Horton 4 Consulting

Review of regulatory finance issues

A report to Guernsey Electricity Ltd by Geoff Horton

 

 26 October 2005

 

CONTENTS

1 Introduction and summary *

2 Regulatory objectives *

3 Economic efficiency *

3.1 Productive efficiency *

3.2 Allocative efficiency *

3.3 Dynamic efficiency *

3.4 Conclusions *

4 Regulatory calculations *

4.1 Capitalisation and depreciation: the standard "building block" method *

4.2 Infrastructure accounting *

4.3 A cash approach *

4.4 OUR’s approach *

5 Company accounts *

6 Numerical examples *

7 Conclusions *

    1. Introduction and summary

      2. The Guernsey Office of Utility Regulation (OUR) has published a draft decision in its review of the price control of Guernsey Electricity Limited (GEL). The decision is based on a method that targets GEL’s cash reserves. GEL has asked Horton 4 Consulting to assess this method, compare it with standard regulatory practice, and recommend any appropriate changes to the method.

      In this paper it is argued that the principles to be applied in assessing the price control are primarily those of economic efficiency. The overall objective is to produce the greatest net benefit - benefit to consumers less overall costs - and this is achieved through the promotion of economic efficiency.

      To satisfy productive and dynamic efficiency it is necessary that an investor should have the expectation of recovering the cost of capital on efficient investment. This requires a method of valuation that ensures an expectation of "financial capital maintenance (FCM)". However, when industries have been privatised and a new regime initiated, the regulatory value adopted for the assets already in existence at that time has often differed from both the historic and current cost values of those assets.

      From a consumer point of view, and to satisfy allocative efficiency, the need is to set a charge that reflects the marginal cost of provision. This is so that the customer will purchase whenever the benefit of consumption exceeds the cost (and so price) and will refrain from purchasing when the benefit is below the cost. Allocative efficiency is promoted by pricing at marginal cost, which may be below the average cost calculated using current cost asset values if there are economies of scale, but is unlikely to be approximated if existing assets are ignored.

      This paper considers three standard methods of utility price control – the building block method used in UK energy and other regulation, infrastructure accounting as used in water and (previously) rail and a cash flow approach – and demonstrates that they all achieve the same result.

      The method used by OUR in its draft decision is a cash flow approach but one that does not consider the cost of the use of existing assets. The effect is that no allowance is made for depreciation of the existing assets or for a return on the undepreciated part. This fails to produce FCM but, given GEL’s position as a government-owned company, it is not certain that it does so in a way that promotes inefficiency. However, the resulting prices are likely to fail to cover the resource cost and so to result in allocative efficiency.

      The argument put forward to support OUR’s method is that the entire asset base and the present cash reserve have been financed by payments by customers in the past, with no contribution from government or taxpayer as the owner of the company, and so to make any further charge in respect of these assets would be to charge twice.

      The argument from fairness, that customers should not pay twice, is less powerful than it at first appears since the payment is a transfer between customers and GEL’s owners, the citizens of Guernsey, who are to a large extent the same people. GEL could distribute any cash surplus it acquires to its owners, thereby compensating them for the higher charges, while at the same time ensuring that electricity prices reflect costs and so that economically efficient choices are made by customers – a situation that is unlikely to occur if a negative value is put on existing assets.

      Moreover, OUR does not appear to be correct in its assumption that customers have financed the existing assets in advance and should therefore be deemed to own them already. When a private sector company finances investment from its cash balances it is said to finance it from retained earnings. It is not said to be financed by customers. According to information provided to me by GEL, this is also the position of that company. GEL is the successor to the States Electricity Board, which was formed by the purchase of an existing power company. The purchase was financed by a loan from the States and a further loan was taken out in 1950 to finance capital expenditure. Customers were not charged for capital expenditure in advance. It was originally financed by borrowing with a state guarantee covering the equity risk. Although interest payments were made on the original loan there have been no dividend payments. Those earnings were retained by GEL and used to finance subsequent investment. As far as I know, there is no evidence that customers have paid a price that exceeds the cost of their supply in order to finance future investment.

      There is a further problem that arises if no revenue is granted in respect of existing assets. They have a positive value (of about £100 million) in the company’s balance sheet and a depreciation charge is levied in respect of them in the profit and loss account. In many circumstances this will result in the company recording a loss. There is a strong argument that, if the proposed price control is introduced, GEL should revalue the balance sheet and set the book value of these assets to zero. If this were to be done, GEL would sustain a large initial loss.

