Further Memorandum submitted by the Director
General of Electricity Supply for Northern Ireland
In theory Capex should be finite, certainly
for transmission investment. There are a specific number of pieces
of work, all of which should be identifiable, to be undertaken
during a regulatory period. Capex planning should be a relatively
exact science with a list of evaluated projects assembled for
consideration before the next price control period.
If it worked like this, then if all the money
were not spent in the first period the explanations would be:
(a) A lower cost solution has been found.
This would mean the same quantum of projects was carried out and
that the target quality of service would have been achieved. This
would be a real efficiency gain. The company would benefit from
the underspend for five years and customers would benefit for
the next 35 years through avoiding the financing and depreciation
charge as NIE's asset base would now be lower than otherwise due
to the degree to which lower cost solutions were found compared
to what originally was anticipated would be spent.
(b) The project was deferred into a later
period. This means that the company benefits from the underspend
in the first period but customers at some stage in the future
will have to finance the project. If the project is merely deferred
into the next regulatory period, the effect is to require customers
to pay for it twice, once in the second period and also in the
first period where they will have already paid up to 30 per cent
of its financing cost if the underspend is in the first year.
The company, in this case, is rewarded twice because it keeps
the first period underspend and earns a return on the project
from the second period once it is constructed.
In practice, Capex is not a list of projects
agreed in advance. It is a sum of money derived from an iterative
bargaining process. It is therefore not possible to know by any
of the usual tests if an underspend represents an efficiency gain.
Some underspend may well do so, and until the last NIE price control
review and subsequent MMC inquiry the convention had been to treat
all underspend as efficiency gains. I believe a greater onus of
proof should be placed on the company to show that underspend
represents an efficiency gain as the greater probability is that
it represents deferral. The MMC assumed that two-thirds of the
underspend represented efficiency gain.
The foregoing should set the answers to your
questions in context. If the actual 93-97 Capex expenditure had
been accurately foreseen, customers would have saved £5.13
million per annum (96-97 prices) for five years. This would have
reduced the average domestic bill by £2.50 per annum. If
the expenditure had proceeded up to the allowed level then the
customer would have had to spend £8.68 million per annum
financing that expenditure in the second period and would be incurring
the financing bill on this expenditure until 2032-37. However,
in that case customers would have an additional £97 million
of network investment in place and logically the second period's
Capex would have been correspondingly smaller.
The ex ante imprecision or black box nature
of Capex is in fact rather neatly illustrated by the way in which
NIE proposes accelerating the refurbishment of the rural network.
This does not involve any additional cost to customers in the
present period and the money will be found by deferring work.
I understand some large sub-stations upgrading or replacement
will be postponed. This deferral should have no adverse effect
on customer minutes lost (CMLs). It is therefore evident that
the sub-stations' work did not, strictly speaking, have to occur
in the present regulatory period. Strictly speaking, perhaps this
work should not have been there at all.
Ideally, Capex should be a rigorously assessed
minimal list of necessary projects and all underspend should be
efficiency gains. In practice it tends to be an affordable black
box in which underspends are mostly deferrals, though with some
element of efficiency gain which is difficult to quantify. It
is this "black box" nature of Capex which probably provides
the reason why the situation described in your second question
of a Capex overspend has never arisen. If a company wanted to
exceed its Capex it would be entitled to ask for a re-opening
of its price control or at least ask that the additional assets
it financed out of its own money would be included in the regulatory
asset base and it would be foolish to proceed with any overspend
without the assurances it needed. But for the reasons given abovethe
benefits to companies of underspending for whatever reason the
underspend arisesI am not aware of any company overspending.
Finally, while I hope I have answered your questions
about Capex effectively, it would be remiss of me to give the
impression that I believe the Boxing Day storm would have been
significantly less damaging if NIE had spent all their Capex in
the first price control period. I am still engaged in analysing
with the company the precise relationship between investment and
vulnerability to storm damage. While the reduced probability of
failure in new components, wire and poles is not in dispute, it
is difficult to establish how much better the network would have
fared in the storms if more money had been spent in an earlier
period to strengthen the rural network.
22 April 1999