Incentive Regulation for Recurrent Expenditure: Benchmarking and CarryOvers

1
Incentive Regulation for Recurrent
ExpenditureBenchmarking and Carry-Overs

• Darryl Biggar
• Consulting Economist
• Australian Competition and Consumer Commission
• 2 April 2004

2
What is the issue?

• In updating the Statement of Regulatory
  Principles two questions naturally arise
• Should the ACCC make more use of external
  benchmarking in setting the expenditure targets?
• Should the ACCC make use of an efficiency
  carry-over mechanism?
• in its revenue cap decisions for TNSPs - either
  for opex alone or for both opex and capex?

3
Incentive regulation

• These issues relate to the design of incentive
  regulation for TNSPs
• Incentive regulation is about giving the firm
  financial rewards for pursuing desirable
  objectives
• such as promoting service quantity and quality
  or reducing expenditure (opex or capex)
• Theory suggests that the power of the incentives
  to promote different objectives should be
  balanced
• Incentives to maintain service quality should be
  balanced with incentive to reduce expenditure
• Incentive to reduce opex may need to be balanced
  with incentive to reduce capex

4
Incentive regulation

• The power of the incentive to reduce
  expenditure depends on the sensitivity of the
  firms profit stream to a change in its own
  expenditure
• The revenue stream is fixed for the remainder of
  the regulatory period so the power of the
  incentive to reduce expenditure depends primarily
  on the sensitivity of the future revenue stream
  to expenditure out-turns in the current period.
• In particular, in the case of recurrent
  expenditure, the power of the incentive to reduce
  expenditure depends primarily on how future
  revenue targets are set in the light of past
  expenditure out-turns

5
Incentive regulation

• The power of the incentive to reduce recurrent
  expenditure depends on how information on past
  expenditure out-turns
• is used in the setting of future revenue
  targets.

6
Need for a mechanistic approach

• If a regulatory regime is to have known and
  predictable incentive properties the regulator
  must commit to a mechanistic approach to using
  past out-turn information when setting the future
  revenue targets.
• To the extent that the scrutiny of external
  consultants acts as a control on the revenue
  targets the precise manner in which the
  consultants determine the appropriate level of
  cost is of primary importance for assessing the
  power of the incentive scheme Any attempt to
  clarify the process will inevitably involve ..
  Replacing this discretionary process with a more
  mechanistic process.
• There must be some formula by which past
  expenditure out-turns are converted into future
  revenue targets

7
Carry-over Mechanism

• Some regulators make use of a carry-over which
  sets the revenue target as the sum of an
  underlying target and a carry-over

Underlying target Year 6
Underlying target Year 7
Underlying target Year 8
Underlying target Year 9
Underlying target Year 10
Carry-over Year 6
Carry-over Year 7
Carry-over Year 8
Carry-over Year 9
Carry-over Year 10


8
Carry-over Mechanism

• But the incentive properties depend only on how
  the total amount of revenue the firm receives
  depends on the past out-turns
• In particular, it is not possible to determine
  the incentive properties of a regulatory regime
  by examining the carry-over alone
• The incentive properties of a regulatory regime
  can only be understood if the regulator commits
  to a mechanistic process for setting both (a) the
  underlying targets and (b) the carry-over.
• But why make things more complicated and less
  transparent?

9
Carry-over Mechanism

• Several submissions said that the efficiency
  carry-over mechanism
• is a tool for overcoming the information
  asymmetry that exists between the regulated firm
  and the regulator
• causes the regulated firm to reveal the
  efficient costs of business to the regulator
• But is this right?
• Theory shows that no mechanism based on the
  firms own costs can induce the firm to reveal
  its fully efficient costs
• But increasing the power of an incentive
  mechanism can increase the incentive to reduce
  expenditure

10
Properties of different mechanisms

• If the regulator is going to commit to a
  mechanistic approach to setting the future
  revenue targets which approach should it choose?
• How should the revenue targets depend on
  information on past expenditure out-turns?
• Ideally we would like constant incentives for
  minimising expenditure over time
• Incentives for minimising expenditure should be
  balanced with incentives to enhance service
  quantity and quality.
• Lets examine the effects of different approaches

11
Example 1

• Suppose first that the regulator sets the future
  revenue targets on the basis of the expenditure
  out-turn in the last year of the last regulatory
  period
• This creates very strong incentives to reduce
  costs in the first four years of the regulatory
  period and very strong incentives to increase
  costs in the last year

12
Example 2

• Now suppose that we add an efficiency carry-over
  mechanism to this approach
• Now we have constant incentives to reduce
  expenditure over time
• But there are other approaches