      I therefore recommend that, if a cash flow approach is used, the price calculation is adjusted to include a dividend payment in respect of the existing assets, with a consequent impact on both P0 and X and so on the expected revenue path. This would accord with standard UK practice and would be more likely to achieve allocative efficiency and to avoid a possibly unsustainable position for GEL’s accounts and financial indicators.

    2. Regulatory objectives

GEL is subject to a price control because it has considerable market power. In setting the price control OUR is seeking to achieve the objectives set out in section 2 of the Regulation of Utilities (Bailiwick of Guernsey) Law 2001. The six duties can be summarised as to:

The duties are similar to those of other utility regulators in the United Kingdom, including those to which I was subject when electricity regulator in Northern Ireland, except that there is no duty to ensure that licence holders can finance the activities for which they are licensed.

However, most UK utility regulators have seen the objectives as complementary, rather than conflicting, and considered that their reordering and revision in the 2000 Utilities Act had little, if any, effect. The main objective is the promotion of customers’ long term interests and this entails many of the other objectives, including that concerning licensees’ ability to obtain finance. Customers’ interests include that their reasonable and economical demands for the service are met and this cannot be done if efficient licence holders are unable to finance their activities. Competitive markets, if they are effective, are the best means of furthering customer’ interests. Environmental costs have to be weighed against the benefits to consumers, as do other costs of providing the service.

The overall objective is to produce the greatest net benefit - benefit to consumers less overall costs - and this is achieved through the promotion of economic efficiency. Regulators’ duties often include the promotion of efficiency and economy in the industry, rather as OUR’s duties include the general promotion of economic well-being in Guernsey through the activities of the regulated utility. Indeed, I would maintain that the principles to be applied are primarily those of economic efficiency.

3. Economic efficiency

Economic efficiency is often considered in three categories:

The first requires, inter alia, that producers are faced with incentives that reflect the true costs of the resources they use so that they can minimise the overall cost, the second that consumers face prices that reflect the full cost of providing the good or service so that they can make efficient choices and the third that investors can recover the costs of efficient investment and innovation.

These considerations also underlie the Bonbright principles, which are widely cited in United States regulation. These involve three primary criteria:
    1. Productive efficiency

      2. To satisfy productive efficiency and capital attraction it is necessary that an investor should have the expectation of recovering the cost of capital on efficient investment. This requires a method of valuation that ensures an expectation of "financial capital maintenance".

      Under financial capital maintenance (FCM) the regulated firm will, if it is operating efficiently, recover the financial capital invested in its assets; payments for return and depreciation (when discounted at the cost of capital) will equal the cost of the investment. This clearly will not apply if assets are excluded from an asset base but, provided they are included, it is a principle that is compatible with many systems of asset valuation and price setting. However, it will have implications for the method of application of each.

    2. Allocative efficiency

      3. From a consumer point of view, and to satisfy allocative efficiency, the need is to set a charge that reflects the marginal cost of provision. This is so that the customer will purchase whenever the benefit of consumption exceeds the cost (and so price) and will refrain from purchasing when the benefit is below the cost.

      The cost valuation that is relevant for this perspective is that of an expansion or contraction of output, which would require purchase of inputs at the present cost and so, as regards capital assets, the purchase of a modern equivalent asset (MEA) or, if less expensive, the use of other means of achieving the same objective without the asset, which is known as the optimal deprival value (ODV).

      Marginal cost should be calculated valuing assets at replacement cost (or ODV) but it may differ from the calculated average cost if there are economies of scale or if the measurement of average cost values assets differently. There is a large literature about the reconciliation of the marginal cost principle with that of raising sufficient revenue to cover all costs (including a reasonable return) in circumstances when there are economies of scale and marginal cost is below average cost.

      There are various studies that have investigated the existence of scale economies in electricity distribution. Some have concluded that they exist but others have concluded that only very small scale is a disadvantage. In New Zealand Meyrick and Associates concluded that economies of scale are unimportant.

      If there are no economies of scale allocative efficiency will require that prices should be based on an assessment of costs that is comprehensive, i.e. including the costs of existing assets. Even if there are economies of scale it is likely that the cost of existing assets would need to be taken into account. For example if the degree of economies of scale is such that costs increase 0.9% when output increases 1% and if capital costs are half of total costs, marginal cost pricing could be approximated by valuing assets at 80% of their real cost.

