Cost of Capital

Methodology for calculation of weighted average cost of capital- WACC Table of contents 1 Introduction 1.1 1.2 1.3 1.4 2 Why the AEC is considering the cost of capital? The significance of cost of capital to the firm and its investors The need for this notice requesting comments on the consultation paper Structure of this document 1 1 1 2 2 4 4 5 5 5 9 9 10 13 21 21 22 23 23 24 24 26 27 27 28 28 29 29 34 Operators for which costs of capital must be calculated 2.1 Operators requiring cost of capital calculations 3 Theoretical principles underpinning cost of capital 3.1 3.2 What does the cost of capital represent? Weighted Average Cost of Capital (Part 1 and Part 2) 4 PART 1 of WACC calculation - Cost of equity 4.1 4.2 4.3 Choosing an appropriate model Capital Asset Pricing Model (CAPM) – AEC’s preferred model Factors of the CAPM equation requiring derivation 5 Possible adjustments to CAPM for emerging markets 5.1 5.2 Testing the assumptions of the traditional CAPM in an emerging market Models to adjust the traditional CAPM to make it relevant in an emerging market 6 PART 2 of WACC equation - Pricing debt 6.1 Factors affecting the cost of debt 7 Factors affecting calculation of WACC in Macedonia 7.1 7.2 7.3 7.4 7.5 7.6 Risk Free Rate Market risk (Equity Risk Premium) Gearing for AD Makedonski Telekomunikacii (MT) Beta for AD Makedonski Telekomunikacii (MT) Debt price for AD Makedonski Telekomunikacii (MT) Corporate Tax rate 8 Aggregate entity level cost of capital 8.1 Calculating the WACC of AD Makedonski Telekomunikacii (MT) Annex 1 – Cost of capital models in an emerging country A1 2 1 Introduction This notice is issued pursuant to Article 105 of the Law on Electronic Communications which requires that, the Agency for Elecotronic Communications (AEC) shall undertake the following duties and responsibilities:    To take appropriate account of opinions of interested parties Publish the proposed instrument before recommending or adopting regulations Make publicly available the obtained opinions and comments 1.1 Why the AEC is considering regulating the cost of capital? Cost of capital represents the financial return on a group of assets, which are employed by a company or business. From a regulatory perspective, licensed telecommunications operators are entitled to a reasonable or fair rate of return within the prices of services they offer whether these are retail or wholesale services. The cost of capital calculation is the most appropriate method of calculation for a reasonable rate of return and is particularly applicable to wholesale services such as interconnection services provided by operators with market power that are therefore under regulatory scrutiny. The regulator has a duty to ensure the price for interconnection services provided by such operators is cost based. The cost of capital of a firm represents an additional cost element within the calculation of the costs of provision of interconnect services. Where an operator is deemed to be in a dominant position, additional regulatory provisions may also be necessary at both the retail level and the wholesale level. Where a dominant operator provides essential facilities to other licensed operators, the price at which such facilities are supplied should in theory replicate the long run incremental cost of an efficient operator. This provision equally applies to a dominant operator‟s cost of capital. The cost of capital should represent a fair return to investors for the provision of capital to the firm. New entrants and existing users of a dominant operator‟s facilities should not be required to over compensate the investors of the dominant operator, which would ultimately feed into higher consumer prices. The regulator therefore has a duty to ensure that the cost of capital used in the calculation of costs / prices of services provided to either end users in the case of retail services or wholesale services in the case of access or interconnect services represents the true efficient / optimal cost of capital. 1.2 The significance of cost of capital to the firm and its investors In calculating the cost of provision of services or elements within an overall service, a business requires assets which are employed by the firm to buy and transform economic inputs into a final or intermediate product. These assets represent an opportunity cost to the firm. If they were not employed for these purposes, they could be sold and the cash used to invest in some other project. Cash is therefore not free, it comes at a price. The price is the cost to the firm of using investors‟ money. The cost of capital represents the return expected by the investors for the capital they supply. Investors do not normally invest directly in projects - they invest in the firms that undertake projects. The return expected from the assets managed by a firm must be the total of the returns expected by debt holders and equity holders, weighted by their respective contribution to the financing of these assets. 1 In normal circumstances a business must seek to make a return on the capital that is actively employed by the business that is at least equal to the recurring cost of that capital. This ensures that investors in the business can be properly compensated for the risk that they take for supplying capital to the business. A return that is significantly greater than the cost of capital represents super-profit and is usually associated with a market structure where entry is restricted or where the market is not contestable, and which therefore is not in the best interests of consumers. Because cost of capital must recognise the expectations of investors, its calculation or measurement must consider factors which are external to the business itself. This can lead to some complexity. It is often the case that a large business actually comprises several individual business streams that have differing risk profiles and thus different costs of capital. Many regulators historically, only calculated a single entity level cost of capital, whilst still acknowledging that the risk profile of different business units within the firm varies significantly. Recently however, many regulators have attempted to calculate the different cost of capital applicable for these different business units. 1.3 The basis and need for this consultation paper This document discusses the principles to be used in the calculation of the cost of capital for licensed operators within Macedonia, who are obliged to provide cost based interconnection services and where the regulator deems necessary, other access, wholesale or retail services. The paper, having discussed these principles foundations for the derivation of cost of capital, provides a calculation of the cost of capital for AD Makedonski Telekomunikacii (MT), as the operator who at this present time, is deemed to have a dominant position within the market. The document considers both the methodology to be used in calculating cost of capital and calculations that relate to AD Makedonski Telekomunikacii (MT). Comment is sought from all interested parties on specific areas of the methodology and the resulting calculations and conclusions. Cost of capital theory is a specialist technical subject, much of which is based on interpretation of historic data, and is subject to a range of plausible estimates. It is not the purpose of this document to provide a detailed discussion on the theoretical models underpinning the cost of capital and its components. In presenting our findings, we have assumed that readers of this document have a basic understanding of the theory underpinning cost of capital and how the cost of capital and its components are derived. 1.4       Structure of this document Section 1 introduced the concept of cost of capital and requirements that lead AEC to address such matters now; Section 2 Lists the current operators which are required to use a regulated cost of capital and thus require a cost of capital calculation; Section 3 provides an explanation of the theory underpinning the calculation of cost of capital, and introduces the concept of WACC; Section 4 discusses the first part of the WACC equation: the cost of equity, and in particular provides a detailed examination of the CAPM methodology; Section 5 discusses the CAPM fundamental building blocks and the applicability of CAPM within an emerging market; Section 6 discusses the second part of the WACC equation: the cost of debt; The paper is structured in the following manner: 2   Section 7 elaborates on the particular factors of the WACC equation in the context of Macedonia; Section 8 calculates the aggregate WACC for AD Makedonski Telekomunikacii (MT); 3 2 Operators for which costs of capital must be calculated This section discusses the requirement for cost of capital calculation for licensed operators within Macedonia. 2.1 Operators requiring cost of capital calculations The Law on Electronic Communications, (Official Gazette of Republic of Macedonia No. 13/2005), states that all licensed operators, which have a dominant position within a relevant market must provide interconnection services which are cost oriented, and further states by July 2007 that such services must be priced based on a LRIC basis. This would imply that an appropriate cost of capital should be used by such an operator when calculating their respective interconnection charges. The AEC within this consultation paper is only considering the case of AD Makedonski Telekomunikacii (MT) as an operator who is presently deemed to possess market power, and therefore have significant impact upon the market. The principles determined within the consultation will determine the methodology and approach that must be taken when calculating the cost of capital for all other licensed operators within Macedonia, that may be required to provide cost based interconnection services at some future date determined by the AEC. 4 3 Theoretical principles underpinning cost of capital This section discusses the theory behind the cost of capital, how it is calculated in practice in many other jurisdictions, assumptions that are inherent in the models proposed and possible changes that may be required to models in light of those general assumptions not holding true for Macedonia as an emerging market. 