Information Paper Financial Capital Maintenance and Price Smoothing February 2014
The QCA wishes to acknowledge the contribution of the following staff to this report: Ralph Donnet, John Fallon and Kian Nam Loke Queensland Competition Authority 2013 The Queensland Competition Authority supports and encourages the dissemination and exchange of information. However, copyright protects this document. The Queensland Competition Authority has no objection to this material being reproduced, made available online or electronically 600182 ver 1 but only if it is recognised as the owner of the copyright 2 and this material remains unaltered. Draft as at 04/02/14 9:06
Queensland Competition Authority Table of Contents Table of Contents GLOSSARY OF ACRONYMS, TERMS AND CONDITIONS 36 THE ROLE OF THE QCA TASK AND CONTACTS EXECUTIVE SUMMARY IV V 1 FINANCIAL CAPITAL MAINTENANCE 1 1.1 Definitions 1 1.2 Literature review 2 1.3 Implications of financial capital maintenance 3 1.4 Summary 7 2 BUILDING BLOCKS APPROACH 8 2.1 Return on and return of capital 8 2.2 Indexation of the asset value 11 2.3 Summary 13 3 PRICE SMOOTHING 14 3.1 Smoothing 14 3.2 Smoothing period 17 3.3 Smoothing period matches the asset life 18 3.4 Smoothing period differs from the asset life 21 3.5 Changing demand 26 3.6 Practical considerations 28 3.7 Summary 29 4 CONCLUSION 31 APPENDIX A 32 REFERENCES 37 600182 ver 1 iii Draft as at 04/02/14 9:06
Queensland Competition Authority The Role of the QCA Task and Contacts THE ROLE OF THE QCA TASK AND CONTACTS The Queensland Competition Authority (QCA) is an independent statutory authority that promotes competition as the basis for enhancing efficiency and growth in the Queensland economy. The QCA s primary role is to ensure that monopoly businesses operating in Queensland, particularly in the provision of key infrastructure, do not abuse their market power through unfair pricing or restrictive access arrangements. In 2012, that role was expanded to allow the QCA to be directed to investigate, and report on, any matter relating to competition, industry, productivity or best practice regulation; and review and report on existing legislation. Task This information paper explains and explores the financial capital maintenance (or NPV=0) principle and documents methods of price smoothing that have been applied by the QCA. The paper was motivated by the need to document methods the QCA has applied and to highlight a number of issues that arise in the context of determining price paths over time. Contacts Comments and enquiries regarding this paper should be directed to: ATTN: Kian Nam Loke Tel (07) 3222 0582 research@qca.org.au www.qca.org.au 600182 ver 1 iv Draft as at 04/02/14 9:06
Queensland Competition Authority Executive Summary EXECUTIVE SUMMARY This paper discusses: (1) the ex ante financial capital maintenance (FCM) principle (also known as the NPV=0 principle); and (2) price smoothing. While an appropriate form of price smoothing can take on various forms, it should always be subject to the FCM principle. The FCM or NPV=0 principle refers to the requirement that the present value of expected capital charges for an asset over its economic life should be equal to the initial asset value or purchase cost. The capital charge for all assets (i.e. the RAB) for a regulatory year comprises both the return on and return of capital. The FCM principle also holds that at any given point in time the present value of expected capital charges over the remaining life of the asset should equal the current value (initial value less allowed depreciation from previous periods) of such an asset. The paper reviews some key papers on rate base and depreciation, and presents a mathematical model that illustrates the implications for depreciation and the evolution of the asset value if the FCM principle is applied over the full life of the asset. This model shows that in order to meet such a principle for a particular asset, the regulator has three options: (a) (b) (c) Specify a schedule for capital charges with a present value that matches the initial asset value. This is typically in the form of an annuity, which is a series of constant or indexed annual capital charges applicable over the life of the asset. Specify a schedule for depreciation amounts that add up to the initial asset cost over the asset life. There are four standard depreciation profiles, namely straight-line, front-end loaded, back-end loaded, and one-hoss shay. In practice, any profile that satisfies the FCM principle can be adopted. Specify the evolution of the asset value. Since the change in the asset value between periods is the applicable depreciation amount, setting the evolution of the asset value is equivalent to setting the depreciation profile provided the FCM principle is satisfied. A smoothed revenue or price path is typically obtained by regulators adjusting the building block annual revenue requirement (ARR) to produce either a smoothed revenue or price path. This paper derives the smoothing formulae and provides various examples of smoothing. In applying smoothing, there can be abrupt and largely arbitrary changes in prices between regulatory periods, depending on the choice of depreciation method adopted when the building blocks revenue is calculated. For example, straight-line depreciation produces capital charges that are larger in absolute terms in the beginning of the asset life relative to those in the later years. This typically means that applying straight-line depreciation will result in lower prices in the future relative to prices in the present. As a result smoothed prices can be materially different between regulatory periods. The differences in prices may be exacerbated when demand is growing over time. While this issue is not as apparent in the case where the firm makes investments over time, the contributions by customers to the capital recovery of regulated assets in one period may differ materially from contributions in another period. The time path of capital charges as a whole often receives less attention than the depreciation profile. While straight-line depreciation, the most commonly applied depreciation profile, produces equal depreciation amounts (in real terms under the QCA's approach) over the life of the asset, the resulting capital charge is in fact decreasing in value over time. This result is often neglected by regulators. A backend loaded depreciation profile can produce a more balanced capital charge that might be considered as 600182 ver 1 v Draft as at 04/02/14 9:06