13
Example 3

• Now suppose that the revenue targets are set on
  the basis of a weighted average of the
  expenditure out-turn in the previous regulatory
  period
• Again we have achieved constant incentives to
  reduce expenditure over time

14
Example 4

• Now suppose that we add an efficiency carry over
  to example 3
• Now we have mild incentives to raise costs in
  years 1-4 and very strong incentives to reduce
  costs in year 5

330
15
Interim conclusions

• These examples re-emphasise the following points
• The incentive properties of an incentive
  mechanism depend not just on the carry-over but
  on how both the carry-over and the underlying
  targets are set
• The use of an efficiency carry-over may make a
  bad outcome better or a good outcome bad,
  depending on how the underlying targets are set
• The use of an efficiency carry-over, when
  combined with setting the underlying targets on
  the basis of the last year of the last regulatory
  period is entirely equivalent to setting the
  underlying targets on the basis of a weighted
  average of all the previous out-turns

16
Interim conclusions

• Conclusion A carry-over mechanism is not
  necessary and is potentially harmful or
  misleading
• The simplest approach to achieve constant
  incentives is to set the revenue targets for the
  next regulatory period on the basis of a weighted
  average of the expenditure out-turns for the
  previous regulatory period.
• But is it the case that a revenue target based on
  a weighted average will always yield sufficient
  revenue?
• Recurrent expenditure could be higher due to
  exogenous factors such as
• Changes in wage rates or exchange rates Changes
  in the size of the network Changes in the age of
  the network

17
Handling scope changes

• These factors can be taken into account by adding
  or subtracting an exogenous amount from the
  simple weighted average
• We can add or subtract an arbitrary amount
  without affecting the incentives as long as the
  amount added or subtracted is exogenous that
  is, completely independent of past expenditure
  out-turns
• This exogenous offset can be used to take
  account of factors which shift expenditure (such
  as changes in wage rates or exchange rates)

18
Handling scope changes

• In addition, as a further protection, the
  regulator could give the regulated firm the
  option of a recovering its actual expenditure ex
  post.
• This approach ensures that the regulated firm is
  able to recover its actual costs, if actual costs
  are higher than expected
• This is an example of the menu of options
  approach mentioned in the discussion paper
• Analogously, both ESC and ESCOSA have chosen to
  not carry-over a negative carry-over amount
  which also increases the likelihood that the firm
  will be adequately compensated

19
Exogenous Cost Measures

• We have discussed how the revenue targets should
  depend on the firms own expenditure out-turns
  but in theory it might be possible to cut the
  link with a firms own expenditure out-turns
  entirely by setting the revenue targets purely on
  exogenous cost measures (also sometimes called
  benchmarking)
• For example, in theory we could use information
  on the performance of other TNSPs to develop an
  estimate of the efficient long-run cost of one
  TNSP
• This would yield high-powered incentives for
  reducing expenditure (not necessarily desirable)
• There would be no need for an efficiency
  carry-over

20
Exogenous Cost Measures

• But there was little enthusiasm for pursuing this
  approach
• Given the large number of factors which affect
  TNSP expenditure and the small number of
  comparator firms it would be difficult to create
  a suitable cost model.
• Capex is highly variable over a long time period
  so benchmarking of capex was viewed as
  infeasible
• Note the ACCCs current use of certain ratios as
  a sanity check is not benchmarking and is of
  limited use contains serious flaws and is
  totally unsuited to be used to imply a TNSP is
  efficient or to reward or penalise a TNSP for its
  performance. SPI PowerNet

21
The Way Forward

• I propose the ACCC adopt a mechanistic approach
  to setting the revenue targets for recurrent
  expenditure
• The underlying target for future periods would
  be set on the basis of the weighted average of
  the expenditure out-turns in previous periods
• In addition, any exogenous differences in cost
  across periods which can be anticipated would be
  addressed through an exogenous offset (positive
  or negative) which is independent of past
  expenditure out-turns
• Finally, the regulated firm would have the option
  ex post of obtaining recovery of actual
  expenditure (through an adjustment to the RAB).

22
The Way Forward

• SPI PowerNet propose a similar approach
• Opex expenditure during the future regulatory
  period would be set at the average of the
  preceding five years expenditure (adjusted for
  inflation) thus ensuring consumers receive the
  benefits of any efficiencies achieved .
• This would be adjusted up or down for movements
  in uncontroversial, simple and clear exogenous
  costs such as OHS regulations and standards,
  environmental regulations, legislative changes,
  property taxes, insurance costs, etc.
• Further adjustments would need to be made for
  volume effects as the assets under management can
  increase substantially during a five-year
  regulatory period .