    3. Dynamic efficiency

      4. There is less consensus over the conditions for dynamic efficiency but it is generally recognised that two important factors are competition, to promote innovation, and FCM, to ensure that efficient investment will be rewarded and so that it takes place.

    4. Conclusions

Financial capital maintenance is important for both productive and dynamic efficiency but, when industries have been privatised and a new regime initiated, the regulatory value adopted for the assets already in existence at that time has often differed from both the historic and current cost values of those assets.

Allocative efficiency is promoted by pricing at marginal cost, which may be below the average cost calculated using current cost asset values if there are economies of scale but is unlikely to be approximated if existing assets are ignored.
  1. Regulatory calculations

    2. Given an initial asset value, the standard regulatory objective in price control calculations has been to provide an expectation of FCM and to attempt to influence the structure of prices to approximate to marginal cost pricing.

    1. Capitalisation and depreciation: the standard "building block" method

Probably the simplest method of securing FCM is when prices are set on a "building block" approach to recover costs using, in respect of the cost of capital assets, straight line depreciation of the historical cost of the asset plus the nominal cost of capital as a return on the undepreciated portion. This recovers the cost of the assets through a process of capitalisation and depreciation. However, use of historic costs will tend to frontload charges relative to an assessment of prices based on marginal cost, particularly when inflation is high, and so will be likely to conflict with allocative efficiency.

An alternative is to index the asset value (and so also the depreciation charge) for inflation and to apply a return equal to the indexed asset value times a real cost of capital. The discounted sum of returns and depreciation charges will equal the cost of the asset and ensure FCM. This indexed value too may differ from a modern equivalent asset or deprival value and so from marginal cost but the difference may not be large.

The "building block method" with indexed values is the standard approach used in energy regulation in the UK to calculate required future revenue. The objective is to set a base price and a value for X so as to produce prices during the price control period which generate revenues equal to the expected value of costs. The sum is done in present value terms discounting future costs and revenues by the company's cost of capital each year. Over a run of years this is equivalent to assuming that the company earns a rate of return equal to its cost of capital. The present value equation can be derived from the more simple form by transformation of the series.

The normal approach is to consider a regulatory period as a whole and set a price control so that the net present value of expected revenue is equal to that of expected costs using an equation similar to that below, where all values are expressed in constant prices, ror is the rate of return, RAB the regulatory asset base, C current spending, I investment spending and the suffixes relate to the year in question.

In other words, revenue during the period should be equal to:

Assets at the end of the period are normally calculated by adding the capitalised new investment to the starting assets, applying a depreciation profile to the starting assets and new investment, and calculating the final asset stock by subtracting that depreciation from the starting assets and new investment.

The full equation (of the kind shown below) sets the net present value of expected revenue, given the revenue driver volume (V) expected in each year, equal to that of the expected costs of an efficient company by setting a base price (Po) and X factor.

    1. Infrastructure accounting

      2. In the water industry, and for a time in railways, there has been a slightly different approach. A value was set on the original network, in both cases well below any estimate of its current cost value, and a return granted on that value at the cost of capital. Other allowable costs were operating costs and, instead of depreciation, a renewals charge designed to cover the cost of replacement and maintenance.

      The present value of the infinite stream of returns will equal the initial value and all other costs are expensed and rewarded as they are incurred, so there is FCM and this "infrastructure" method produces the same expectation for the net present value of revenue as the building block method used in energy.

      The system is complicated by the fact that "enhancements" to the network’s capacity (either in quantity or quality) are treated differently. These are capitalised and rewarded on the basis of a return and depreciation.

    2. A cash approach

      3. A further extrapolation of the infrastructure approach would be to grant a return on the original asset value and pay for all expenditure, whether on maintenance, renewal or enhancement as it occurs.

      The costs to be considered would therefore be all future expenses plus a perpetual dividend stream whose net present value (at the cost of capital) would equal the initial network value. The net present value of revenue would be set equal to the expected net present value of these costs. This should achieve the same NPV as the other methods and has been advocated for use in the UK.

    3. OUR’s approach

OUR appears to have adopted a form of such a cash flow approach but with the difference that no value or dividend stream been assigned to the existing assets, on the grounds that under the "save to pay" policy customers have already paid for them. Indeed, in practice, a negative value has been assigned since GEL’s cash balance is to be reduced from around £20 million now to £10 million in real terms in 2016.