Queensland Competition Authority Executive Summary more 'equitable', especially if there is excess capacity and the demand is expected to rise significantly over time. The paper concludes that the optimal profile of capital charges needs to be determined by separate efficiency or equity considerations. For example, considerations of excess capacity, growth of demand, potential for asset stranding and intergenerational equity would all be relevant in determining the optimal time profile of prices for a regulated entity. 600182 ver 1 vi Draft as at 04/02/14 9:06
Financial Capital Maintenance 1 FINANCIAL CAPITAL MAINTENANCE Biggar (2011) prepared brief summaries of The Fifty Most Important Papers in the Economics of Regulation'. These papers are divided into fifteen topics one of which, 'Rate Base and Depreciation' is considered here. This topic focuses on the regulated firm's recovery of invested capital. Biggar highlights the conclusion drawn by Greenwald (1984) that 'as long as the regulatory regime ensures that the change in the asset base over the regulatory period is equal to the present value of the payments to investors (discounted at the appropriate cost of capital), investors will be adequately compensated for their investments in the long run'. This is commonly referred to as the financial capital maintenance (FCM) principle, or the NPV=0 principle, in regulatory economics. This: (a) (b) (c) 1.1 Definitions provides an initial understanding of the concept of financial capital maintenance and its origins outlines the findings of some key papers on rate base and depreciation discusses the implications for depreciation and the asset value if the FCM principle is applied over the full life of an asset. Capital maintenance is a longstanding financial accounting concept that can be related to economic notions of the recovery of capital. Financial capital refers to the value of business assets as measured by the value of equity and debt, while physical capital is the productive or operating capacity of the assets. Financial capital maintenance (FCM) is the maintenance of the initial value of a business as measured by the value of assets at the time of investment. Under the accounting definition of FCM, a profit is earned only if the financial value of net assets at the end of the period exceeds the financial value of net assets at the beginning of the period, after excluding any distributions to, and contributions from owners during the period. In the regulatory context, FCM is applied in an ex ante sense, meaning that investors of a regulated firm can expect to recover the opportunity cost of their capital and the nominal value of their initial investment over time. This is referred to as the FCM principle. As long as the present value (PV) of future regulated returns, calculated on the basis of an appropriate opportunity cost discount rate, is equal to the value of the regulatory asset base (RAB), the FCM principle is achieved. The FCM principle in an exact sense is often referred to as the NPV=0 principle. In contrast, physical or operating capital maintenance (OCM) focuses on maintaining the productive capacity of physical assets over time, rather than the financial value of assets. In the accounting context, this means that profit is only recognised after the operating capacity of assets has been maintained or when the operating capacity of the enterprise at the end of the period exceeds the operating capacity at the beginning of the period, after excluding any changes to capacity associated with additions or disposals of assets during the period. OCM determines asset prices and depreciation charges based on the cost of replacing assets in order to maintain operational capability at a defined level. 600182 ver 1 1 Draft as at 04/02/14 9:06
Financial Capital Maintenance For a defined level of operational capability, FCM differs from OCM in an accounting sense by recognising capital gains and losses associated with holding the assets as well as the standard OCM charge. 1.2 Literature review This section reviews the four seminal papers highlighted by Biggar (2011) on the cost recovery of sunk investment in the context of economic regulation. To achieve this cost recovery, Australian regulators typically apply a form of building blocks model, where prices are set to allow regulated firms to earn expected revenue (after deducting operating expenditure and taxes) that provides a 'fair' rate of return on the RAB plus an allowance for economic depreciation. Biggar (2011) notes that there are at least three approaches for setting the RAB: (a) the historical cost of the assets of the firm; (b) the replacement cost of the assets of the firm; and (c) the historical cost of the assets of the firm adjusted for inflation. 12 The first paper reviewed by Biggar is Greenwald (1984). Greenwald's model, which focuses on the evolution of the RAB, shows that under certain conditions, as long as (a) depreciation is defined as the change in the RAB each period, and (b) the regulator sets the allowed rate of return equal to the firm's true cost of capital, the three asset valuation approaches above are consistent with investors earning a fair and normal return. According to Greenwald, the RAB should satisfy the following conditions: (a) (b) (c) The RAB cannot be negative. The RAB must be at all times be less than the market value of a profit-maximising unconstrained monopolist, hence ensuring that no regulated firm can earn more than it could if it were unregulated. The RAB should start at zero before any investment is sunk and should finish at zero when the firm ceases to exist. Greenwald (1984) argues that the concept of the RAB avoids the "dynamic consistency" problem as it ensures that the firm will always be adequately compensated for its investments in the long run, regardless of the form of regulatory regime in place. For example, suppose that a regulator had traditionally set prices that would yield optimal behaviour by the firm, but for a specific period, perhaps due to political pressure, the regulator decided to alter its approach and specified prices that sought to minimise customer costs. As long as the change in the RAB over this period reflected the present value of the payments to the firm, the firm would still earn a normal return on its investments. This provides a strong argument for the use of the building blocks model, as it ensures that the regulated firm will always be allowed an expected return consistent with the normal return on investments, regardless of any future changes in operating and borrowing costs, as well as demand. Biggar (2011) also reviews research by Schmalensee (2011). Schmalensee shows that, provided the regulator sets the allowed rate of return equal to the firm s true cost of capital, virtually any path of depreciation is consistent with investors receiving a normal return on capital. When it comes to ensuring that investors are adequately compensated, the choice of the path of 1 Options (a) and (b) are equivalent if the cost of capital parameter applied is adjusted appropriately for inflation. 2 An application of option (b) in Australia is known as depreciated optimised replacement cost, commonly abbreviated to DORC. 600182 ver 1 2 Draft as at 04/02/14 9:06
