                                         2002 average
                                                                                             2003 average
                 500
     € per
 flight-hour
                 400

                 300

                 200

                 100

                    0




                                                                               e
                         l




                                                                   AV
                                                           na




                                                                                         A




                                                                                                                    C
                                            FS
                               S




                                                                                                            en
                                                                                                 IR
                                                                                    A
                                     L




                                                       l
                      ro




                                                     ro




                                                                            id
                             AT


                                    N




                                                                                                                 UA
                                                                                        IA
                                                                                   N
                                                           Ae




                                                                                                          d
                    nt




                                          D




                                                                                               IA
                                                   nt




                                                                          gu
                                                                 EN
                                   LV




                                                                                   D




                                                                                                       we
                             N
                   co




                                                 Co




                                                                                                                 M
                                                                                             AV
                                                                           y
                                                                        Sk




                                                                                                    SS
                lgo




                                                 o




                                                                                             N

                                                                                                  AN
                                             st r
               Be




                                           Au




               Breakdown of financial cost-effectiveness

2.133 It is not possible to break down rigorously the financial cost-effectiveness for en-
route in the same manner as that for gate-to-gate, using the ACE data, because Information
Disclosure does not require the data concerning the employment costs of operational
ATCOs to be split into en-route and terminal. With the data available, the most useful split
is that between operating costs and the capital-related components of cost.

2.134 Figure 2.20 shows the comparison between the en-route costs of NATS and the
twelve in 2001 and 2003. Note that unlike in Figure 2.16 above, the operating costs include
the costs of operational ATCOs. The broad picture is very similar to that for gate-to-gate.
Operating costs for NATS are higher than the average, but fall in 2002 because of NATS’s
cost-cutting. In 2003 once truly exceptional items (types (b) and (c)) have been eliminated,
the en-route operating cost of NATS per unit output is very close to that of the twelve.
Note, however, that we understand that a large element of DFS’s costs in 2003 represented
a top-up of the pension fund, which might well be regarded as “exceptional” even though
it was not described as such in their PRU submission. Because of the size of DFS, this
would have an appreciable impact on the average.

2.135 Capital-related costs are substantially higher than in the twelve, and depreciation
costs show a large increase in 2002, related to the start of operation at Swanwick.




                                                            40
                                       NATS cost benchmarking: Final report on benchmarking with existing data


                  Figure 2.20      En-route unit costs (including ATCO costs), 2001-2003


           700
                                                                                      Finance
                                                                                      Depreciation
           600                                                                        Exceptional items (b) & (c)
                                                                                      Exceptional items (a)
                                                                                      Operating costs (less exceptionals)
           500


           400

     € per
     flight 300
     -hour

           200


           100


             0
                   Twelve - 2001            NATS - 2001    NATS - 2002    NATS - 2003                    Twelve - 2003
          -100



Sensitivity of the comparison to exchange rates
2.136 The comparison of performance between NATS and the twelve is sensitive to the
exchange rate used in the comparison. Nine of the twelve, comprising the vast majority of
the activity, use the euro. The relevant exchange rate is therefore that between the pound
and the euro.

2.137 Between 2001 and 2002 the exchange rate used by the PRU changed from €1.609
to the pound in 2001 to €1.558 in 200213. This change made a difference of 3% to the
results – had there been no exchange rate changes, NATS’s performance would have
appeared 3% worse relative to the nine €-zone ANSPs.

2.138 The change between 2002 and 2003 was more marked. The exchange rate used in
2003 was €1.439 to the pound, a fall in the pound of nearly 8%. For comparability with
our Interim Report, we have used the 2002 exchange rate throughout the analysis. It is
worth noting, however, that if 2003 exchange rates had been used throughout, the ranking
of NATS in terms of indicators with a financial component would have been more
favourable for all years examined. For example, the costs per unit output for NATS in




13According to the Specification for Information disclosure these should have been the average September rates, as used
by the CRCO. NATS followed this approach for 2001. For 2002 and 2003, however, NATS used an exchange rate
corresponding to the average for the year in question (April to March), which in our view gives a more representative
picture.