3.1 What does the cost of capital represent? The cost of capital that a firm faces represents the equilibrium return investors expect from investing in a firm with a specific set of risks. The risks that an investor faces, in addition to market risks, are influenced by the ratio of debt versus equity that comprises the capital structure of the company. Debt, by virtue of the fact that it has a higher priority on claims, in the event that a firm goes into bankruptcy, in addition to normally having fixed interest payments, implies a lower risk for lenders than for holders of equity, who face higher levels of uncertainty and lower priority on claims in the event of liquidation. Interest payments on debt are generally deductible for the purposes of corporate tax calculation and hence a profitable company can benefit from the effects of these tax shields in improving the value of the resulting equity. It may therefore be advantageous for a profitable company to hold a reasonable level of debt in order to maximise this effect without causing undue risk to the business. Accordingly, in its considerations of the cost of capital, it becomes necessary for the AEC to consider matters such as the debt/equity ratio that should apply and the risks and costs that arise from employing the particular mix of debt and equity. These issues are considered within the umbrella of the Weighted Average Cost of Capital (WACC) concept. As the term suggests it is necessary to determine both the cost of debt and the cost of equity; the latter also varies with the level of debt within the capital structure of the company as the fundamental business risk is allocated between equity and debt holders. The AEC would like to stress that it is not attempting to control or recommend a particular type of financing arrangement for a particular operator. These are matters entirely within an operator‟s own control. The AEC is however concerned that a particular financing arrangement may unduly increase the cost of services charged to other interconnecting parties, such a prospect would theoretically not arise within a competitive market. The AEC is therefore proposing using a particular mix of equity and debt in far as calculating the price of regulated services is concerned. 3.2 Weighted Average Cost of Capital (Part 1 and Part 2) The combined level of returns expected, by definition also represent the combined level of risk between debt and equity - known as the Weighted Average Cost of Capital (WACC). WACC is the internationally accepted basis for the calculation of cost of capital by regulators, financial institutions, businesses and the academic community. The WACC model makes certain fundamental assumptions about the nature of the environment that the WACC is been calculated for. It is therefore important to understand these assumptions in determining the appropriateness of the WACC model for a regulated telecommunications business within a particular country. One of these fundamental assumptions include an assumption that the firm has a long-term optimal, constant debt / equity (D/E) ratio and does not face any floatation costs or the effect of subsidized financing. 5 It is important to emphasis that the AEC is using a forward looking efficient approach to the calculation of the cost of capital. In doing this, the AEC must use parameters that would be observed within an efficient competitive market. Where an operator is competing within a competitive market, where there are no perceived market power advantages and where the market can be considered to some extent contestable, the operator must ensure its cost of service provision is minimised if it is to continue to be profitable. Faced with the challenges of a competitive market, a firm will be forced to use an optimal financing mix that lowers its cost of capital. In making a cost of capital determination for regulated services, it would not be economically justifiable to simply use the current financing mix of operators, as this may not represent the optimal mix that would be observed in a competitive market. A market structure with a limited number of players allows an operator to choose a financing mix, which is may not be optimal and using the status quo would force interconnecting parties (and indeed end users) to pay for what may be inefficiencies of the operator. As debt finance can have tax advantages over equity finance (because interest payments, unlike dividends, are normally a tax deductible expense for a company), it is possible to reduce the overall cost of capital by switching from equity to debt. Higher gearing will increase the firm's equity cost (due to increased volatility of equity earnings) but over a certain range this will be more than compensated for by the benefits from tax shields obtained with debt. When the cost of capital calculation is likely to be used for setting interconnection rates or price caps, the regulator has an incentive to use a cost of capital based on efficient or optimal capital structure. Assuming that the firm is financed by debt and equity, its WACC is equal to the weighted average of the cost of these two means of financing. Weights are equal to the relative proportions of debt and equity used in financing the firm‟s assets. This is explained by the formulae: WACC  re Where: re rd E D V tc E D  rd (1  t c ) V V = return on equity = return on debt = market value of equity = market value of debt = market value of Firm (D+E) = marginal corporate tax rate It is usually the post tax WACC that is calculated. It represents the return required for investors to take on the risks of investing in the company. The pre tax WACC reflects the returns that the company must earn to be able to pay investors the post tax WACC and finance tax liabilities. The cost of debt can be estimated if the regulated entity has issued traded debt securities. Estimating the cost of equity capital raises more complex issues. Calculating the cost of equity requires a forward looking view of the expected return, it is normally estimated by reviewing actual, or ex-post returns which can fluctuate substantially from one year to the next. It must be appreciated that the basic form of WACC is appropriate where there are no market imperfections and the investment opportunity set is static- market imperfections may take a variety 6 of forms. Such market imperfections require an assessment of their effects as they influence the validation of the basic model. Such imperfections include: 1. Regulatory uncertainty; 2. Financial distress costs; 3. Information asymmetries; 4. Funding constraints; and 5. Imperfect capital markets These factors are each discussed below. 1. Regulatory uncertainty The possibility of the introduction of additional regulatory measures in the future may cause investors to require a return on long term investments which is in excess of the long term equilibrium WACC. In the case of AD Makedonski Telekomunikacii (MT), allowance for additional regulatory risk has been made, as it will be seen later, by increasing the historic beta. For new operators, these issues may be of more importance. 2. Financial distress AD Makedonski Telekomunikacii (MT) is profitable and appears to have no indications of financial distress, and as such there does not appear to be a necessity to incorporate additional distress risk allowance. For new operators, these issues may be of more importance. 3. Information asymmetries Within the context of “Regulation” as is within many “principal - agent” situations, there is likely to be information asymmetry and the regulator must base its decisions on the information available at that time. It is because of these information asymmetries that a consultation with all stakeholders is essential. The AEC invites all participants to provide comments including those which may be considered confidential (which should be provided separately and marked as being confidential) so as to help the AEC make decisions on the cost of capital of AD Makedonski Telekomunikacii (MT) with full information at hand. 4. Funding constraints For established operators, particularly those already part funded by international organisations, it is unlikely that acquiring funding at “market rates” will be particularly difficult. For new operators, these issues may be of more importance. 5. Imperfect capital markets There are very few capital markets which are considered as being highly efficient and integrated to other world financial markets. Those that are considered to be close proxies to efficient capital markets are primarily based in the USA or Western Europe. Certain countries in Eastern Europe are becoming increasingly integrated to the world market. Within Macedonia and certain other states it is difficult to claim that these markets are integrated. The level of integration of a particular country with world markets include analysis based on indicators such as the market capitalisation of the local index, the relative integration of local financial markets with global markets and the ratio of market capitalisation to GDP. It may therefore be necessary to adjust the traditional basic form of the cost of equity calculations, or allowance be made for the use of a higher factor within a developing country context, to allow for possible imperfections within the financial market of an emerging market. 7 These issues will be further elaborated subsequently in the context of the Macedonian market in section 5. The next three chapters will discuss the two fundamental elements of the WACC calculation: 1. the cost of equity (section 4) and how it may need adjustment in the context of an emerging market (section 5); and 2. the cost of debt (section 6) 8 4 PART 1 of WACC calculation - Cost of equity This section discusses the first part of the WACC calculation; that is the cost of equity that a firm has within its capital structure. There are numerous methods that have been formulated over time to assess this cost. This section discusses these briefly and proposes to use a model which appears appropriate for Macedonia. This model is then described in detail and methods of derivation of factors within the model are explored in some detail. The section forms an important base upon which adjustments to international data can be made to accommodate for Macedonia specific factors, which are subsequently determined in section 7. 4.1 Choosing an appropriate model There are various methods that can be used to price equity; some are theoretically stronger than others. Financial assets are acquired for the cash flows expected from owning them. Consequently perceptions of value have to be backed up by reality, which implies that the price paid for any asset should reflect the cash flows it is expected to generate, and the risks involved in undertaking this obligation. The most simplistic method of establishing a cost of equity is the dividend discount model (DDM). According to the model, the price of a share should be equal to the present value of the stream of future cash dividends discounted at the firm‟s cost of equity. The dividend discount model cannot be used to solve the cost of equity, unless simplifying assumptions regarding the dividend growth rate are made. A firm‟s cost of equity is the sum of its expected dividend yield and the expected dividend growth rate. If it is assumed the dividend that a firm is expected to pay next year will grow at a constant rate forever, then the cost of the equity will be: re  Div1 G P0 = return on equity = dividend payment in period 1 = price of asset in period 0 = growth of dividends in the future Where: re Div1 P0 G For the vast majority of companies, the simplistic assumptions underlying the reduced version of the dividend discount model are unacceptable, thus, an alternative approach is needed. There are four broad alternative approaches to the pricing of equity: 1. Relative valuation by which the value of an asset is estimated by looking at the pricing of comparable assets relative to a common variable such as earnings, cash flows, book value; 2. Arbitrage Pricing Theory (APT), allows the actual return R(i) on asset-i to be influenced by a number of market-wide variables or “factors”, such as interest rates, exchange rates etc; 3. Capital Asset Pricing Model (CAPM), which relates the value of equity to the implied risk investors must bear, and is effectively a shorter form of the APT model; 4. Contingent claim valuation, which uses options pricing models to measure the value of assets with share option characteristics. 