Financial Capital Maintenance depreciation is essentially arbitrary (from an NPV=0 perspective). This point is illustrated in the next section. However, Schmalensee emphasises that this does not necessarily mean that all schedules of depreciation are socially equally desirable. This is because the sequence of depreciation has implications for the allowed revenues and prices. In this respect, Biggar quotes Baumol (1971):... there may be many alternative revenue streams, each of which can give investors their required returns. The choice of depreciation policy may then be defined as the selection of one of these inter-temporal patterns of prices which will yield one of the revenue streams adequate to compensate investors It should now be clear in what sense one depreciation policy may be better than another. One set of inter-temporal patterns of product prices may yield a better allocation of resources than another. According to Biggar (2011), some authors, such as Rogerson (1992), have suggested the choice of depreciation should result in 'inter-temporal Ramsey prices'. The third paper discussed is Crew and Kleindorfer (1992). In this paper, the authors consider the possibility of an emergence of external forces, such as market competition, that may affect the firm's ability to recover its sunk costs in the future. Crew and Kleindorfer (1992) argue that the regulator must take such a possibility into account when setting the schedule of depreciation. If there was an expectation that the regulated firm would be constrained in how much it could recover in the future, a front-end loaded capital recovery, such as a tilted annuity could be applied. The final paper reviewed in Biggar is Salinger (1998). According to Salinger, since cost-based prices are associated with zero economic profit, and there are many streams of future prices that are consistent with such a condition, there is a need to impose further constraints on the path of desired prices. Salinger suggests that if minimising deadweight loss is the ultimate goal, the regulator should first specify those prices that will achieve such a goal (e.g. Ramsey prices), then work backwards to work out the depreciation. Nevertheless, this approach is not supported by Biggar (2011), who points out that such 'price-based costs' will incur high transaction cost and hence may not be practical. Salinger also makes the point that the required revenue is higher when there is a chance that the asset will be stranded in the future. Biggar (2011) recognises that none of these papers addresses how the regulator should set the 'opening RAB' at the time of transition into a new regulatory regime. However, these papers all adopt or are consistent with the FCM principle. The latter is the focus of this paper. 1.3 Implications of financial capital maintenance This section presents a model illustrating the implications for depreciation and the evolution of the asset value if the FCM principle is applied over the full life of the asset. This model is adopted from Diewert, Lawrence and Fallon (2009). To simplify, the focus is on the cost recovery of a single asset. Suppose that at the beginning of period one, the regulated firm completes a project and the total cost is. 3 This asset is expected to yield a stream of services for periods before it is retired. The salvage value and decommissioning costs associated with disposing of the asset are assumed to be zero. 3 Alternatively, the firm may have purchased a long-lived asset that costs at the beginning of period one. 600182 ver 1 3 Draft as at 04/02/14 9:06
Financial Capital Maintenance The regulated firm faces a weighted average cost of capital (WACC) of in period, where. 4 The WACC parameter of a given period can represent either the actual or expected opportunity cost of capital applicable to such a period, depending on the time it is evaluated. For instance, evaluated at the beginning of period one, here is the actual cost of capital in period one, while the rest of represent the regulator's view of future cost of capital. To recover the sunk cost of this investment, the firm is allowed by the regulator to impose capital charges over periods. Period 's capital charge amount is denoted by. One can interpret as the sum of charges paid by all customers in period for services provided by such an asset, excluding charges for the recovery of other costs. Assume that these capital charges are paid at the end of each period. The FCM principle will be met if the period-by-period capital charges satisfy the following fundamental equation: (1) The present value (PV) of future income generated from capital charges is equivalent to the initial cost of the asset. That is, the firm can expect to fully recover its sunk investment over time, including the opportunity cost of capital. Equation (1) shows that if the series of future opportunity cost of capital is reasonably accurate, then the regulator can choose any schedule of capital charges over the asset life that satisfies equation (1). The regulated firm should have no grounds for complaint, as it will recover the cost of its sunk investment in present value terms under this condition. Box 1: Indexation of capital charges Suppose that the regulator decides to index the periodic capital charge for a new asset by the expected inflation rate. Let the anticipated rate of inflation for period be, where. The capital charge in period can be defined as: (a) Using the equation above, the FCM principle will imply: (b) In order to satisfy the NPV=0 principle, the base capital charge has to be set such that equation (b) holds. Once is determined, the capital charge in period is then calculated using equation (a). Regulators typically apply a constant indexation rate, such that the capital charge grows at a constant rate annually. This is akin to the annuity approach explored in Chapter 2. It is useful to convert the choice of paths for capital charges into a choice of paths for asset value or depreciation. If the expected cash flow for period one turns out to be accurate (i.e. there is no significant under- or over-recovery of ), the discounted stream of capital charges that the regulator will allow over periods two to is equivalent to the asset value at the beginning of period two,, and it is defined as follows: 4 Either a real or a nominal discount rate can be applied. If a real discount rate is used, the capital charges calculated will be expressed in real terms, while if a nominal discount rate is applied, the capital charges will be nominal values. 600182 ver 1 4 Draft as at 04/02/14 9:06
