Proposals on the Development of the LRIC Model for Identification in Japan Presentation to the LRIC Working Group by Cable

1
Proposals on the Development of the LRIC Model
for Identification in Japan Presentation
to the LRIC Working Group by Cable
Wireless IDCJonathan Sandbach, Ph.DDirector,
Economic Regulatory AnalysisCable Wireless
Global
1
2
Contents

• Methodologies for distinction of traffic
  sensitive vs non-traffic sensitive costs
• Estimating the cost of Universal Service
• Asset lives
• Unbundled copper fibre
• Forward looking charges

2
3
Methodologies for Distinction of Traffic
Sensitive vs Non-Traffic Sensitive Costs
3
4
Principles used on other countries
Hong Kong The charge will be based on LRAIC
Long Run Average Incremental Cost, reflecting
the use of the traffic dependent portion of the
terminating carriers network. UK The exchange
local switch performs a dual function of
customer access to the network and the handling
of different types of calls. The cost drivers
for expenditure are those factors responsible for
the levels of provision of specific elements of
equipment within the exchange. The main cost
drivers are connections of access lines,
traffic and call attempts or a mix of these. The
costs of equipment items whose provision levels
are dictated by customer connections are clearly
those associated with the function of providing
access to the network, and are thus driven by
exchange lines. This proportion of the primary
plant group costs are therefore apportioned to
Access activities. The costs of equipment
provided to satisfy the traffic and call attempts
required are clearly driven by the number and
duration of those call types contributing to
those activities.
4
5
Principles used on other countries

• US
• Element rates shall be structured consistently
  with the manner in which the costs of providing
  the elements are incurred.
• Australia
• Cost classification For the purposes of
  calculating costs, Telstras PSTN is divided into
  two components
• the customer access network, which is the
  network connecting each end user to a node (e.g.
  an IRIM, RSS, RSU). It consists of
  infrastructure such as network termination
  points, copper lines, trenches, pillars, and line
  cards and
• the inter-exchange network, which is the
  network connecting nodes. It consists of
  infrastructure such as optical fibre, trenches,
  multiplexers, remote concentrators and switches.
• Broadly, the costs of the customer access network
  are regarded as line related costs, whereas the
  costs of the inter-exchange network are regarded
  as conveyance costs (i.e. call related costs).

5
6
Why align the model with cost causation?

• This alignment of cost causation with revenue
  recovery is one of the underlying principles of
  an economically efficient interconnection charge
  structure. Restricting interconnection charges
  to traffic related costs is necessary to
  encourage
• efficient building of alternative access
  infrastructures. Carriers should only make these
  investments where it constitutes an efficient use
  of investment resources. The level of
  interconnection charges will determine when new
  carriers decide to invest in new local loop
  infrastructure, and when to purchase
  interconnection from the incumbent. If
  interconnection charges include non-traffic
  sensitive costs, new carriers will be forced into
  making sub-optimal decisions, opting to build
  their own infrastructures when, in fact, it would
  be more efficient to use the incumbents network
  through interconnection
• efficient use of the long distance and
  international networks when prices reflect costs.
  Whenever prices exceed the underlying resource
  costs, efficient usage of the network by
  interconnect customers (and ultimately retail
  customers) is curtailed. If interconnection
  charges (either those charged by NTT-East or
  NTT-West to new carriers, or those paid by the
  NTT group to itself) exceed traffic sensitive
  costs of long distance or international usage,
  use of these services will be curtailed to the
  detriment of consumer welfare.

6
7
Analysing network components for traffic
sensitive non-traffic sensitive costs
Local Exchange
Dimensioned by traffic (traffic sensitive)
Remote switching / concentrator units (may be
collocated with LE)
Subscriber line cards
Cable junctions/ distribution points
Dimensioned by access lines (non-traffic
sensitive)
7
8
Revenue recovery of traffic sensitive and
non-traffic sensitive costs
Total Network Cost
Incremental to access Common costs
Incremental to conveyance
Allocated to Allocated to
Allocated to Allocated to leased leased
lines etc exchange lines Common costs
PSTN lines, data, etc.
Allocated to Allocated to
Al located to Allocated to leased
leased lines etc exchange lines
PSTN lines, data services, etc
Allocated to Allocated to
Allocated to Allocated to
Allocated to leased leased lines etc
exchange lines retail
PSTN interconnect lines, data services, etc
Exchange line rental
Interconnect charges
Call Charges
8
9
Estimating the cost of Universal Service
9
10
Cost of Universal Service

• Cost of universal service
• avoidable net cost of unprofitable
  subscribers
• avoidable net cost of unprofitable
  exchanges
• - benefits to the universal service
  provider
• Note avoidable costs will exclude
• many retail costs (including sales)
• most overheads
• all RD costs

10
11
Avoidable net cost of unprofitable subscribers
Analysis for each Message Area
Iterations Adjustment for loss of calls to/from
unprofitable subscribers Ri R i-1 l i / l
i-1 where Ri is subscriber revenue (from
distribution) at iteration i l i is number of
lines of lines not defined as USO at iteration
i Iterations stop when l i l i-1
Distribution for each Message Area of
Subscriber Bills Incoming Call Revenue At
Iteration 0, this will be based on actual
subscriber data (or samples thereof) supplied by
NTT/E NTT/W
illustrative
Iteration 1
Iteration 2
Iteration 3
USO
Yen
Subscriber pure Incremental costs (excluding
common costs) for Message Area SourceLRIC
Model Note calculation of this distribution
requires the LRIC model to analyse traffic
sensitive non-traffic sensitive costs
11
12
illustrative
12
13
13
14
14
15
Asset Lives
15
16
International benchmarks