9 There is much academic material that explores the relative merits of the methods. This is accordingly not reproduced in this document. The general consensus amongst academics and in particular regulatory bodies is that the CAPM is the most suitable at this present time. However, recently a number of academics have suggested that the CAPM is not entirely relevant for emerging markets, although have accepted the fundamental idea of the model. These academics have advanced that the APT could be more representative of factors of relevance in an emerging market, as fundamental drivers with the APT are macroeconomic level factors. The AEC believes however that in emerging markets, macroeconomic level data are not readily collected, and where it is, it suffers from being incomplete, short and volatile. The importance of the CAPM is still found to be widely used within emerging markets. Based on the academic merits, proven track record and availability of data to implement the method, the AEC is of the opinion that the CAPM is the most appropriate method to be used for the calculation of the equity price to be used at present, although accepts that it may need minor adjustments in the context of Macedonia being classified as an emerging country. 4.2 Capital Asset Pricing Model (CAPM) – AEC’s preferred model The CAPM is based on portfolio theory, which recognises that investors are broadly risk averse and seek to limit the impact of exposure to the risks associated with individual businesses by creating a diversified investment portfolio. The theory is based on the Markowitz Mean-Variance model. The model assumes: 1. Investors are risk averse; 2. It is a one period world, and investors maximize expected utility of end-of-period wealth; 3. A risk-free security exists and everyone can invest in, or borrow, at this risk-free rate; 4. There are perfect capital markets: no frictions to trading (commissions, taxes, continuous time trading) and costless access to information (everybody will have same information). Asset pricing theory is a framework designed to identify and measure risk as well as assign rewards for risk bearing. This theory helps to explain why the expected return on a short-term government bond is less than the expected return on a stock, why two different stocks have different expected returns, and also why expected returns change through time. The asset pricing framework usually begins with a number of premises such as: investors like higher rather than lower expected returns, investors dislike risk, and investors hold well-diversified portfolios. A major implication of holding a diversified portfolio of securities is that the risk of a single stock can be divided into two components: 1. Unsystematic or diversifiable risk - can be eliminated through portfolio diversification (such risk includes company-specific events such as the discovery of a new product (positive effect) or a labour strike (negative effect)). The fact that a company operates in a competitive sector of the market place does not necessarily mean that it will have a high cost of capital; 2. Systematic or non-diversifiable risk - cannot be eliminated through portfolio diversification. Events that affect the entire economy instead of only one firm, such as changes in the economy‟s growth rate, inflation rate and interest rates Theory predicts that financial markets will not reward unsystematic risk, because it can be eliminated through diversification at practically no cost. Thus, the only risk that matters in determining the required return on a financial asset is the asset‟s systematic risk. In other words, the required rate of return on a financial asset depends only on its systematic risk. The underlying premise of portfolio theory is that, as more assets / securities (with varying levels of risk) are added to a portfolio, the risk of the overall portfolio falls. Theoretically as the number of 10 assets approach infinity, the diversifiable risk tends to zero. However, in practical terms, a portfolio with around 40 different assets sufficiently tends towards zero diversifiable risk (this implies that even a stock exchange with less than 100 companies listed, can be sufficient in achieving a locally diversified portfolio, provided that the industrial composition of the exchange is not biased towards a particular industry). It is the impact that a given asset has upon a portfolio that is of concern, and not the individual risk of that asset. The diagram below illustrates the impact of reduced volatility on a portfolio by the addition of shares / firms in a portfolio: Diversifyable risk Total risk Non diversiftable risk 0 5 10 15 20 25 30 35 40 45 Number of securities (assets) in portfolio CAPM recognises research that suggests that investors require a premium for investing in equities rather than in risk free investments. The premium is commonly known as the market risk premium (MRP) and notionally represents the premium required to compensate for investment in the equity market in general. A firm‟s systematic risk is usually measured relative to the market portfolio (the portfolio that contains sufficient assets such that the diversifiable risk tends towards zero). Systematic risk of a stock is estimated by measuring the sensitivity of its returns to changes in a broad stock market index. This sensitivity is called the stock‟s “beta coefficient” (beta). A firm‟s risk depends on the risk of the cash flows generated by the firm‟s assets (business risk) and the risks associated from the use of debt (financial risk). A firm‟s asset beta and corresponding return captures its business risk, whereas a firm‟s levered equity beta and corresponding return captures both its business risk and finance risk. The formula below shows the relationship between the asset return and the levered equity return against changing levels of debt and is know as Modigliani & Miller proposition II1 with taxes. re  ra  ( ra  rd )(1  tc ) 1 D E F. Modigliani and M. H. Miller, “The Cost of Capital, Corporate Finance and the Theory of Investment,” American Economic Review 48 (June 1958), PP. 261-297. 11 Where: re ra rd E D tc = return on equity = return on assets, equivalent to the unlevered return on equity = return on debt = market value of equity = market value of debt = marginal corporate tax rate Conceptually the first term of the above equation represents the return equity holders demand from holding the asset, while the second term represents the additional return equity holders demand from the increased financial risk of equity due to the inclusion of higher priority debt instruments. The above equation is valid for both returns and betas and in many cases it is the beta which is levered according to the equation above with the simplifying assumption that the debt beta is equal to zero. Since beta measures a security‟s risk relative to the market portfolio, a security‟s risk premium equals the market risk premium × the security‟s beta. The CAPM states that the expected return on any security is the risk-free rate, plus the market risk premium multiplied by the security‟s beta: re  rf   e,l ( rm  rf ) Where: rf βe,l rm = risk free return = equity beta = Market return. [(rm-rf) if often referred to as the market risk premium] In graphical form, the security market line is shown as a linear line with intercept through the risk free rate. The intersection point of the security market line and the point at which beta is “1” represents the market portfolio rate. Changes in the beta of an asset imply a higher or lower return above or below the market rate (i.e. above / below the ERP). Expected Return ne t Li rke Ma ty Market portfolio rate uri Sec Rm Rf Risk free rate e.g. Government bond rate 1 Beta 12 In essence (and in mathematical terms) the beta of an asset measures the covariance of the returns of that asset with respect to the returns from the “market” for a unit change in the market‟s own return: i  Cov( ri , rm ) Var ( rm ) = β coefficient of investment asset i = Covariance of the return on asset i (rm) with the return on the market portfolio = Variance of the return on the market portfolio (rm) Where: βi Cov (ri, rm) (rm) Var (rm) Estimates of beta vary according to the period examined. Betas calculated over a long period can be inaccurate, because the nature of the firm‟s activities may have changed. Similarly, shorter horizons may give too much weight to single events, and can result in large variation in the cost of capital between reviews because of changes in the estimated beta. Typically 60-month data are used to derive betas, although some analysts also use 36-month data. 4.3 Factors of the CAPM equation requiring derivation 1. The return that can be obtained from a „risk free‟ investment; 2. The return that reflects diversified market risk (market risk premium, “MRP”) i.e. the extra return required by an investor to put his/her money into equities rather than risk free investments; 3. The Beta factor for the individual asset under consideration, which determines whether the risk associated with that investment (share) is less or greater than the risk of the market overall. It follows from the above that in determining the cost of equity it is necessary to consider: Each of these elements is discussed below. 4.3.1 Derivation of the Risk Free Rate The risk free factor, as the name suggests is an investment in an asset that provides a return for an investor with no inherent risk. It is a safe investment which has a stated return, set for a given time period, and where the return is to a large extent guaranteed, such that there is not the possibility that the party to the consideration will default on payment. An investment which has the lowest possibility of default is commonly regarded as being the country‟s sovereign bond rate. Portfolio theory states that the true risk free rate is one that has zero correlation with the market portfolio. In practice, however, there are few good proxies for a risk free asset since inflation and other factors have been shown to lead to covariance between bond and stock markets2. Companies invest for the long term and should be permitted to choose the term of financing that they see as most likely to maximise shareholder wealth. Setting the estimate on the basis of a shorter period would seem to be undue interference in the financial management decisions of the companies. 