Financial Capital Maintenance (2) Period one's depreciation,, can be defined as the decline in the asset value going from the beginning of period one to the beginning of period two (which is also the end of period one): (3) Using equations (1) and (2), the initial asset value, (4), can be expressed as: Equation (4) can now be rearranged, and used together with equation (3) to provide the following equation for period one's capital charge : 5 (5) Thus period one's capital charge can be expressed as times the asset value at the beginning of period one,, plus period one's depreciation amount,. A general formula similar to (5) can be obtained for the capital charge of any period. Define the asset value at the beginning of period, and period 's depreciation,, respectively by equations (6) and (7) below: (6) (7) Note that is defined to be zero (i.e. the asset has no residual value). Period 's counterpart to equation (4) is: (8). Equation (8) can now be rearranged and used together with equation (7) to provide the following formula for period 's capital charge : (9) The last line of equation (9) applies (7) repeatedly. Equation (9) shows that the sequence of capital charges is completely determined by the initial cost of the asset ( ), the sequence of costs of capital ( ), and the sequence of nonnegative depreciation amounts where the sum over periods must equal to the initial value of the asset (i.e. ). 6 5 See Schmalensee (1989, p. 294) and Johnstone (2003,p. 4) for a similar formula. 6 Consider a three-period model where:,, and. Then:. This means that when a nominal discount rate is applied, the compensation for inflation is included in the discount rate, rather than depreciation. However, many regulators (including the QCA) index the RAB value (using the CPI), use a nominal discount rate and then deduct the inflationary gain to avoid double 600182 ver 1 5 Draft as at 04/02/14 9:06
Financial Capital Maintenance Thus, to meet the FCM principle, the regulator has three options: (a) Specify a schedule for capital charges that satisfies equation (1) over the life of the asset. (b) Specify a schedule for depreciation amounts that add up to the initial asset cost over the asset life. The capital charge for each period is then calculated using equation (9). (c) Specify the evolution of the asset value (i.e. the schedule for over the asset life). Since the change in the asset value between periods is the applicable depreciation amount, setting the evolution of the asset value is equivalent to setting the schedule for depreciation. The capital charge is calculated using equation (9). Note that the third line in equation (9) comprises the two capital cost components of the building blocks methodology applied by Australian regulators in deriving the allowed revenue for a particular regulatory year. The product of represents the return on capital, while the depreciation amount is referred to as the return of capital. As a regulated firm has more than one asset, the return on capital is calculated on the firm's regulatory asset base (RAB), which records the initial value of all assets less the total depreciated amount from previous periods. This is equivalent to calculating the return on capital separately for each asset in the RAB and subsequently summing them up. The return of capital is the sum of depreciation amounts for all assets of the firm for a particular regulatory year. It is easier to work with equation (9) in practice than to work with equation (1). Applying equation (9) repeatedly at the beginning of each period (or prior to the commencement of the next regulatory pricing period) requires only the knowledge of the prevailing opportunity cost of capital (or expected cost of capital over the pricing period) and the depreciation amount for this period (or over the pricing period). This means that the regulator will not have to commit ex ante to a single depreciation path each period's depreciation (or depreciation over the regulatory period) can be determined separately as long as at the end of the asset life the sum of the depreciation amounts from all periods equals the initial value. Conversely, working with equation (1) at the beginning of the period where the asset is commissioned requires the knowledge of the entire sequence of expected cost of capital over the life of the asset, which can easily change over time. 7 The use of forecast cost of capital over an extended period can provoke strong disagreements from stakeholders. The regulator is also required to commit to a path of capital charges determined ex ante. Note that the regulated firm will always recover its going opportunity cost of capital if the period-by-period capital charges satisfy equation (1) and the anticipated future period actually materialise. To see this, rearrange the first equation in (9) as follows: (10). This equation shows that the asset value at the beginning of period,, times one plus the prevailing WACC,, is equivalent to the regulator s allowable period 's capital charge,, plus counting of inflation. In this case, the depreciation amounts over the life of the asset will not sum up to the initial asset value,. Rather, the FCM principle requires the sum of depreciation amounts less the nominal asset value gains arising from indexation over the life of the asset to be equal to. This is explored further in Section 2.2. 7 If equation (9) is used and the cost of capital changes over time, then the resulting user charges will not be the same as the original planned sequence of capital charges which satisfied equation (1) using the original sequence of anticipated. 600182 ver 1 6 Draft as at 04/02/14 9:06