• PwC has conducted an analysis of asset lives for
  over 30 telecommunication operators globally, 20
  of which are in the top 50 (based on company
  revenues)
• The following table summarises asset lives used
  for different types of equipment

16
17
International benchmarks
1
17
18
International benchmarks switching
18
19
International benchmarks transmission equipment
Years
19
20
International benchmarks cable

• The life attributed to cabling will mainly depend
  on the type of technology and cable used (copper
  v. fibre optic). The life of aerial cable is
  shorter due to likely damage and technical
  obsolescence.
• The following chart shows the number of companies
  using various asset lives for underground copper
  and fibre cables.

20
21
International benchmarks duct
Low asset lives for duct is sometimes a result of
extensive competition in local infrastructure
provision. Therefore, we would expect appropriate
asset lives for NTT to be at top end of this
range.
21
22
International benchmarks
Benchmark survey performed in June 1996 on plant
asset lives Asset Average Minimum Maximum Standar
d Deviation Conduit/duct 51.56 45.00 60.00 4
.81 Source INDETEC The results compare
to an estimate of 56 years used in the HAI model
(another of the major US cost models).
22
23
International benchmarks power equipment

• Research indicates that 10 years appears to be a
  reasonable figure for the average in the industry
  with regards to DC power equipment.
• In Europe and the US Telcos have adopted service
  lives for DC power in the range of 10 to 12
  years, and in Asia 9 to 15 years.

23
24
Main drivers of economic depreciation
Electronics (e.g. switching transmission) t
echnical obsolescence Duct cable ... physical
life
24
25
25
26

Long asset life
Shorter asset life
26
27
Possible Model (1)
Let P be price of asset at year 0
p be annual price increase of asset
M be annual operational cost of asset in 1st
year, as of asset price m be annual
increase in operational cost in subsequent years
of assets life r be discount rate
d be asset life N be net present
cost of infinite renewal of asset after d years


N P



Provided p lt m m ltgt r
27
28
Possible Model (2)
Maximise N with respect to d First order
condition is
Provided p lt r m ltgt r
Can be solved for d
28
29
Example (1) low maintenance equipment
Assume 8 discount rate Assume 0.5
operational cost as of asset price in year 1 of
asset life rising to 1.0-1.5 after r years
duct
Annual price increase of asset -5 -2 0 1
optical fibre cable
18.6 20.9 22.9 24.0 13.6 14.7 15.5
15.9
Annual increase 15 in opex in each year of
asset life 25
Assume 1 operational cost as of asset price
in year 1 of asset life rising to 2-3 after 5
years
Annual price increase of asset -5 -2 0 1
15.5 17.4 19.0 19.9 11.5 12.4 13.1
13.4
Annual increase 15 in opex in each year of
asset life 25
29
30
Example (2) high maintenance cost
Assume 8 discount rate Assume 2 operational
cost as of asset price in year 1 of asset
life rising to 4-6 after 5 years
Annual price increase of asset -5 -2 0 1
12.8 14.3 15.5 16.2 9.6 10.3 10.8
11.1
Annual increase 15 in opex in each year of
asset life 25
Assume 5 operational cost as of asset price
in year 1 of asset life rising to 10-15 after 5
years
Electronic equipment
Annual price increase of asset -5 -2 0 1
9.5 10.6 11.4 12.0 7.3 7.8 8.2
8.4
Annual increase 15 in opex in each year of
asset life 25
30
31
Asset life data way forward
LRIC Working Group should formerly survey
overseas regulatory authorities to collect
up-to-date estimates of asset lives.
CW IDC would be glad to do this on behalf of the
WG.
31
32
Unbundled Copper Fibre
32
33
LRIC for unbundled copper fibre

• Important points
• duct sharing
• optimal cable sizes
• sharing of common costs with other services
  (e.g.PSTN)
• cable asset life
• exclude subscriber line card (not required for
  unbundled loops)

Example of unbundled copper model see
www.analysys.com
33
34
Forward looking charges
If the model is to be used to set charges going
forward in time it is essential that it reflect
projections of - equipment costs - network
usage volumes This last element (usage volumes)
is important to capture the correct economies of
scale. In other countries, one of the most
important drivers of local network usage growth
is Internet access traffic. An example is given
on the next slide.
Volume sensitive costs
Total conveyance costs
Fixed costs
now
Internet access
Volumes
PSTN
now
Unit costs
now
34
35
Internet access volume data

Internet access (0845) traffic growth on PSTN
This annex provides an illustration of the
growth experienced in 0845 traffic. Since the
introduction of the NTS formula and the
wholesale adoption of the Virtual Point of
Presence (VPOP) model by the industry for
Internet access this may be used a a
quantification of Internet access traffic since
it will be predominantly such. Figure C1 is a
direct plot of the traffic, measured in minutes
for each quarter over the period Q1 1994/1995
to Q3 to 1999/2000. This gives a graphic
indication of the explosive nature of the rate of
growth.
Source OFTEL
35