2 Harrington , 1987 13 The relevant question is what benchmark bond rate an investor would look at when investing in long life assets. It is common to match the duration of the cash flows being evaluated with the term of the bond. Duration is the weighted average period of time which will elapse before the cash flows form a bond are received, the weights used being the present values of the individual cash flows expressed as a fraction of the total present value of the bonds. In practice, investors often use the 10-year bond terms as an approximation of the duration of cash flows. Long-term bonds have an advantage as a benchmark because long-bond prices reflect not only today‟s short-term interest rate, but also future expected interest rates. However, as it is the desire to get the best proxy for a „risk-free‟ rate, it is best to consider an income return rather than a total return, such as the redemption yield of a government bond. As stock markets are always looking forward, yields must equate to anticipated levels of market interest rates. Price changes in bonds are due to unanticipated yield changes. There appears to be a general consensus that the rate of return on an index linked bond is the best proxy for the expected risk free rate, primarily because the yield on index linked bonds are immune from the effects of unanticipated inflation. Consequently the 10-year swap interest rate3 should ideally be used as a proxy for the risk-free rate. In emerging markets the risk free rate is generally not as simple to determine as in developed markets. There are three main problems with emerging market derivation. Firstly the sovereign debt is generally not risk-free, secondly it is often difficult to find long–term debt to match the duration of the cash flows being reviewed, and thirdly any longer term debt which does exist is generally denominated in US dollars or a stable European currency. These concepts together with methods to estimate an appropriate risk-free rate are discussed in the context of Macedonia in section 7. 4.3.2 Derivation of the Equity Risk Premium (ERP) The equity risk premium, as has already been discussed, represents the additional return that an investor would require to invest in equities as a general asset category. It is the premium required above the risk free rate that an investor would require to bear the additional risk inherent in equity returns versus returns on a risk free asset. The traditional methodology of estimating such a parameter is through historic interpretation of data on returns from risk free rates and returns from a “market”. It is however, important to recognise that the cost of capital calculation is a forward looking concept and data based on historic interpretation need to be considered in this light. The historic risk premium and the prospective (forward looking) equity risk premium are discussed below. 4.3.2.1 The historic Equity Risk Premium The historic Equity Risk Premium (ERP) can be measured by comparing the return on equities with the return from risk free investments. The return on equity can be decomposed into a capital gain, representing the price appreciation over the holding period and an income return, i.e. the dividend. Practitioners and academics have carried out a large number of investigations into the value of the ERP, using both quantitative techniques and surveys. These studies have produced a range of widely differing estimates, meaning that the AEC needs to choose a value from within the plausible range implied by these studies. 3 An interest rate swap is in essence a series of forward contracts on interest rates. Under the commonest form of interest rate swap, a series of payments calculated by applying a fixed rate of interest to a notional principal amount is exchanged for a stream of payments similarly calculated but using a floating rate of interest. Differences in the credit quality between entities borrowing money motivate the interest rate swap market. 14 The most widely accepted study of market risk premium is by Dimson, Marsh and Staunton (DMS) of the London Business School (LBS). The most recent example of their work on the equity risk premium can be found in their 2006 paper “The World Equity Premium: A Smaller Puzzle”, revised April 2006. This study expands the data source to a world index covering 17 countries over a period of 106-years. The study also discusses prior sources of the Equity Risk Premium such as studies and databases created by Ibbotson Associates, Barclays Capital and CSFB. These databases often contained flaws relating to survivorship bias4, success bias5 and incomplete data which, although frequently used, tended to overestimate the equity premium. The AEC believes the data supplied by DMS to be the most complete and accurate data source on developed market equity risk premium and is minded to use this data in the calculation of the Equity Risk Premium. Historical Equity Premium Relative to Bills Country Geometric Arithmetic Standard Standard % p.a. Mean Mean Error Deviation Australia 7.08 8.49 1.65 17.00 Belgium 2.80 4.99 2.24 23.06 Canada 4.54 5.88 1.62 16.71 Denmark 2.87 4.51 1.93 19.85 France 6.79 9.27 2.35 24.19 Germany* 3.83 9.07 3.28 33.49 Ireland 4.09 5.98 1.97 20.33 Italy 6.55 10.46 3.12 32.09 Japan 6.67 9.84 2.70 27.82 Netherlands 4.55 6.61 2.17 22.36 Norway 3.07 5.70 2.52 25.90 South Africa 6.20 8.25 2.15 22.09 Spain 3.40 5.46 2.08 21.45 Sweden 5.73 7.98 2.15 22.09 Switzerland 3.63 5.29 1.82 18.79 U.K. 4.43 6.14 1.93 19.84 U.S. 5.51 7.41 1.91 19.64 Average 4.81 7.14 2.21 22.75 World-ex U.S. 4.23 5.93 1.88 19.33 World 4.74 6.07 1.62 16.65 * Germany omits 1922–23 Historical Equity Premium Relative to Bonds Geometric Arithmetic Standard Standard Mean Mean Error Deviation 6.22 7.81 1.83 18.80 2.57 4.37 1.95 20.10 4.15 5.67 1.74 17.95 2.07 3.27 1.57 16.18 3.86 6.03 2.16 22.29 5.28 8.35 2.69 27.41 3.62 5.18 1.78 18.37 4.30 7.68 2.89 29.73 5.91 9.98 3.21 33.06 3.86 5.95 2.10 21.63 2.55 5.26 2.66 27.43 5.35 7.03 1.88 19.32 2.32 4.21 1.96 20.20 5.21 7.51 2.17 22.34 1.80 3.28 1.70 17.52 4.06 5.29 1.61 16.60 4.52 6.49 1.96 20.16 3.98 6.08 2.11 21.71 4.10 5.18 1.48 15.19 4.04 5.15 1.45 14.96 Source: Dimson, Marsh and Staunton (2006) The table above indicates the historic equity risk premium for the 17 markets along with the statistics relating to standard deviation and standard error together with a world market summary. The choice between a domestic or world index depends upon the degree to which capital markets are integrated or segmented and how internationally diversified investors are. The extent of such “home bias” will be considered in the context of Macedonia subsequently in section 5 along with recommended adjustments to a mature equity market premium. There is significant danger from using short term time horizons in the estimation of a forward looking expected equity risk premium. Historically there have been relatively long periods where the actual returns experienced have significantly exceeded or underperformed reasonable expected returns. The table below, an extract from DMS 2006, serves to explain the significance of the variation of observed risk premium over different time periods from 1900 to 2002. 4 Studies which review the historical performance of current listed companies, by definition, will not include the returns experienced by companies which have failed and cease to operate. These companies would have formed part of a well diversified portfolio and as such should have been included in the calculation of historical returns. Survivorship bias overestimates historical returns. Where an analysis is performed of only large companies, e.g. those comprising a stock market index such as the FTSE100, there is the likelihood that these companies have been successful relative to others in the market. Success bias overestimates historical returns. 15 5 Period Description U.S. U.K. Rates of return (%) over the period France Germany Japan World World ex-US (21) 107 (47) (47) 670 (37) 326 40 (46) Selected Episodes 1914–18: World War I 1919–28 Post-WWI recovery 1929–31 Wall Street Crash 1939–48 World War II 1949–59 Post-WWII recovery 1973–74 Oil shock/recession 1980–89 Expansionary 80s 1990–99 90s tech boom 2000–02 Internet ‘bust’ (18) 372 (60) 24 426 (52) 184 279 (42) (36) 234 (31) 34 212 (71) 319 188 (40) (50) 171 (44) (41) 269 (35) 318 226 (46) (66) 18 (59) (88) 4,094 (26) 272 157 (57) 66 30 11 (96) 1,565 (49) 431 (42) (49) (20) 209 (54) (13) 517 (47) 255 113 (44) Source: Dimson, Marsh and Staunton (2006) The table above clearly indicates the different historic equity risk premium observed over significant historical periods. It can be see that taking a short time period to estimate the future equity risk premium can be misleading. For example, taking a period which began in 2000 would expose the short term historic mean to the excessive negative performance of the stock market within that period driven by the internet „bust‟ and market readjustment. Similarly, a 10-year period covering either the 1980‟s or 1990‟s would expose the mean estimate to periods of relatively strong growth. It is therefore clear that an appropriately long time series estimate is required to average out the effects of the periods of either superior or poor performance which may be witnessed other what may in some cases be decades. It is also important to note that estimates of the equity risk premium have been reducing over time and there are a number of possible explanations for this. First there has been productivity and efficiency growth within the industrial sector, particularly post 1945 allowing significantly higher returns for a given fixed asset base. Second, there have been significant improvements in management and corporate governance, fuelled by advances in technology both at the senior management level, where control is necessary and at the operational level where efficiency improvements are driven by technological changes. Third, international trade and investment flows have increased and forth, the ending of the cold war may have led to a more stable business environment. However recent world events may have replaced the cold war as a destabilising force in international affairs. More generally it appears that the growth of the stock market and diffusion of such assets from institutions to the general masses has meant that there may have been a fall in the required rate of return due to diminishing investment risk, this is further influenced by improved opportunities for diversification at both domestic and international levels. Such factors have reduced the volatility historically associated with the financial markets. These factors have contributed to, and are now built into higher stock prices, and hence lower equity risk premiums, specifically within well developed markets. It could also be argued that sector specific events may have again changed the forward looking risk premium recently, such as the diminishing appetite for investment within high technology areas; the overcapacity within the market and the substantial gearing for many telecommunications companies which led to financial distress costs. However, the telecommunications sector is not isolated in this respect, with many other sectors also suffering from similar effects - the equity risk premium is considered for an investor with a well-diversified portfolio and thus the impact of the telecommunications sector should not be dramatic. The AEC considers that the use of short time series is not acceptable in the calculation of the future mean equity risk premium from the observed historical mean. The AEC therefore considers that the dissect observation of the equity risk premium from the Macedonian stock market does not have sufficient history to allow its use in the determination of the appropriate future equity risk premium and that reference should be made to established analysis of developed markets. 