                                                          41
                                          NATS cost benchmarking: Final report on benchmarking with existing data


2003, as shown in Figure 2.4, would fall below those of Aena and DFS. This is illustrated
in Figure 2.21.

                   Figure 2.21        Financial cost-effectiveness indicator for 2003 (2003 prices)

               700


               600                                                           Average for thirteen


               500


     € per     400
     flight
     hour      300


               200


               100


                   0
                                                                      e
                        l




                                                        AV
                                          na




                                                                                              A




                                                                                                                        C
                                                 FS
                                    S




                                                                                                                en
                                                                                                      IR
                              L




                                                                                     A
                                                                                l
                     ro




                                                                              ro
                                                                  uid
                                  AT
                             N




                                                                                                                     UA
                                                                                            IA
                                                                                    N
                                         Ae




                                                                                                               d
                   nt




                                                D




                                                                                                      IA
                                                                            nt
                                                      EN
                            LV




                                                                                    D




                                                                                                            we
                                                               yg
                                  N
                 co




                                                                          Co




                                                                                                                     M
                                                                                                    AV
                                                             Sk




                                                                                                           SS
              lgo




                                                                        o




                                                                                                 N

                                                                                                       AN
                                                                    st r
              Be




                                                                  Au




2.139 In practice, exchange rates are subject to considerable volatility. Figure 2.22
shows the daily exchange rates, €/£ for the period 1 January 2001 to 31 March 200414.
The exchange rates used in the ACE data are also shown for comparison.




14 These exchange rates were obtained in a convenient electronic form from the website of Pacific Exchange Rate Service
(fx.sauder.ubc.ca) and are used with their permission




                                                             42
                                                   NATS cost benchmarking: Final report on benchmarking with existing data


              Figure 2.22                 Daily exchange rates in 2002 and 2003

        1.7



       1.65



        1.6



       1.55

 €/£

        1.5



       1.45



        1.4



       1.35
            1 1 1 1 1 1 1 1 1 2 2 2 2 2 2 2 2 2 2 2 2 3 3 3 3 3 3 3 3 3 3 3 3 4 4 4
         r 0 y 0 n 0 ul 0 g 0 p 0 ct 0 v 0 c 0 n 0 b 0 ar 0 r 0 y 0 n 0 ul 0 g 0 p 0 ct 0 v 0 c 0 n 0 b 0 ar 0 r 0 y 0 n 0 ul 0 g 0 p 0 ct 0 v 0 c 0 n 0 b 0 ar 0
       Ap Ma Ju J Au Se O No De Ja Fe M Ap Ma Ju J Au Se O No De Ja Fe M Ap Ma Ju J Au Se O No De Ja Fe M



2.140 Sampling the exchange rate at different points in the year could give substantial
differences. The exchange rate in the year to 31 March 2002 (the most volatile of the years
considered) varied between 1.45 and 1.64. This means that NATS unit costs in euros for
2002 could have been calculated as anywhere between 8% lower, if the lowest figure had
been used, and 5% higher, if the highest figure had been used.

2.141 Volatility of exchange rates therefore can have a major impact on the comparison.
Taking the highest possible value of the pound results in a substantially greater
performance gap. Even taking the lowest possible value, however, still leaves an
appreciable overall performance gap.

Purchasing Power Parity
2.142 One possible objective way to determine how to compare costs between different
countries, and at the same time correct for exchange rate volatility, is to use Purchasing
Power Parities (PPPs). PPP exchange rates are intended to represent the relative
purchasing power of one currency unit in different locations. Within the €-zone different
countries can have different PPP exchange rates. For example, €100 will buy a bigger
basket of goods in Spain than in Germany or Ireland.

2.143 We obtained PPP indicators from Eurostat for the years 2001 and 2002. They
show, for example, that a euro in Spain had 20% more purchasing power than the average
of the €-zone, and in Germany it had 6% less. Using the exchange rates used by the PRU,
we found also that the purchasing power of the euro in the UK was 6% less, and that in
Switzerland 21% less.