The effect is that no allowance is made for depreciation of the existing assets or for a return on the undepreciated part.

The argument put forward to support this is that the entire asset base and the present cash reserve have been financed by payments by customers in the past, with no contribution from government or taxpayer as the owner of the company, and so to make any further charge in respect of these assets would be to charge twice.

If this were the case, customers would have been paying prices that were probably well above marginal cost, particularly in GEL’s early days. The argument is that they should now be compensated by paying lower prices. However, this is likely to result in prices that are below costs and so in allocative inefficiency.

The argument from fairness, that customers should not pay twice, is less powerful than it at first appears since the payment is a transfer between customers and GEL’s owners, the citizens of Guernsey, who are by and large the same people. GEL could distribute any cash surplus it acquires to its owners, thereby compensating them for the higher charges, while at the same time ensuring that electricity prices reflect costs and so that economically efficient choices are made by customers – a situation that is unlikely to occur if a negative value is put on existing assets. However, there will be some mismatch between the customer base and taxpayers.

Moreover, OUR does not appear to be correct in its assumption that customers have financed the existing assets in advance and should therefore be deemed to own them already. When a private sector company finances investment from its cash balances it is said to finance it from retained earnings. It is not said to be financed by customers.

According to information provided to me by GEL, this is also the position of that company. GEL is the successor to the States Electricity Board, which was formed by the purchase of an existing power company. The purchase (plus an initial cash balance) was financed by a loan from the States. This was subsequently repaid but a further loan was taken out in 1950 to finance capital expenditure. Customers were not charged for capital expenditure in advance. It was originally financed by borrowing. It would be unusual for a private sector electricity company to be entirely debt financed but state-owned companies are, because the state guarantees the company and so takes what would otherwise be the equity risk. Although interest payments were made on the original loan there have been no dividend payments. Those earnings were retained by GEL and used to finance subsequent investment. As far as I know, there is no evidence that customers have paid a price that exceeds the cost of their supply in order to finance future investment.

OUR’s proposals would seem therefore not to accord with FCM. However, it is possible that the owner of a state-owned company might agree to a breach of FCM without there being adverse impact on the incentive to invest and so on productive and dynamic efficiency, provided that it is made clear that the circumstances are exceptional and that the breach could not occur without the owner’s consent. However, such a step – particularly one involving assigning a zero value to existing assets – would be without precedent. Apparently similar occasions, when regulators have used asset values below the book value at the time of a first industry price control, have been ones where the state has privatised the assets below book value and FCM has been with respect to what the shareholders paid for the assets and not what the state originally paid.

In summary, it would appear that:
  1. Company accounts

    2. There is a further problem that arises if negative revenue is granted in respect of existing assets. They have a positive value (of about £100 million) in the company’s balance sheet and a depreciation charge is levied in respect of them in the profit and loss account. In many circumstances this will result in the company recording a loss. This can be seen by considering that revenue is calculated as operating costs plus investment less interest earnings less a share in the decline in cash balance. Costs in the profit and loss account are operating costs plus depreciation. The latter often exceeds investment, let alone investment less interest earnings and cash reduction. Costs will often, perhaps even normally, exceed revenue.

    There is a strong argument that, if the proposed price control is introduced, GEL should revalue the balance sheet. Assets should be priced according to the "value-to-the-owner" rules. These set the value at the lower of the replacement cost or the economic value, where the economic value is the higher of the net present value of the asset’s expected future earnings and its possible sale value. The future earnings of the existing assets are set to zero (on the grounds that they have already been paid for) and they have little resale value (not least because GEL are required to use them to produce electricity) so their "value-to-the-owner" is zero. This is therefore the appropriate balance sheet value. If this were to be done, GEL would sustain a large initial loss.

    I have not investigated the likely financial consequences for GEL of OUR’s proposals in any detail but it seems likely that such circumstances could result in outcomes for GEL’s accounts and for financial indicators (such as interest cover), which are commonly taken into account in UK utility regulation, that could be potentially unsustainable.

  2. Numerical examples

These points can be illustrated using the figures for operating costs, investment, units sold, the interest rate and the cost of capital adopted by OUR in their draft decision. The use of these values, and of the assumption that there are no other expenditures required, does not imply endorsement of them. Their assessment is not within the scope of this study.