16 The most accurate and long dated data tends to be available for developed countries (as highlighted by the Dimson, Marsh and Staunton study) and its applicability to emerging countries needs to be assessed. Issues of possible adjustments to the ERP for an emerging market will be considered in section 5. 4.3.2.2 The prospective Equity Risk Premium As was stated earlier, cost of capital is a forward looking exercise. Looking forward, it is the “prospective” risk premium that must be considered, i.e. the reward investors require now, and in the near future for taking on risk, not the historic equity risk premium. Where the geometric mean is appropriate for historic results, the arithmetic mean is an appropriate measure for forward looking results, since it presents the mean of all the returns that may possibly occur over the investment holding period6. The historic arithmetic means are influenced by the periods of extreme volatility during the 20th century; periods of extreme hyperinflation etc may not occur in the future. Historic data therefore should be taken with care when trying to determine future risk premia. In their most resent paper, Dimson, Marsh and Staunton decompose the historic geometric equity premium into its constituent parts. The table below shows a summary of this decomposition. % p.a. Country Australia Belgium Canada Denmark France Germany Ireland Italy Japan Netherlands Norway South Africa Spain Sweden Switzerland U.K. U.S. Average World (USD) Real dividend growth rate 1.30 (1.57) 0.72 (0.87) (0.74) (1.54) (0.25) (1.46) (2.39) (0.16) (0.25) 0.91 (0.62) 2.88 0.32 0.61 1.32 (0.10) 0.77 plus* plus plus Expansion in Geometric mean Change in real the P/D ratio dividend yield exchange rate 0.46 5.83 (0.24) 0.08 3.95 0.62 0.98 4.46 (0.04) 1.43 4.68 0.47 0.42 3.93 (0.14) 0.97 3.69 0.23 0.38 4.66 0.25 (0.08) 4.05 0.10 1.59 5.39 0.32 0.41 5.00 0.27 0.50 4.02 0.25 0.31 5.95 (0.80) 0.24 4.13 0.00 0.67 4.09 (0.05) 0.60 3.52 0.72 0.18 4.68 (0.03) 0.75 4.36 0.00 0.58 4.49 0.11 0.68 4.23 0.00 minus equals U.S. real Equity premium interest rate for U.S. investors 0.96 6.42 0.96 2.05 0.96 5.18 0.96 4.74 0.96 2.47 0.96 2.35 0.96 4.05 0.96 1.58 0.96 3.85 0.96 4.54 0.96 3.54 0.96 5.38 0.96 2.75 0.96 6.72 0.96 4.22 0.96 4.46 0.96 5.51 0.96 4.11 0.96 4.74 Source: Dimson, Marsh and Staunton (2006) Dimson, Marsh and Staunton have argued that it is not appropriate to simply use the unadjusted historic risk premium as an estimate of the future expected risk premium since certain factors which caused higher than anticipated equity market returns will not be repeated. They argue, for example, that the expansion in the price/dividend (P/D) ratio occurred, to a certain extent, due to the reduction of risk associated with integrating markets which resulted in an increase in the value of stocks and hence returns. By definition, this return based on improved market integration cannot be repeated once markets have become fully integrated. Removing this non-repeatable occurrence would result in a reduction in risk premium for the world of 0.68%. Further, the real exchange rate movements within integrated markets could reasonably be assumed to be zero given integrated 6 The more volatile or risky the sequence of returns, the greater will be the difference between the arithmetic and geometric means. The key use of equity risk premium is to determine investors‟ required returns, and hence the cost of capital to use as the discount rate in valuing assets. For discounting uncertain future cash flows it is necessary to use the expected risk premium. The expected risk premium is the arithmetic mean of the one-year premia. 17 market assumptions of parity. The world average change in real exchange rate average is already 0.0% and so no further adjustment is required. These two non-repeatable impacts would result in an equity risk premium reduction for the world from 4.7% to 4.0%. Furthermore, DMS also note that current geometric mean dividend yield is less than the historical average and the adjustment for current expectations suggest a further reduction factor of between 0.5% and 1.0%. The average world equity premium is therefore within the range of 3.0% - 3.5% on a geometric basis and between 4.5% - 5.0% on an arithmetic mean basis. The results of previous studies from Dimson, Marsh and Staunton have been used by the Competition Commission in the UK, for their investigation of mobile termination charges in the UK, and more recently have been considered as a basis for OFCOM‟s consultation on cost of capital. We therefore propose to use the Dimson, Marsh and Staunton study with relation to the world index in determining a mature market future equity risk premium used in the determination of an appropriate cost of capital for Macedonian. The AEC will use the future estimate of the world arithmetic mean equity risk premium range of 4.5% to 5%. 4.3.3 Beta The Beta coefficient of an asset represents the sensitivity of the asset‟s return to changes in the market index - it is the firm‟s systematic risk, measured relative to the market portfolio. In essence (and in mathematical terms) the beta of an asset measures the covariance of the returns of that asset with respect to the returns from the “market” for a unit change in the market‟s own return. The beta should be calculated using rolling daily/monthly data measured over a period of say 5 years. However, historic beta may not give an accurate measure of the current level of risk in the telecoms sector. A point of particular concern when measuring betas is the influence of the “technology bubble”. The non-diversifiable risk (beta) of a company is measured by calculating the way its stock price moves, both in speed and volatility, in relation to market indexes (usually the national financial stock market, such as the FTSE, DAX etc). During the 1999 to 2001 bubble, the influence of telecom, media, and technology (TMT) share prices on the indexes had a distorting effect on the calculated betas. From December 1997 to August 2000, the three sectors together accounted for about 68% of the change in the total S&P 500 index, for example. By 2000, the stocks of such companies represented 45% of its value and were responsible for as many as 21% points of its 26% annual growth. In such an environment, the effective beta of individual companies within the TMT sector results in a distorted value relative to normal conditions (i.e. the covariance between a TMT beta relative to a heavily biased TMT market index would be higher, than a normal weighted market index, thus resulting in a artificially higher beta). However, this is only relevant to a greater extent within those countries that experienced the bubble, and where the market index consisted of a large number of TMT stocks. A directly observable beta against an appropriate market is the most preferable way to calculate the beta for a specific company. The AEC notes that none of the AD Makedonski Telekomunikacii (MT) share capital is traded on an exchange and therefore no directly observable beta can be obtained. International benchmarks must therefore be used to arrive at a proxy and care therefore needs to be taken in evaluating the benchmark companies to ensure that they reflect an appropriate level of risk. Particular areas of concern may be the benchmark companies susceptibility to the „internet bubble‟ and the effects of deregulation and increased competition. Increased competition would be expected to increase the beta as the increased risk of the companies future cash flows increases the volatility of the companies returns in comparison to the general market. Furthermore, 18 Betas and Gearing care needs to be taken to ensure that the chosen beta reflects accurately the business lines being evaluated. The AEC believes, that by careful selection of the benchmark companies, appropriate levels of risk can be accounted for, and believes it has done this as discussed subsequently.. Gearing As has been discussed previously, the AEC needs to consider the optimal gearing level for a company (Debt / Debt + Equity), in recognition that debt benefits from certain tax advantages). This is only true up to an optimum level, beyond, which the higher levels of debt adversely effect the financial stability of the company. As the debt passes an optimum level, the risk on debt repayment increases which results in an increase in the return required by the debt holders - WACC trends upwards after this point. The following diagram illustrates the effects of increased leverage on returns and WACC. 30.0% 25.0% return on equity 20.0% 15.0% return on assets WACC 10.0% return on debt 5.0% 0.0% Debt/Equity In assessing the prudent level of debt, it is also the firm‟s interest cover ratio which must be considered. As the level of debt is increased, the interest cover ratio falls, and it is likely that the cost of debt will rise. Where a firm is highly profitable, its interest cover ratio may be high, even with a large amount of debt, whereas with new entrant, for the same proportion of debt, its interest cover ratio will be much lower. It could therefore be argued that an optimal gearing level will be relatively higher for a larger more profitable firm than a new entrant. The typical gearing for established telecommunications operators within a competitive market is approximately 28% over a period of time (see table below): D/V Telecom Electric utility Water utility 28% 32% 50% Levered Beta 1.4 1.1 0.6 Unlevered Beta 1.0 0.8 0.4 Source: Damodaran It is the market values of the debt that is important not the book value. In many cases, it may be the case that telecom operators may have significant debt on their books, but the actual market value of such debt is much lower. The above table is a simple average over time. In any one year it changes, and some years it may be lower or higher than those stated. 