Financial Capital Maintenance the asset value of which the regulator will allow the firm to have at the beginning of period. Thus period 's capital charge,, plus the asset value at the end of period,, will be just large enough for the regulated firm to recover its initial asset value plus earn a rate of return of on its investment in this asset the asset will earn the going cost of capital during period. The period-by-period capital charges are largely arbitrary at this point (they need only satisfy equation (1) in order for the firm to recover its cost of capital and the sunk cost associated with the investment). Similarly, the period-by-period depreciation amounts are arbitrary (except that they must sum to the initial asset value over the asset life). This proposition was demonstrated in Schmalensee (1989). At this point, if the allowed rates of return are 'fair' from an investor's perspective and the investors are not concerned about stranded assets, then the regulated firm has no particular incentive to contest whatever pattern of periodic capital charges or depreciation that the regulator chooses. However, it is unlikely that the will in fact correspond exactly to appropriate weighted costs of capital, and thus depending on whether the allowed rates are too high or too low, the firm will have incentives to either postpone depreciation (the allowed is too high) or to lobby for accelerated depreciation (the allowed is too low). In reality, the regulator s choice problem is also typically much more complex, particularly where there is scope for asset stranding or excess capacity. 1.4 Summary In the regulatory context, the FCM principle is equivalent to the NPV=0 principle, where the investors of a regulated firm can expect to earn a series of regulated returns with a present value (discounted at the appropriate cost of capital) equal to the initial capital investment. The model shows that, given the FCM principle is satisfied, the regulator should be indifferent in terms of specifying a sequence for capital charges (to cover return on and return of capital), a sequence for depreciation, or the evolution of the asset value. This point has been made in the existing literature. However, in practice, it is easier to apply the building blocks approach, where the allowed cost of capital can be updated at the beginning of the pricing period, and the depreciation amount for each period can be determined separately (the regulator does not have to commit to a path of capital charges or depreciation). Such an approach is commonly applied by Australian regulators. The existing literature also highlights that while there can be many schedules of depreciation amounts (capital charges or the evolution of asset value) that meet the FCM principle, this does not necessarily imply that all paths are equally desirable. Some paths produce more efficient outcomes than the others, but at the same time they may not be considered as socially equitable as others. These factors should be taken into account when the regulator determines the appropriate series of allowable returns for the investors. 600182 ver 1 7 Draft as at 04/02/14 9:06
Building Blocks Approach 2 BUILDING BLOCKS APPROACH This chapter presents four standard conventional methods for setting the depreciation profile of an asset, as well as the annuity approach. This is followed by a discussion of the method for calculating capital charges when the RAB value is indexed by the inflation rate and nominal discount rates are used, where adjustments to the return on capital are required to avoid double counting of inflation. 2.1 Return on and return of capital In Australia, regulated prices are commonly set to allow the regulated firm to recover the annual revenue requirement (ARR). The ARR is a series of notional annual revenue allowances, and is calculated by using a building blocks approach. The allowed building blocks revenue for a particular regulatory year is made up of the following components: (a) (b) (c) (d) return on capital return of capital efficient operating costs taxation allowance. As most regulated firms tend to be capital-intensive, the largest components of the building blocks ARR are typically the return on and return of capital. These two components, amounting to the total capital charge for the RAB in a regulatory year, enable the firm to recover its sunk capital investments. As described in Chapter 1, the return on capital is calculated on the firm's RAB, while the return of capital is the sum of all assets' depreciation amounts. There are four common approaches for setting the depreciation profile of an asset: (a) (b) (c) (d) Straight-line where the initial value of an asset is recovered in equal annual amounts over its productive life. For a given asset, straight-line depreciation is calculated by dividing the prevailing value of the asset (i.e. the initial value less the sum of depreciated amounts from previous periods) by its remaining productive life. Front-end loaded where the amount of depreciation decreases progressively for each year of the productive life of an asset. Back-end loaded where the amount of depreciation increases progressively for each year of the productive life of an asset. One-hoss shay where the value of an asset is maintained until the end of its productive life, at which the value declines significantly or to zero. That is, there is no deprecation until the final year of the life of the asset. Alternatively, rather than setting its depreciation profile, the regulator can specify a schedule of periodic capital charges with a present value equal to the initial asset value. Recall that the FCM principle provides substantial flexibility for the regulator to specify a sequence for either capital charges or depreciation, as long as over the life of the asset the NPV=0 principle is satisfied. One way of calculating the capital charge streams is to apply the annuity approach, which produces a series of constant or indexed capital charges applicable over the asset life. Each periodic amount embodies both the return on and return of capital components for such an asset. In the case where the annuity approach is applied to some but not all assets, the return 600182 ver 1 8 Draft as at 04/02/14 9:06
