APPENDIX Cost of capital for UK home credit lenders Introduction

APPENDIX 3.5 Cost of capital for UK home credit lenders Introduction 1. This appendix sets out our approach to estimating a WACC for a large UK home credit business. It is based on information on the five largest home credit lenders: Provident, Cattles, LSB, S&U and Park. The main purpose of the appendix is to provide a figure for the cost of capital which can be compared with the returns of the large lenders. We also consider the applicability of our estimated cost of capital to smaller lenders in Annex B. Approach to measuring cost of capital of larger home credit lenders 2. We considered the appropriate costs of equity and debt for the large home credit companies. None of the five companies had the supply of home credit in the UK as its only activity. Three (Provident, Cattles and S&U) had car finance businesses, LSB provided other credit services and debt collection, and Park had an established savings-based hamper supply business. Provident also had small but growing home credit operations in the Republic of Ireland, Eastern Europe and Mexico, and had been piloting a sub-prime credit card through its Vanquis Bank subsidiary. The existence of these and other activities meant that the corporate entity’s cost of capital was likely to be different from those of its home credit operations. We discuss the effects of these other activities in paragraphs 16, 22 and 37. In calculating an estimated cost of capital for a large UK home credit business we used data such as betas, capital structure and debt premiums that relate to the five largest lenders. As they were the only UK listed companies in the industry, they were the only companies for which data such as betas was readily available. We also considered the appropriate capital structure and determined an appropriate gearing which provided the weightings for the costs of equity and debt in the WACC. The WACC calculation relies on several variables that cannot be calculated with complete accuracy. Therefore we did not express the cost of capital as a single figure, but rather as a range to compare with company returns. Provident submitted an analysis comparing our calculated WACC range with the WACC figures or ranges from other recent inquiries or price determinations made by the CC or other regulators. The components of WACC (particularly the risk-free rate and beta) change over time and WACC will vary between industries. Furthermore, WACCs used in price determinations are forward looking and use estimates for the expected future values of the components, whereas we are able to use observed historic data. For these reasons we did not consider WACC ranges from other inquiries or price determinations to provide useful evidence to support the determination of a home credit WACC. 3. 4. 5. A3(5)-1 Summary 6. We calculated the average pre-tax nominal WACC for a large home credit lender between 1999 and 2004 to be between 8.4 and 11.2 per cent. We calculated the average pre-tax real WACC for the period to be between 6.0 and 8.7 per cent. 1 We also calculated the real pre-tax cost of capital for 2004 to be between 5.4 and 8.1 per cent. This was derived by deflating the 2004 nominal pre-tax cost of capital range of between 8.0 and 10.8 per cent by the RPIX for 2004 of 2.5 per cent. This was derived from an average pre-tax cost of equity of between 12.0 and 20.9 per cent, an average cost of debt of 6.9 per cent and a gearing of between 60 and 70 per cent debt. The detail of the calculations is set out in the sections below. 7. 8. Lenders’ estimates of cost of capital 9. Table 1 shows the lenders’ calculations of their WACC, submitted as part of their responses to the financial questionnaire (or in the case of Provident, its response to the issues statement). [ ] per cent 1999 Provident Cattles LSB S&U Source: Responses to the financial questionnaire, response to the issues statement. 2000 2001 2002 2003 2004 Simple average TABLE 1 Lenders' estimates of their costs of capital 10. [ ] estimates for cost of capital lay within the range of our calculations. The simple averages of [ ] estimates of their cost of capital were both in excess of our range. This was mainly due to different assumptions about gearing: [ ] assumed only [ ] per cent debt in its calculations (a significant difference from its target book gearing of [ ] per cent debt) and [ ] considered its home credit business to be principally financed by retained profits, which would have made its cost of capital approximate to its cost of equity. Calculating the WACC 11. The pre-tax WACC is calculated from the following formula: WACC = (1 – g) x Ke x (1/(1 – t)) + g x Kd Where: g is the gearing level (debt divided by the sum of debt and equity) Ke is the cost of equity Kd is the cost of debt t is the Corporation Tax rate. This is assumed to be 30 per cent. Each of the elements of the WACC is discussed below. 