A price path that is very close to that proposed by OUR can be obtained by calculating the constant real decrease in prices (the X in RPI-X) from 2006 that produces a revenue that, given the assumed expenditures on operating costs and investment (net of customer contributions) together with a starting cash balance of about £20 million, results in a balance of about £10 million at the end of 2016/2017. This calculation has been labelled "Cash flow I". X is 0.0026. Prices fall at just over a quarter of a per cent a year in real terms. The path of proposed prices does not appear to have a constant real change over the entire period. It is initially slightly higher, then lower and finally (after 2020) higher again; but the differences are small.
    1. Comparison of OUR proposals with a simple cash flow

    2. Comparison of proposals with standard methods rewarding the asset base

However, such an approach does not give any reward to the existing capital asset base. Figure 2 therefore compares these two price series (this time with general inflation removed so as to express them in 2004/05 prices) with three forms of the standard method:

      1. The same cash flow approach but adding a dividend payment each year equal to the cost of capital times the initial net asset stock: the dividend is constant in real terms, i.e. it increases with inflation. Cash flow II

      2. The building block approach that sets revenue equal to operating costs plus depreciation plus a return of the cost of capital on the asset base. The calculation of revenue and of the regulatory asset base has assumed a 40 year asset life and straight line depreciation. A further assumption about the rate of retirement of gross assets was also required. The resulting depreciation charge was lower than the accounting figures given to me. Building block

      3. An infrastructure approach that arbitrarily separated investment into replacement (70%) and enhancement (30%). Infrastructure

Unsurprisingly, the methods rewarding the asset base produce higher prices.

Calculating a price based on the costs each year produces a rather unstable price. This is particularly the case with the infrastructure method which adjusts revenue to fund 70% of a highly volatile investment path. Figure 3 therefore calculates an RPI-X price path for the building block and infrastructure approaches that produces the same net present value of revenue as the volatile price paths in Figure 2. As theory predicts, the three methods (cash flow II, building block, infrastructure) produce the same path, RPI+1.2%.
    1. RPI-X price paths for the standard UK method and the proposals

      2. The upward price path that results when the asset base is rewarded is a little misleading. It produces early HCA losses, particularly when the actual depreciation figures from GEL’s accounts are used, followed by large profits in later years. A "P0 adjustment" that raised prices by 13% and then kept them flat would produce the same net present value of revenue and flatter profits. This is shown as a "level" price in Figure 4. There are many other possible combinations of P0 adjustments and X factors.

    2. Different P0 adjustments and X factors
  1. Conclusions

Economic efficiency is the appropriate criterion by which to assess the price control proposal.

The standard methods of utility price control seek to ensure that an efficient company will achieve financial capital maintenance (FCM), which promotes productive and dynamic efficiency, and that customers face prices that reflect the true costs of their purchases, which promotes allocative efficiency. Possible standard methods include capitalisation and depreciation (as used in the UK in energy regulation), infrastructure accounting (as used in water) and cash flow. If used correctly these can all promote efficiency and be reconciled with each other.

The method used by OUR in its draft decision is a cash flow approach but one that does not consider the cost of the use of existing assets. The effect is that no allowance is made for depreciation of the existing assets or for a return on the undepreciated part. This fails to produce FCM but, given GEL’s position as a government-owned company, it is not certain that it does so in a way that promotes inefficiency. However, the resulting prices are likely to fail to cover the resource cost and so to result in allocative efficiency

Allocative economic efficiency is unlikely to be served by setting a price that does not take the cost of the existing assets into account. The argument from fairness for failing to include them (that customers have already paid for the assets)

    1. does not appear to be supported by any evidence that investment has been financed from payments from customers that exceed the costs of their supply rather than from the retained earnings of GEL and its predecessor and

    2. would be weakened, even if true, by the fact that GEL is state owned and so the beneficiaries of the additional payment would be in large part electricity customers in their role as citizens.

An additional argument against failing to include existing assets as a determinant of price is that GEL would probably make accounting losses and may have to write down its existing asset value to zero. This could put it in a very difficult financial position.

I therefore recommend that, if a cash flow approach is used, the price calculation is adjusted to include a dividend payment in respect of the existing assets, with a consequent impact on both P0 and X and so on the expected revenue path. This would accord with standard UK practice and would be more likely to achieve allocative efficiency and to avoid a possibly unsustainable position for GEL’s accounts and financial indicators.

Prices and profits