Building Blocks Approach on and return of capital will only be calculated separately on the assets where the annuity approach is not applied. This means that a record of asset values, adjusted for previous allowed depreciation, such as the RAB, will no longer be required for pricing purposes if the annuity approach is applied to all assets. A simple example below illustrates the application of the four standard depreciation methods as well as an annuity. It relies on the following assumptions: (a) All values are expressed in real terms. (b) Initial purchase price of asset = $1000. (c) (d) (e) Asset life = 5 years. Allowed real rate of return = 10 per cent. Capital charges are received at the end of each period. The different outcomes are shown in the Table 1. The present value is evaluated at the beginning of Year 1. Table 1 Comparison of recovery of capital Method Year 1 Year 2 Year 3 Year 4 Year 5 1. Straight-line depreciation Opening asset value $1000 $800 $600 $400 $200 Depreciation $200 $200 $200 $200 $200 Return on capital $100 $80 $60 $40 $20 Capital charge $300 $280 $260 $240 $220 Present value $1000 2. Front-end loaded depreciation Opening asset value $1000 $700 $450 $250 $100 Depreciation $300 $250 $200 $150 $100 Return on capital $100 $70 $45 $25 $10 Capital charge $400 $320 $245 $175 $110 Present value $1000 3. Back-end loaded depreciation Opening asset value $1000 $900 $750 $550 $300 Depreciation $100 $150 $200 $250 $300 600182 ver 1 9 Draft as at 04/02/14 9:06
Building Blocks Approach Method Year 1 Year 2 Year 3 Year 4 Year 5 Return on capital $100 $90 $75 $55 $30 Capital charge $200 $240 $275 $305 $330 Present value $1000 4. One-hoss shay depreciation Opening asset value $1000 $1000 $1000 $1000 $1000 Depreciation $0 $0 $0 $0 $1000 Return on capital $100 $100 $100 $100 $100 Capital charge $100 $100 $100 $100 $1100 Present value $1000 5. Constant annuity Annual charge $264 $264 $264 $264 $264 Capital charge $264 $264 $264 $264 $264 Present value $1000 The streams of capital charges produced under the four methods of depreciation and a constant annuity are all equivalent in a present value sense, hence satisfying the FCM principle. Regardless of the depreciation method applied, depreciation over five years must total $1000 in order to satisfy the NPV=0 principle. Under the depreciation approach, the capital charge for each year is calculated using equation (9). Given that customers are liable for capital charges, not just the depreciation allowances, it is important to take into account the resulting paths of capital charges under varying methods of depreciation. For example, straight-line depreciation does not lead to a path of constant capital charges, but rather the capital charge decreases over time, as shown in Table 1. This has implications for pricing, and is discussed further in Chapter 3. In the example above, a constant annuity has been applied (fifth scenario in Table 1). The annuity approach provides the flexibility for the annual charge to be defined to have any growth patterns. Suppose the regulator decides to escalate the annual charge at a constant rate of 20 per cent per annum. Then, the capital charge for period will be defined as: where. Now substitute the equation above into equation (1) to get the following equation: A base annuity charge of approximately $153 solves the equation above. Hence, the annual charges will be as follows: 600182 ver 1 10 Draft as at 04/02/14 9:06
Building Blocks Approach Year 1 Year 2 Year 3 Year 4 Year 5 Annual charge $183 $220 $264 $317 $380 The discounted stream of annual charges above is equivalent to $1000. Compared with the constant annuity in Table 1, the annual charges of this indexed annuity are lower in the first two years, the same in year three, and higher in the final two years. 2.2 Indexation of the asset value Thus far, indexation of the asset value has not been considered. The equations in Chapter 1 and the example above assume no indexation of the asset value (the change in the asset value between periods is simply the depreciation amount). However, in practice the RAB value is typically adjusted for inflation. Furthermore, nominal values are commonly used as it is more convenient to report building blocks components such as operating costs as well as prices on a nominal basis. The QCA's approach is to index the RAB value by the CPI inflation rate and use a nominal WACC to calculate the return on capital. An adjustment is then made to eliminate the double counting of inflation that would otherwise occur from the use of an asset value indexed for inflation and a nominal rate of return that also includes compensation for inflation. Over the full life of an asset, the PVs of capital charges will be identical if the calculations are done correctly. Table 2 presents a comparison of nominal capital charges under two scenarios: with (the QCA's approach) and without indexation of the asset value respectively. Both apply straight-line depreciation. The asset value is $1000, with a productive life of five years, a residual value of zero, an inflation rate of 2.5 per cent and a nominal WACC of 12.75 per cent. The indexation rate for the asset value is the inflation rate. Table 2 Comparison of capital charges under two scenarios Scenario Year 1 Year 2 Year 3 Year 4 Year 5 1. Without indexation of the asset value Opening asset value $1000 $800 $600 $400 $200 less Depreciation 8 $200 $200 $200 $200 $200 Closing asset value $800 $600 $400 $200 $0 Return of capital (depreciation) $200 $200 $200 $200 $200 Return on capital $128 $102 $77 $51 $26 8 To calculate the depreciation amount for a given year, divide the opening asset value in that year by the length of the remaining asset life. If the asset value is escalated for inflation, include the inflationary gain amount in the opening asset value before performing the division. 600182 ver 1 11 Draft as at 04/02/14 9:06