1 The real WACC is calculated by discounting the nominal WACC by the average annual RPIX between 1999 and 2004 of 2.28 per cent. A3(5)-2 Gearing and capital structure 12. Our approach to determining the capital structure of a large home credit lender which minimized its WACC was to examine the gearings of the five listed lenders and come up with a range of possible gearings. Market versus book values 13. Gearing can be based on market or book values of debt and equity, or from firms’ estimates of their target gearing. If the book value of equity is lower than market value, then a business will have a higher gearing using book values (ie a greater proportion of debt in its capital structure) than it would using market values. It would generally follow that the WACC would be lower because of the higher proportion of debt in the capital structure, although this higher book gearing would lead to higher costs of both equity and debt than under market gearing. 2 Therefore, in addition to evidence on book values we also considered possible market values and target gearings in our assessment. One lender [ ] said that gearing should be calculated from the market values of debt and equity, since these are the values on which investors’ expectations of returns are based and also from which other variables in the WACC such as beta and debt premium are based. However, accurate market values of debt and equity may be difficult to obtain where lenders have operations other than UK home credit within their corporate group. All the five large lenders we examined had other activities besides UK home credit. The difference between market and book value of the assets employed in these other activities may have been different from that for UK home credit assets. This is especially true where the other activities had higher growth expectations than UK home credit. Even if it were possible to isolate the market value of UK home credit assets, it would not necessarily be reasonable to use this value as a basis for gearing. This is because if companies were earning persistently high returns as a result of a lack of effective competition in the market, the market value of their equity would be higher than if they were operating in a very competitive market. Our objective in calculating a cost of capital for a typical large UK home credit lender is to provide a comparator to lenders’ actual returns and determine whether excessive profits are being made. It would therefore be somewhat self-defeating to reflect these possibly excessive profits in our comparator. Because of the problems with using market values set out above, we decided to use book values as an alternative to market values. We tested the reliability of this alternative by examining evidence on the differences between book and market values of equity. We looked at acquisitions of home credit companies and debt portfolios with book values above £10,000 between 1999 and 2004 to see the difference the acquisition price and the book value of assets acquired. We used data submitted by three lenders [ ]. One lender [ ] did not provide detailed information, and in any case it said that its acquisitions were small both in volume and value terms. One lender [ ] had made no acquisitions of home credit companies or portfolios over £10,000. 14. 15. 16. 17. 18. This takes bankruptcy costs into account. If these costs are ignored then the cost of debt would remain constant at different gearings. 2 A3(5)-3 19. Table 2 shows the results of our analysis. For the 13 acquisitions of companies (which will include assets other than customer debt) between 1999 and 2004, the ratio of market value (represented by the acquisition cost) to book value ranged between 2.1 and 0.5. The average ratio was 1.18 or 1.52 when weighted by assets acquired. These ratios recognize any possible excess profits purchased. TABLE 2 Market to book value ratios in acquired home credit companies and portfolios, 1999 to 2004 Companies and portfolios Number of acquisitions Total value of assets acquired (£m) Total consideration (£m) Highest market to book ratio Lowest market to book ratio Average market to book ratio Weighted average market to book ratio (by assets) Source: CC analysis of lender submissions. 