19 It appears that for those sectors which are more established and are considered to be able to sustain a higher debt load, the gearing is higher, approaching approximately 32-50% (i.e. utility sectors). In assessing the network of AD Makedonski Telekomunikacii (MT), it could be argued that the business unit relating to interconnect is more akin to a utility business and, as highlighted above, could sustain a higher level of gearing than the rest of the network. The ability of a company to support debt is determined by the stability of the projected cash flows of the business and the availability and rate of financing and it could be argued that a higher gearing of up to 35% could be assumed for AD Makedonski Telekomunikacii (MT. The AEC, however is minded to use the average gearing level of 28% indicated above for the telecommunications sector as a whole. 4.3.4 Weighted Average Cost of Capital and preferential rates The Weighted Average Cost of Capital (WACC) of a company may be influenced by the leverage of its owners or parents. The analysis by J Gregory Sidak, in January 2001 in the publication: “Capital subsidies, profit maximisation, and acquisitions by partially privatised telecommunications carriers”, suggests the effects of parentage have little effect upon the WACC. This may be primarily due to the relative size of incumbent operators; whether they have parentage or not, their size may imply a lower default probability and thus access to the debt market may be on par whether parental advantage is present or not. In calculating the cost of capital of an operator, it is the forward looking scenario that is of primary interest and assessment must be made of whether the company would continue to benefit from such government financing in the future and the sustainability of such schemes, where it is evident that such preferential rates have anticompetitive effects upon the market. The AEC does not believe operators are gaining particular advantages from preferential rates and proposes not to consider such issues here and will instead rely upon forward looking market rates. 20 5 Possible adjustments to CAPM for emerging markets This section discusses the CAPM model in the context of Macedonia as an emerging market. It focuses on how international data can be adjusted to reflect the reality of the state of the market within Macedonia. The conclusions from this analysis will be used to derive an appropriate ERP for Macedonia, to be used in section 7. 5.1 Testing the assumptions of the traditional CAPM in an emerging market The application of CAPM in emerging markets has drawbacks specifically with the assumption that investors can readily diversify their portfolio. Within the traditional CAPM, there is an assumption of normally distributed asset returns, which may not be appropriate in an emerging context. The model also assumes perfect market integration - markets that are "emerging" would seem, by definition, to be imperfect in their integration:  Financial markets are integrated if financial assets with the same risk characteristics have identical expected returns, irrespective of the market to which they belong. The widespread use of domestic market indices as a proxy for the true market portfolio assumes that markets are internationally segmented. The cost of equity should fall when markets become integrated; Investors can transfer some part of country-specific risks to foreign investors, while taking some foreign risk in exchange; Each country‟s cost of capital is reduced by making risk diversifiable that would not be otherwise diversifiable.   A sufficient strong condition for this to work is that there are no effective barriers to portfolio investment across borders. That is, local investors are free to add any stock in the world to their portfolio and international investors are free to choose any stock within a particular country. With capital market integration, a world version of the Capital Asset Pricing Model is achieved. In the integrated world, a country portfolio's risk is its covariance with world returns. This covariance is rewarded with a common world price, which is linked to weighted average risk aversion in the world. In the segmented world, a country portfolio's risk is it variance. The variance is rewarded with a country specific price, which is linked to a weighted average risk aversion within the particular country. These scenarios: integrated or segmented are polar extremes. Roughly speaking, the expected returns in a partially segmented emerging market reflects some reward for the covariance with world returns as well as some reward for the market's own variance, i.e. a hybrid CAPM that includes both variance as well as covariance with the world. The integration process is gradual, as regulations are changed such that local investors can purchase stocks outside their country and foreigners are allowed into the local market. The most important impact will come from foreign investors. When the foreign investors invest into the market to take advantage of the low correlation - high expected return opportunities, prices rise and expected returns decrease. So, one immediate implication of capital market liberalization is that expected returns should decrease. There are powerful implications to this decrease in expected returns. An immediate implication is that the cost of equity capital decreases. As financial markets are liberalised globally and information technology provides access to information and transactions services more easily and cost effectively, it would seem, it is more likely that some version of the world capital asset pricing model holds more today than in the past for emerging markets. Nevertheless, as a result of segmentation, a situation may arise where the traditional CAPM underestimates the cost of capital for emerging markets. To incorporate all additional factors 21 affecting the return in emerging markets, ideally the use of multi-factor models should be made, but the information required for such a task is significant, and without compete and accurate data, it seems there is little advantage of using a multi-factor model. 5.2 Models to adjust the traditional CAPM to make it relevant in an emerging market There are various models that attempt to incorporate elements of increased risk premium in emerging markets as detailed in Annex 1. As it can be seen there are numerous attempts to incorporate some element of additional risk, however, no consensus seems to exist. In many of the models proposed, subjective adjustments seem to be applied. The AEC considers that it is appropriate to base emerging market adjustments on common regulatory practice. The AEC believes that any additional risk allowance must fundamentally be aligned with potential changes to the fundamental assumptions used within the CAPM. The AEC believes the most significant reason for additional emerging market risk is correlated to a large extent from the nature of the local market, and how segmented it is. The AEC furthermore believes that the risk associated with investments in emerging markets may be appropriately adjusted from an investigation into the sovereign default spread and is minded therefore to use the following model for the determination of return on equity for the Macedonian market: re  rf ,US   e ,US * ( MRPUS )  CRPcountry Where, country risk premium is calculated as the default spread between a stable economy and an emerging market, both denominated in the same stable currency. 22 6 PART 2 of WACC equation - Pricing debt This section discusses the second part of the WACC calculation; the cost of debt. This section briefly highlights the factors which contribute to the different rates of debt observed in the market. 6.1 Factors affecting the cost of debt If a firm takes out a loan, the firm‟s cost of debt is the rate charged by the bank. If we know a sufficient amount of information, the valuation formula can be solved for the investors‟ required rate of return. If the firm has no bonds outstanding, its cost of debt can be estimated by adding a credit risk spread to the yield on government securities of the same maturity. Since interest expenses are tax deductible, the after tax cost of debt is the relevant cost of debt: After-tax cost of debt = Pre-tax cost of debt × (1 – marginal corporate tax rate) However, the after tax cost of debt is a valid estimator only if the firm is profitable, or carry back or carry forward tax rules apply to interest expenses. The value of a particular issue of corporate debt depends essentially on three items: 1. the required rate of return on riskless (in terms of default) debt (e.g., government bonds or the very high grade corporate bonds); 2. the various provisions and restrictions contained in the debt covenants (e.g., maturity date, coupon rate, call terms, seniority in the event of default, etc.); 3. the probability that the firm will be unable to satisfy some or all of the indenture requirements (i.e., the probability of default) - particular regard is paid to a firms interest coverage ratio. The debt rate will thus correspondingly vary for new entrants and established operators. New entrants will not be able to tap into international capital markets for finance in early network operation phase, until rating agencies have established the operator‟s credentials. Gearing therefore, as has already been discussed previously, could be lower in the earlier years for new operators. 23 7 Factors affecting calculation of WACC in Macedonia This section applies the theory that has been explored in the previous sections in the context of Macedonia. This section will determine the appropriate parameter values of the WACC formulae for application in Macedonia. These parameters include: 1. Risk Free Rate in Macedonia - which varies depending on economic activity and government monetary policies 2. Equity Risk Premium in Macedonia - which varies depending upon the size and liquidity of the stock market and the type of investors in Macedonia and their location 3. Gearing levels for Macedonian telecom companies – based on international practice 4. Beta for AD Makedonski Telekomunikacii (MT) – based on international practice 5. Debt rate for AD Makedonski Telekomunikacii (MT) – based on the estimated sovereign rates and appropriate estimates of default spread. 6. Corporate marginal tax rate – based on current legislation 7.1 Risk Free Rate and Country Risk Premium The risk free rate in emerging markets is generally not as simple to determine as it is in developed markets. There are three main problems in determining emerging market risk free rate. Firstly the sovereign debt is generally not risk-free, secondly it is often difficult to find long–term debt to match the duration of the cash flows being reviewed, and thirdly any longer term debt which does exist is generally denominated in US dollars or a stable European currency. There are three potential models that have been proposed for estimating the risk free rate in emerging markets:  Method 1 – Local Bond Yield. This method starts with the yield on a local currency denominated bond and adjusts for sovereign risk and possibly extension of duration if appropriate; Method 2 – International currency denominated local bond yield. This method uses the yield of a local currency denominated bond and adjusts for sovereign risk, duration (if appropriate) and currency effects; Method 3 – Mature market bond yield. This method starts with the yield on a mature stable market bond and adjusts for inflation differential.   