Building Blocks Approach Scenario Year 1 Year 2 Year 3 Year 4 Year 5 Capital charges $328 $302 $277 $251 $226 Present value (12.75% nominal WACC) $1000 2. With indexation of the asset value (the QCA's approach) Opening asset value $1000 $820 $630 $431 $221 add Inflationary gain $25 $21 $16 $11 $6 less Depreciation $205 $210 $215 $221 $226 Closing asset value $820 $630 $431 $221 $0 Return of capital (depreciation) $205 $210 $215 $221 $226 Return on capital $128 $105 $80 $55 $28 less Inflationary gain $25 $21 $16 $11 $6 Capital charges $308 $294 $280 $265 $249 Present value (12.75% nominal WACC) $1000 When there is no indexation of the asset value (first scenario in Table 2), the total capital charge is calculated using equation (9) the return on capital is the product of the opening asset value and nominal WACC, while the return of capital is the specified depreciation amount. Note that each year's depreciation amount is $200, and over the life of the asset they add up to the initial asset value. In contrast, under the QCA's approach (second scenario in Table 2), an inflationary gain component is added annually to the opening asset value. This component, amounting to the product of the opening asset value and inflation rate, ensures that the asset value is indexed by the inflation rate. For example, in year one the inflationary gain is $25. The FCM principle requires the sum of depreciation amounts less the inflationary gains over the life of the asset to be equal to $1000, as above. When the asset value is escalated annually by the inflation rate, and a nominal rate of return is applied, it is necessary to adjust the return on capital (the product of the opening asset value and nominal WACC) by deducting an amount equivalent to the gain in the RAB value from indexation. This avoids the double counting of inflation that would otherwise occur from indexing the capital base by inflation and applying a nominal rate of return that embodies the same inflation rate. The downward adjustment in the first year is $25. Note that straight-line depreciation has been applied in both scenarios, but there are some technical differences between them. When there is no indexation of the asset value, the 600182 ver 1 12 Draft as at 04/02/14 9:06
Building Blocks Approach nominal depreciation allowance remains at $200 over the five-year period. Under the QCA's approach, the nominal depreciation allowance grows at a rate of 2.5 per cent per year, but the real depreciation allowance is a constant real value of $200. In this example, the indexation rate is also the CPI inflation rate, and this has resulted in a path of annual depreciation allowances that maintain the same value in real terms over time. Given both scenarios follow different paths of depreciation, their respective paths of return on capital (after deducting the inflationary gain under the QCA's approach) are also different. Nevertheless, as seen above, the PVs of capital charges are identical under both scenarios, and they are equal to the initial value of the asset ($1000). To be consistent with the normal practice and for expositional convenience, the QCA's approach is adopted for the remainder of this paper. 2.3 Summary This chapter presents four standard methods of depreciation, and also the annuity approach. The four standard methods of depreciation are: (1) straight-line; (2) front-end loaded; (3) backend loaded; and (4) one-hoss shay. Regardless of the depreciation method applied, the depreciation amounts over the life of the asset must sum to the initial asset value in order to satisfy the NPV=0 principle (when there is no indexation of the asset value). The annuity approach produces a series of constant or indexed capital charges applicable over the asset life. Each periodic amount resembles both the return on and return of capital for a particular asset. The annuity approach provides the flexibility for the annual capital charge to be defined to have any growth pattern. This chapter also introduces the method for calculating capital charges when the asset value is indexed. The QCA's approach is to index the RAB value by the CPI inflation rate and use the nominal WACC to calculate the return on capital. Under such an approach, it is necessary to adjust the return on capital by deducting an amount equivalent to the change in the RAB value from indexation. 600182 ver 1 13 Draft as at 04/02/14 9:06
Price Smoothing 3 PRICE SMOOTHING This chapter explains how price or revenue smoothing is typically implemented in a regulatory context. Depending on the estimates of the building blocks ARR and forecast demand, prices can fluctuate significantly between years without smoothing. The smoothing technique is used to achieve a smooth price (or revenue) path within a regulatory period. However, smoothing per se does not fully address the issue of price (or revenue) fluctuations over a longer time frame then the current regulatory period. As smoothing is typically limited to the building blocks ARR within each regulatory period, there may be abrupt changes in prices (or revenues) from one period to the next. The choice of depreciation method adopted also has implications for smoothed prices (or revenues).additionally, should demand grow over time, prices or (revenue) may show increased fluctuation as a result. 3.1 Smoothing The previous chapter explained the building blocks approach typically applied by Australian regulators for calculating the ARR. Once the ARR has been determined, prices can then be generated. There are two approaches for calculating prices: (a) (b) Building blocks prices dividing the building blocks revenues by their respective forecast annual volumes. Depending on fluctuations of the building blocks revenues and forecast demand, prices determined in this manner can vary significantly from year to year. Smoothing in order to counter fluctuations described above, the regulator can either apply a smoothed price or revenue path for the regulated firm. Smoothed revenues (or prices) may be subject to a defined rate of escalation (e.g. the expected inflation rate), and are calculated on the basis that the present value (PV) of smoothed revenues (or revenues arising from smoothed prices) is equivalent to the PV of the building blocks revenues. However, if smoothed revenues are applied, resulting prices can still fluctuate between years depending on the forecast demand. Likewise, forecast demand patterns may result in fluctuating revenues despite price smoothing.. The derivation of smoothed revenues and prices is presented below. The focus is on price smoothing as the calculation of smoothed revenues is relatively simple. The formulae below are applicable to both nominal and real analyses. For a nominal analysis, smoothed prices are calculated using a nominal discount rate, while for a real analysis they are calculated using a real discount rate. Calculation of constant prices This subsection demonstrates the derivation of constant prices. The starting proposition is revenue equivalence the PV of revenues arising from smoothed prices (denoted by ) must equal the PV of the building blocks allowed revenues (denoted by ). Consider the following: (11) (12) 600182 ver 1 14 Draft as at 04/02/14 9:06