82 37.8 53.1 2.47 0.42 1.08 1.41 Companies only 13 24.3 36.9 2.10 0.50 1.18 1.52 20. One lender [ ] submitted an analysis of gearing which used market values of equity. By comparing this with our existing calculations of book gearings we were able to determine the market to book ratios of equity for each year for each of these lenders. Figure 1 shows these market to book ratios. The chart shows a high degree of variation in ratios, with the highest [ ] being 6.4 and the lowest [ ] being 0.3. There was also significant variation between years for some lenders (for example, [ ] varied between 6.4 and 1.3), and the average ratio for each year fell from 3.2 in 2000 to 1.3 in 2004. We did not see any significant change in the cost of [ ] debt as a result of the large change in market to book ratios and hence to gearing. This suggested to us that the cost of debt would be similar for the range of gearings we used in our calculations. 21. A3(5)-4 FIGURE 1 Market to book value ratios for four large home credit lenders’ corporate groups 7 6 5 4 3 2 1 0 2000 2001 2002 2003 2004 Source: [ ], CC analysis. 22. From these two sets of data we concluded that market and book values of home credit lenders’ equity did sometimes show significant differences. However, these differences varied markedly between companies and years, and in a number of instances the market value of assets was below the book value. The variations in the market to book ratios of the four large listed lenders appeared to support our view that the ratios for home credit are difficult to separate from the lenders’ other activities. The substantial variations in the market to book ratios for acquired home credit businesses also made it difficult for us to determine a suitable value to adjust book equity values to reflect potentially different market values. We therefore have used the book values of equity and debt in our assessment of gearing. However, we recognized that even in a market where no excess profits were being made, the market value of a company may be greater than its book value since the latter values assets prudently and does not recognize certain intangible assets. This issue is discussed further in Appendix 3.6. From our provisional work on estimating lenders’ intangibles, we believed that were intangible assets to be recognized, a typical large home credit lender’s assets would increase by approximately 5 per cent. This would alter the gearing we used in our calculations only by a very small amount (for example, a 40 per cent equity and 60 per cent debt structure would change to 41 per cent equity and 59 per cent debt) and would have an immaterial effect on the WACC. Therefore, we did not propose to make any adjustment to book values. 23. Estimates of gearing 24. Information on gearing came from two sources. First, the lenders were asked how their UK home credit business was financed, to provide an estimate of their group’s capital structure and cost of capital, and to state whether different costs of capital were applied to their UK home credit and other businesses. Secondly, the lenders A3(5)-5 sent us detailed financial information for their home credit operations that showed equity shareholders’ funds and debt, from which we calculated gearing ratios for each financial year. 25. 26. 27. 28. 29. 30. [ [ [ [ [ ]3 ] 4,5,6 ]7 ] 8,9,10 ] Table 3 summarizes the gearing ratios discussed above. The gearings varied from year to year, so we have presented them as a range showing the figures for the years where the proportion of debt was highest and lowest. There were similarities between the three largest lenders’ own estimates, suggesting an optimum gearing of between 70 and 78 per cent. From the CC’s calculations using financial information submitted by the lenders the range appeared to be broader, between 39 and 67 per cent. We consider that the former range is a better indication of optimum gearing for two reasons: first, this is the range that the lenders themselves suggested, and secondly, that the lower number arising from the financial information submitted by the lenders may have been a function of differences in the allocation of debt and equity between group businesses. per cent As submitted in response to the financial questionnaire Highest Lowest From CC calculations of financial information supplied by the parties Highest Lowest TABLE 3 Estimates of lenders' gearing Debt as a % of debt + equity As submitted in response to the financial questionnaire: Provident submitted Cattles submitted LSB submitted S&U from annual accounts Source: Responses to the financial questionnaire, CC analysis. 31. 32. [ ] 11 The gearing range of 70 to 78 per cent debt applied to the three largest companies in the home credit market. However, gearing levels may change over time as companies usually are not able to optimize their capital structure in the short term. We therefore used a gearing range of between 60 and 70 per cent debt in our calculations. Our use of a gearing range with a greater equity component also covered the 3 4 [ [ 5 [ 6 [ 7 [ 8 [ 9 [ 10 11 ] ] ] ] ] ] ] ] ] response to cost of capital working paper, 17 August 2005, p5. [ [ A3(5)-6 possibility that the market values of home credit assets are greater than their book values. 