The method that is ultimately chosen is generally determined by the information available. In Macedonia, only foreign currency debt has been issued in sufficient size and is actively traded. Local currency denominated debt is very short term in duration. The table below shows lists the issues of sovereign debt by Macedonia which are traded on the Macedonian Stock Exchange and for which there are readily accessible prices. Name RM01 RMDEN01 RMDEN02 RMDEN03 RMDEN04 Issue Date 19.09.2000 17 Jun 2002 20 Mar 2003 25 Feb 2004 2 Mar 2005 Amount €550ml €2.5ml €38.1ml €47ml €58ml Duration 1.9 2.1 2.3 2.5 2.7 Yield 7.8% 7.8% 7.9% 7.7% 7.6% 24 The debt is dominated by foreign currency issues in Euros and the term of repayments includes periodic principal repayments, which has the effect of reducing the bond duration. As was discussed earlier, ideally it would be appropriate to use 10-year bonds, as these typically match the generally accepted duration of the company cash flows, and are considered appropriate by other Regulatory Authorities. In the case of Macedonia such bonds are not available and the traded bonds indicated above have durations which range from 1.9 to 2.7 years as at end October 2006. Given the information available, the AEC proposes to use Method 2 above, where the foreign currency denominated government debt is adjusted for sovereign risk, currency and duration effects. It is thus necessary to determine suitable adjustments to enable the AEC to estimate a 10-year local currency risk-free bond. Such adjustments require the assessment of the following parameters: 1. Euro area sovereign debt and term structure. 2. Sovereign risk premium. 3. Duration and inflation adjustments. It is relatively easy to ascertain Euro area sovereign debt and term structure from published financial reports. The table below lists bonds issued by various countries of differing terms. Country Australia Britain Canada Denmark Japan Sweden Switzerland US Euro area 3-month 6.39% 5.19% 4.18% 3.62% 0.35% 2.77% 1.87% 5.33% 3.58% Source: Economist 2-year 6.08% 5.00% 3.97% 3.78% 0.78% 3.48% 2.19% 4.76% 3.75% 10-year 5.64% 4.56% 4.03% 3.73% 1.72% 3.64% 2.34% 4.63% 3.72% As can be seen, Euro area debt is currently traded at 3.75% YTM for 2-year bonds and 3.72% YTM for 10-year bonds. The first step is to remove the sovereign risk embedded in the Macedonian bond yield. This is achieved by comparing a Euro denominated Macedonian bond yield with a Euro denominated Euro area bond yield of the same duration. In using similar yield and issues in the same currency, the there can be no differences attributable to either a duration or currency effect. The difference must therefore substantially relate to the difference on sovereign risk between the Euro area and Macedonia. The difference between the 2-year Euro Area bond yield (3.7%) and the average of the 1.9 and 2.1 year duration Macedonian bond (7.8%) gives us the Macedonian Country Risk premium of approximately 4.1% and a Euro denominated 2-year risk-free rate of 3.7%. The second step involves converting the Euro denominated 2-year risk-free rate to a 10-year MKD denominated risk-free rate. Where real rates of return are comparable between two countries, forward looking inflation predictions can provide a method of estimating the movement in currency and the forecasts of individual inflation within a country can provide an indication of the term structure of interest rates. To use such a methodology, an assessment of Macedonian inflation rates over the 10-year period is required. The appropriate currency and duration adjustments which enable an estimate of a longer term duration to be estimated, (i.e. to derive a ten year MKD denominated risk free rate from the 2-year 25 Euro denominated risk free rate) involves estimation of real rates of return and inflation rates for both the Euro area and for Macedonia. The Euro area inflation forecasts indicated by CPI are approximately 2.2% and 2.1%7 and the real return p.a. over the 2-year period is therefore inferred from the bond yield and inflation estimates to be 1.3%. The Euro inflation estimate for the period from 2-years to 10-years is calculated from the 10-year bond yield and by assuming the real return remains constant over the period. The inflation estimates for Macedonia were estimated to be 5% p.a. This gives a currency and duration adjustment of approximately 2.9%. The table below summarises these adjustments and calculates the Macedonian 10-year risk free rate in MKD at approximately 6.6%. 2-year sovereign bond € 7.8% Less country risk 4.1% Currency and duraton adjustment 2.9% Risk free rate - MKD 6.6% 0.0% 1.0% 2.0% 3.0% 4.0% 5.0% 6.0% 7.0% 8.0% 9.0% Source: AEC estimates 7.2 Market Risk (Equity Risk Premium) The AEC needs to choose an appropriate Equity Risk Premium (ERP) within the plausible range of estimates that balance the relative risks of setting a cost of capital that is too high with one that is too low. A rate that is too low, may deter further investment by the operator and could potentially affect infrastructure competition, however, it has immediate benefits for consumers, through lower prices. A rate that is too high, may lead to consumers paying prices that are above the competitive level, leading to an overall welfare loss, and possible over investment by operators into unworthy investments. The AEC has concluded that the Macedonian equity market is not sufficiently large or liquid and does not have a long enough history to allow meaningful estimation of the market risk premium through direct observation to be made. The AEC is therefore minded to use international 7 Source: Economist 26 benchmarks obtained from recognised international financial reports such as the Dimson, Marsh and Staunton study8. SUMMARY (ERP) Appropriate Equity Risk Premium 4.5 – 5.0% AD Makedonski Telekomunikacii (MT) 7.3 Gearing for AD Makedonski Telekomunikacii (MT) As has been discussed earlier, the AEC is not proposing to treat the different businesses units as different financial entities, each with its own debt rate and gearing ratios, as this will ultimately introduce additional assumptions that may deter from an accurate assessment. The AEC proposes to use the average of 28% gearing level as representing an optimum level. SUMMARY GEARING [D / (D+E)] Appropriate Gearing AD Makedonski Telekomunikacii (MT) 28% 7.4 Beta for AD Makedonski Telekomunikacii (MT) Asset betas reflect the contribution to portfolio risk of the investment, independent of financing, i.e. the riskiness of the business itself. The AEC will use the comparable asset betas to determine an appropriate return on assets. The return on assets will then be levered according to the formula below to determine the appropriate costs of equity: re  ra  ( ra  rd )(1  tc ) Where: re ra rd E D tc = return on equity D E = return on assets, equivalent to the unlevered return on equity = return on debt = market value of equity = market value of debt = marginal corporate tax rate 8 Dimson, Marsh and Staunton (DMS) of the London Business School (LBS). The most recent example of their work on the equity risk premium can be found in their 2006 paper “The World Equity Premium: A Smaller Puzzle”, revised April 2006. 27 The effect of liberalisation of the telecom sector will impact the volatility of the incumbent operator with respect to the market - this is the effect of the increased regulatory risk. The effects of the onset of further competition and possible undue extra regulatory pressure, many have argued should be factored into measured betas. By using observed reference betas we can reasonably assume that, since the comparable companies operate within a more competitive market, they currently reflect the effects of increased competition. Furthermore, the AEC is of the opinion that due to the commitments relating to EU accession and the commitments given to the WTO, such regulatory risk is small and therefore no adjustment is required. SUMMARY BETA AD Makedonski Telekomunikacii (MT) Appropriate asset betas (Basset) 1.0 – 1.1 7.5 Debt price for AD Makedonski Telekomunikacii (MT) The cost of debt is the interest that needs to be offered to raise capital in the form of debt. It should reflect the risk-free interest rate that investors would require for lending their money, adjusted to reward investors for the risk that the borrower will default. The cost of debt is the market interest the firm has to pay on its borrowing and depends upon the general interest rates, the firms default premium and possible tax advantages. AD Makedonski Telekomunikacii currently has no outstanding external debt and therefore an appropriate interest rate to use must be calculated independently. The AEC has estimated an appropriate return on debt using the following methodology. Macedonian risk free rate of 6.6% is used as a base rate. This rate must then be adjusted to account for the company default risk. AD Makedonski Telekomunikacii financial statements indicate an ability to sustain a relatively high credit rating indicated by the assumed levels of gearing and interest coverage ratios. It is unlikely, however, that this credit rating will be higher than the sovereign ceiling and hence propose to base the additional default premium on the sovereign premium. Country risk, including sovereign default risk, was previously estimated at 4.1%. The AEC believes that it is appropriate to calculate an indicative level of debt based on the risk free rate and the country risk premium. This would result in a return on debt of 10.7%. SUMMARY DEBT RATE (rd) Appropriate debt rate AD Makedonski Telekomunikacii (MT) 10.7% 7.6 Corporate Tax rate The current marginal tax rate for both operators9 is stated to be 15% of profits and is the value AEC will use in the calculation of the cost of capital for AD Makedonski Telekomunikacii (MT). In the future, the corporate tax rate will be decreased to 12% in 2007 and 10% in 2008. 9 Law on corporate tax (Official Gazette 80/93) 28 8 Aggregate entity level cost of capital In this section, the parameters as determined in the previous sections will be applied to derive a cost of capital for AD Makedonski Telekomunikacii (MT) using the WACC approach. 