Price Smoothing where: = building blocks revenue for period = constant price = forecast volume for period = total length of the smoothing period = period within the smoothing period, where = discount rate (real or nominal) applicable. Note that the price,, is fixed over time in this case. The present value equivalence proposition can be expressed as follows: (13) (14) The constant price is then calculated as: (15) The fixed price is derived by dividing the PV of the building blocks revenues by the PV of the forecast demand. If the building blocks revenues and the discount rate are expressed in real (nominal) terms, the resulting prices will also be in real (nominal) terms. Note that if nominal prices are constant over time, they will be declining in value in real terms as a result of inflation. To calculate a path of constant revenues, simply replace the term ( ) in equation (12) with a constant revenue term (denoted by ), and solve for this term. It can be shown that the constant revenue is calculated as: (16). The annual prices are then calculated by dividing the constant revenue term by their respective forecast annual volumes. The prices in this case can fluctuate significantly depending on the forecast demand. Calculation of smoothed prices escalated at a specific rate This subsection illustrates how prices can be defined to escalate annually at a specific rate. In practice, prices typically rise annually at the expected percentage change in the CPI to reflect the general movement in prices within the economy. Similar to the case of constant prices, the starting proposition is that the PV of the revenues arising from smoothed prices must equal the PV of the building blocks revenues. Consider the following: (17) (18) where: = building blocks revenue for period = base price 600182 ver 1 15 Draft as at 04/02/14 9:06
Price Smoothing = forecast volume for period = total length of the smoothing period = price escalation rate = period within the smoothing period, where = discount rate (real or nominal) applicable. Note in equation (18) the annual price is set to grow at the rate of each year. Following similar steps illustrated in the previous subsection, the base price can be derived as: (19) This can be re-expressed as: (20) The indexed price for period can then be expressed as follows: (21) Note that if the escalation rate is set at 0%, equations (15) and (19) are identical the smoothed price for each period is simply the base price and hence is constant over time. To calculate an indexed revenue stream, replace the term, and solve for. Then the base revenue is: in equation (18) by, and the smoothed revenue for period is: (22). The annual prices are then calculated by dividing the smoothed revenues by their respective forecast annual volumes. Similar to the case of constant revenues, prices can fluctuate significantly depending on the forecast demand. Prices in nominal terms escalated at the inflation rate In a nominal analysis, where nominal values and nominal discount rate are applied, equation (19) can be further simplified if the price escalation rate is set to equal the expected percentage change of the CPI. Firstly, note that the Fisher equation specifies the following relationship between nominal and real interest rates:, where is the nominal discount rate, is the inflation rate, and is the real discount rate. Given the Fisher equation, if the escalation rate is the inflation rate, and is the nominal discount rate, equation (19) can be re-stated as follows: (23), where the term has been replaced by. 600182 ver 1 16 Draft as at 04/02/14 9:06
Price Smoothing Note that the term appears in both the denominator and the numerator in equation (11). Hence, the formula can be simplified to be: (24). This shows that when nominal values are applied and smoothed prices escalate at the expected inflation rate ( ), the real discount rate should be applied to discount the forecast demand in a nominal analysis (but note that the nominal discount rate is still being used to discount the nominal building blocks revenues). Prices escalated at the discount rate In contrast, if the escalation rate is set to be equal the discount rate equation (19) then becomes: (real or nominal), (25). In this special case, there is no need to discount the stream of volumes. The denominator of the base price formula is simply the summation of the forecast volumes over the smoothing period. 3.2 Smoothing period Smoothing period (also known as planning period) refers to the period used for smoothing of revenues or prices. Denoted by in the smoothing formulae, this parameter effectively determines the number of years of building blocks revenues that are included in the applicable smoothing formula to generate a base price (or revenue). In this paper, target prices (or revenues) refer to the final allowed prices (or revenues) adopted by the regulator. This is to distinguish between the smoothed prices (or revenues) calculated and the actual allowed prices (or revenues) adopted. Consider the following example of price smoothing. Suppose that the regulatory pricing period is five years and the building blocks revenues for the next 10 years have been specified. In determining the regulated prices for the next 10 years, the regulator can apply a 10-year smoothing period. In this case, the building blocks revenues and forecast volumes for this period will be used to produce a single base price, which is then escalated annually to generate prices for the next 10 years. Under this approach, target prices for the two successive pricing periods are made up of a single smoothed path. Alternatively, the regulator can apply a five-year smoothing period. This involves firstly applying the applicable smoothing formula to the building blocks revenues and forecast volumes for the first five years; then repeating it for years six to 10. This will produce two base prices, in which the first is escalated to derive prices for the first five years, and the second is used to determine prices for the remaining years. Under this approach, target prices for the 10 years consist of two separate smoothed paths. As shown below, the smoothed prices (or revenues) under varying smoothing periods can be significantly different, depending on the underlying building blocks revenues the smoothing applies to. While the smoothing period applied by the regulator may be equal or greater than the regulatory pricing period, in practice this choice is constrained by the number of years of building blocks revenues available to the regulator. Regulators typically only forecast building blocks revenues for the current regulatory period because of the risk of unavailability of data needed for forecasts in the long term. 600182 ver 1 17 Draft as at 04/02/14 9:06