12 The sensitivity of the WACC to changes in gearing is tested in Annex A. The cost of equity General approach and conclusions 33. We followed the CC’s market investigation guidelines, which state that we generally look to the capital asset pricing model (CAPM) when considering the cost of capital, but have regard to alternative models where appropriate. Given the availability of data and the nature of the industry, we decided to use CAPM to consider the cost of equity. The CAPM postulates that the opportunity cost of equity (Ke) is equal to the return on risk-free securities (Rf) plus the company’s systematic risk as measured by beta (β) multiplied by the equity risk premium (Rm – Rf): Ke = Rf + β (Rm – Rf) 35. We estimated that the beta for a UK home credit company would be between 0.85 and 0.95. We calculated that the nominal risk-free rate had averaged 4.47 per cent during the period from 1999 to 2004 based on redemption yields for UK index-linked gilts. We estimated a range for the equity risk premium of between 3 and 5 per cent based on studies of long-term historical returns and expected returns on international equity markets. Detailed explanations of these numbers are set out in the following sections. Using these three elements, we estimated a range for the average nominal pre-tax cost of equity of between 10.0 and 13.2 per cent during the period from 1999 to 2004. 34. 36. Estimates of beta 37. We tried to determine the beta for a large UK home credit operation. As noted above, none of the listed lenders had the supply of home credit in the UK as its sole operation. This meant that the beta for their UK home credit business may differ from their company beta. Betas also may change according to the particular historic data observed—for example, if average monthly returns are being examined then the beta will vary according to the day on which the monthly period begins. We therefore sought to estimate a range for a large UK home credit beta using company betas. The effect of this range on the cost of equity and WACC was then tested using sensitivity analysis. We looked at betas calculated by the risk measurement service of London Business School. Table 4 shows the betas of the five listed home credit lenders, 13 together with simple and weighted averages. The betas ranged between 0.7 and 0.95, with the simple average ranging between 0.81 and 0.85 and the average weighted by market capitalization ranging between 0.86 and 0.89. It should be noted that this form of weighted average was heavily influenced by Provident’s beta. 38. For example, a business with a book gearing of 75 per cent debt and a market:book value ratio of 1.5:1 would have a market gearing of 67 per cent debt, which is within our 60 to 70 per cent range. 13 The betas are based on the previous 60 months of returns, and so cover the period 1999 to 2004. 12 A3(5)-7 TABLE 4 Home credit betas Dec 2003 Provident Cattles LSB S&U Park Group Simple average Weighted average 0.86 0.95 0.70 0.87 0.73 0.82 0.88 Mar 2004 0.86 0.94 0.69 0.90 0.75 0.83 0.88 Jun 2004 0.84 0.92 0.70 0.87 0.78 0.82 0.86 Sep 2004 0.83 0.93 0.71 0.84 0.76 0.81 0.86 Dec 2004 0.88 0.94 0.73 0.90 0.82 0.85 0.89 Source: London Business School Risk Measurement Service. 39. The five listed home credit lenders derived differing proportions of their total income from UK home credit operations, as shown in Table 5. The effect of other operations on each company’s beta is difficult to calculate, but we considered that the effect was likely to be small for Provident and S&U, which derived 70 per cent or more of their group earnings from UK home credit. Cattles told us that it did not consider a home credit beta to be significantly different to its group beta. 14 Therefore, we considered it unlikely that these non-home-credit operations would cause each company’s beta to be significantly different from a UK home credit beta. TABLE 5 Percentage of profit before tax from UK home credit, 2004 % Provident Cattles LSB S&U Park Group Source: Financial Questionnaire responses (Provident, Cattles, LSB), Statutory accounts (S&U, Park). Note: Park's home credit operation made a loss in 2004. Its contribution to group turnover was 7 per cent. 40. Another possible guide to a large UK home credit beta was provided by examining the beta for the market sectors in which the home credit companies operate. This would enable us to use the betas of firms with similar operations to home credit