8.1 Calculating the WACC of AD Makedonski Telekomunikacii (MT) E D  rd (1  t c ) V V WACC  re Return on assets (ra) ra  rf   a * ( MRP )  CRPcountry , where rf = 6.6% β = 1.1 to 1.1 MRP = 4.5 to 5.0 CRP = 4.1% ra = 15.2% to 16.2% Levered return on equity (re) re  ra  ( ra  rd )(1  tc ) D/V = 28% rd = 10.7% D , where E ra = 15.2% to 16.2% re = 16.7% to 18.0% WACC WACC  re E D  rd (1  t c ) , where V V D/V = 28% rd = 10.7% re = 16.7% to 18.0% tc = 15% WACC = 14.6% to 15.5% Summary Aggregate WACC AD Makedonski Telekomunikacii (MT) Appropriate WACC 14.6 – 15.5% 29 The AEC considers that setting a rate which is to the upper end of the range indicated above is appropriate. The AEC therefore proposes to set a WACC of 15.5% for AD Makedonski Telekomunikacii (MT) 30 Annex 1 – Cost of capital models in an emerging country Many different models have been proposed by academics and practitioners to determine an appropriate cost of capital for an emerging market. There does not appear however, to be a consensus of what the real implications of an emerging country are upon the required equity risk premium. Listed below are some of the models that have been proposed with a summary of the advantages and disadvantages of such models. The models generally range from an integrated world model to a segmented local model. The choice of whether to use the integrated or segmented model depends primarily on the level of integration of the particular country to the rest of the world and the perspective of the marginal investor in the company: 1. Direct observation from the local market The direct observation of the local market involves an analysis of the local trading conditions in the country of observation. The formula below is used to determine the observed historical market return. rm  P1  d 1 1 P0 = return on market = price of asset in period 0 = price of asset in period 1 = dividend payment in period 1 Where: rm P0 P1 d1 The equity risk premium is calculated by subtracting the risk free rate from the market return calculated above. The key advantage of this method is that it allows us to directly observe the return characteristics for the specific market. The major disadvantages are the requirement for large amounts of historical data required to determine the long run market risk premiums. This is unfortunately often not available in emerging markets. Specifically for Macedonia, the stock exchange was established in and maintains only stocks. The graph below shows the number of firms listed in the stock exchange during the relevant period together with the value and number of transactions and the market capitalisation vs. GDP. These indicate a rather illiquid market, insufficient historic data10 and a market which is unrepresentative of a well diversified portfolio. In addition there is a requirement to determine an appropriate risk free asset for each period which requires a liquid and deep market for risk free securities. Government bonds are often used as the risk free asset, however in emerging markets these often contain elements of default risk and are therefore, by definition, not risk free. AEC has therefore concluded that it is not appropriate to use a directly observed market risk premium for Macedonia. 10 Long time series of data are generally required. By way of example, recent studies of the US market calculated a standard deviation (σ ) of the historical mean risk premium of approximately 20%. Given that we can calculate the standard error of this mean by using the formula σ/√T we can see that the standard error of a 10-year time series is approximately 6%. This error is significant in terms of the observed mean and illustrates the requirement for very long time series data. Over 100-years of data the standard error falls to 2%. 31 2. The International Capital Asset Pricing Model (ICAPM) The ICAPM is based on the traditional CAPM but assumes that world markets are perfectly integrated. If investors can invest in any markets certain risk associated with particular countries and regions can be diversified away and the only risk remaining is the unsystematic risk on the world portfolio. The ICAPM tends to give results in returns which are lower than a single country model. This is a reasonable conclusion given the assumption that integrated world markets allow the investor to diversify country or region specific risks as part of a world portfolio. The β and the ERP are both calculated with reference to the world portfolio. re  rf   w * ERPw The model appears to work for integrated markets, but where markets are segmented, as is the case in emerging markets, the model fails to adequately price risk and underestimates the required return. The AEC therefore believe that this model is not appropriate for Macedonia. 3. Globally nested capital asset pricing model This model is again based upon the CAPM. Further to the ICAPM this model recognises that in addition to the world risk as priced in the ICAPM, the model incorporates both regional and country specific risk. This reflects the degrees of integration of the region and country to the world market and when the country and region become fully integrated, the model collapses to the ICAPM. Indications of segmentation of the region and the country with the world are estimated by regressing the excess returns of the region against the excess returns of the world, and by regressing the excess returns of the country over the regional risk premium against the world excess returns. The model has the advantage of recognising and pricing regional and country specific risk, but requires regional and country market data to allow the coefficients to be calculated from regression. The AEC believes that insignificant data exists to support this model. 4. Country-spread models Country spread models are particularly well knows and are used extensively in practice. These models follow the following formula. re  rfUS   * ERPUS  Country Risk Premium The model adds a country risk premium to the cost of capital determined from a mature market, normally the US (as is the case of the above formula). The model uses a mature market risk free rate and ERP, while the β is determined with reference to the local market portfolio. A significant and subjective variable in the above equation is the calculation of the country risk premium and there are many models which have been proposed to allow the user to calculate this. A common method of approximating the country risk premium is to use the default spread as a proxy and this is generally measured as the difference between a stable-currency denominated bond issued by the country and the mature market treasury bond rate. The spread represents an ex-ante assessment of a country risk premium which reflects the credit worthiness of the government. The default spread requires the existence of a stable-currency denominated bond issued by the emerging market. Another method is to estimate the country risk with reference to the countries credit rating. These ratings measure default risk rather than equity risk, but both are generally affected by the same factors, namely country currency stability, budget and trade balances and political stability. The rating can be used to estimate default spread over a mature market treasury bond. Using ratings has the advantage of being relatively easily accessible and simple to use. The ratings do, however, tend to lag behind the market expectations (especially when ratings reviews are due). 32 Another alternative is to estimate the level of country risk by looking at fundamental metrics for the country and comparing these to integrated and mature markets. In understanding the degree of integration it is important to understand the following drivers:    Market capitalisation estimator - Market capitalisation to Gross Domestic Product (GDP) ratio Foreign investment estimator – Foreign Direct Investment (FDI) to Gross National Investment (GNI) ratio Stock market liquidity estimator - Stock turnover ratio The AEC believes that there is significant subjectivity to the weightings applied to this method and therefore does not propose to use it. Damodaran also suggests a further adjustment to the basic country-spread model. As the model stands, the country premium is assumed to impact all companies equally. Damodaran suggests that this is often not the case and proposes the inclusion of a separate weighting factor to allow for different risk impacts on different businesses. A summary of the basic models suggested by Damodaran are as follows The following formula assumes that the company‟s exposure to country risk is proportional to its exposure to all other risk re  rf ,US   e,US * ( MRPUS  CRPcountry ) (1) A model which assumes that all company‟s in a country are equally exposed to country risk is as follows re  rf ,US   e ,US * ( MRPUS )  CRPcountry (2) A model which assumes that exposure to country risk is different for each company and that it is not in proportion to all other risks. re  rf ,US   e,US * ( MRPUS )   * CRPcountry (3) The essential difference between these three models is the assumption relating to the effect of the country risk premium (referred to as λ in the above equation) on the individual company. Equation (1) assumes that it is weighted according to the company β. Equation (2) assumes that it is equal to 1, while equation (3) assumes that it is related to a separate and identifiable factor which is specific to the company concerned. Indications of a low λ would be given by issues such as a high proportion of revenues or business operations were not related to the country specific risks. This may be the case in, for example, commodity companies where revenues are largely determined by global markets. It the specific case of Maktel, we believe that the majority of Maketel business is dependent on the local market and discount equation (3) since it requires the use of further subjective estimates. We further assume that Maktel is exposed to country risk in the same way as all other companies in Macedonia and therefore favour the use of equation (2). This method has the benefit that the risk free rate and the ERP are both from a developed market with significant data series history and academic studies from which to source data. Where sovereign yield spreads are available, these can be used to estimate the country risk premium in raw form, or adjusted as indicated above. The beta, calculated with reference to the local market, is generally difficult to obtain for emerging markets due to the limited availability of data, however, the beta may be estimated with reference to comparable companies. 33 5. Relative standard deviation model The relative standard deviation model estimates the ERP of a country by multiplying the ERP of a mature market by the relative standard deviation of that market vs. the mature market using the following formulas ERPcountry = ERPUS * Relative SD of Country Relative SD of Country  SDCcountry SDUS This approach requires the calculation of standard deviations in both equity markets. While the standard deviation is generally available for a mature market, it may not be so readily available for an emerging market. The model has the further disadvantage in that if a market is illiquid, the standard deviation may reduce the volatility and hence underestimate the country specific ERP. The AEC does not favour the use of this model. 34