                                                                                43
                                       NATS cost benchmarking: Final report on benchmarking with existing data


2.144 It was therefore informative to look at the comparison of unit costs in terms of
their purchasing power. We have looked only at the financial unit costs of the gate-to-gate
business.

2.145 Figure 2.23 shows the results of the PPP adjustment. Adjusting for PPP makes
major differences for some ANSPs, notably Aena and ENAV, whose apparent
performance is substantially worsened, and Skyguide, whose apparent performance is
greatly improved. However, the overall qualitative assessment of NATS’s position is not
greatly changed. It is still a relatively costly ANSP. Among the twelve the only more costly
ANSPs are LVNL and Belgocontrol, whose costs are overstated because they bear CNS
infrastructure costs shared with Maastricht, and Aena and ENAV. Its costs are still well
above the average of the twelve.

            Figure 2.23            Unit costs comparison corrected for purchasing power

            700


            600

                                                                                         2002
            500                                                                          2002 adjusted for PPP


  € per     400
  flight
  -hour     300


            200


            100


                0
                                                                       e
                     l




                                               AV


                                                                     na




                                                                                                    A


                                                                                                            C
                                       FS
                                 S




                                                                                    en


                                                                                            IR
                                                                            A
                           L




                                                  l
                  ro




                                                ro




                                                                   uid
                               AT
                          N




                                                                                                         UA
                                                                                                  IA
                                                                           N
                                                          Ae




                                                                                   d
                nt




                                       D




                                                                                           IA
                                              nt

                                            EN
                         LV




                                                                           D


                                                                                we
                                                                yg
                               N
              co




                                            Co




                                                                                                        M
                                                                                         AV
                                                             Sk




                                                                               SS
           lgo




                                             o




                                                                                       N
                                                                           AN
                                         st r
           Be




                                       Au




2.146 In the light of these results, which show that no fundamental changes to the
comparison result from the use of PPP exchange rates, we have not updated the PPP
analysis for 2003.




                                                        44
                          NATS cost benchmarking: Final report on benchmarking with existing data



Glossary
ACC              Area Control Centre
ACE              Air Traffic Management Cost-Effectiveness
Aena             Aeropuertos Españoles y Navegación Aérea, Spain
AIS              Aeronautical Information Services
ANS              Air Navigation Services
ANS Sweden       ANS department of Swedish Civil Aviation Administration (LFV)
ANSP             Air Navigation Service Provider
APP              Approach Control Unit
ATC              Air Traffic Control
ATCO             Air Traffic Control Officer
ATFM             Air Traffic Flow Management
ATM              Air Traffic Management
Austro Control   Austro Control Österreichische Gesellschaft für Zivilluftfahrt mbH, Austria
Belgocontrol     Belgocontrol, Belgium
CEATS            Central European Air Traffic System
CFMU             Central Flow Management Unit
CNS              Communications, Navigation and Surveillance
CRCO             Central Route Charges Office
DFS              Deutsche Flugsicherung GmbH, Germany
DNA              Direction de la Navigation Aérienne, France
ENAV             Ente Nazionale di Assistenza al Volo S.p.A., Italy
EU               European Union
FTE              Full-Time Equivalent
IAA              Irish Aviation Authority
ICAO             International Civil Aviation Organisation
IFR              Instrument Flight Rules
LVNL             Luchtverkeersleiding Nederland, Netherlands
MET              Aeronautical Meteorology
MUAC             Maastricht Upper Air Centre
NATS             National Air Traffic Services, UK
NAVIAIR          Air Navigation Services – Flyvesikringstjenesten, Denmark
NERL             NATS En Route Limited
PPP              Purchasing Power Parity
PRC              Performance Review Commission
PRU              Performance Review Unit
SAR              Search and Rescue
Skyguide         Skyguide, Switzerland
TCC              (London) Terminal Control Centre
TMA              Terminal Manoeuvring Area
TWR              Traffic Control Tower
UK CAA           United Kingdom Civil Aviation Authority
VFR              Visual Flight Rules




                                           45

				
DOCUMENT INFO
Shared By:
Categories:
Stats:
views:8
posted:3/11/2010
language:English
pages:49
Description: Cost benchmarking NATS relative to selected ANSPs