as a guide for a home credit beta. Provident, Cattles, S&U and Park were all listed in the ‘Consumer Finance’ sector; LSB was in the ‘Other Finance’ sector. In 2004 there were 13 companies in the Consumer Finance sector. However, the nine other companies had very different businesses to home credit, including small business leasing and financial advice. This suggested that the betas of these companies would not be indicative of a home credit beta. Furthermore, Provident and Cattles accounted for approximately three-quarters of the sector, measured by market capitalization, so sector betas would be heavily influenced by the betas of these companies. For the reasons above, we did not consider sector betas to offer an effective insight into home credit betas. In its own calculations of WACC, Provident used a beta of [ ], which was similar to our observed betas. Cattles used a range of betas between [ ] and [ ], which was above our range. We considered that our observed betas for Cattles were reasonable and took comfort from its view that its home credit and group betas were likely to be similar. LSB and S&U did not use the CAPM to calculate WACC; and Park did not calculate a WACC. 41. 14 Source: Cattles response to the Cost of Capital working paper. A3(5)-8 42. To summarize, we used a range for the observed beta of between 0.85 and 0.95, based on the range of lender company betas observed. We also considered the weighted (0.87) and simple (0.83) averages of the betas of the five listed home credit companies over five quarters to give comfort that we were not underestimating beta. The range for the beta set out above was based on observed gearing (ie the actual gearing observed using the market values of debt and equity). However, our estimate of optimum gearing was based on target gearings. Because these target gearings were higher than the observed gearings the cost of equity would be expected to rise to reflect the greater financial risk of the higher gearing. In the CAPM equation this is achieved through adjusting the beta. Betas can be regeared to give a theoretical beta for a company at different levels of gearing to market gearing. We took the average beta for each lender for the period 1999 to 2004 and re-geared it from the average market gearing for the period to the target book gearing. The regearing formula is set out in the annex. The results (set out in Table 6) showed an increase in the range of the beta. At the low end of our gearing range (40 per cent equity, 60 per cent debt) betas ranged between 1.03 and 1.58. At the high end of our gearing range (30 per cent equity, 70 per cent debt) betas ranged between 1.32 and 2.03. 43. 44. TABLE 6 Observed and regeared betas for listed UK home credit lenders Observed market gearing (0–44% debt) Provident Cattles LSB S&U Park Average 0.85 0.94 0.71 0.88 0.77 0.83 1.35 1.73 Regeared to 60% debt Regeared to 70% debt Source: LBS risk measurement service, Datastream, CC analysis. 45. Regearing has the effect of increasing the range of betas. For each level of gearing, we considered the beta for a typical home credit company to lie in a range between the highest and lowest values for the five large lenders in the table. This produced a range of 1.03 to 1.58 for the lower gearing (40 per cent equity) and a range of 1.32 to 2.03 for the higher gearing (30 per cent equity). A wide range has been used, reflecting the differing capital structures of the listed lenders. We would expect a typical ‘pureplay’ home credit lender’s beta to lie towards the middle of these ranges. The risk-free rate 46. The risk-free rate is observable from trading in liquid securities markets. The UK Government has issued index-linked securities (index-linked gilts), which are generally considered to have negligible default risk and inflation risk 15 and therefore provide a proxy for the risk-free rate. The redemption yield on these index-linked gilts provides a direct estimate of the real risk-free rate for differing maturities. The Bank of England makes regular estimates of index-linked yields for various maturities. Figure 2 shows the index-linked yields from February 1982 to March 2005. 15 Some inflation risk arises from the RPI being lagged by eight months. A3(5)-9 FIGURE 2 Redemption yields on index-linked gilts (estimated by Bank of England from a smoothed zero coupon yield curve) 5 4.5 4 3.5 3 Per cent 2.5 2 1.5 1 0.5 0 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 APR APR APR APR APR APR APR APR APR APR APR APR APR APR APR APR APR APR APR APR APR APR APR Short term (5yrs) Real Medium term (10yrs) Real Long term (20yrs) Real Source: Bank of England. 47. Figure 2 shows that the yields have been on a downward trend since the early 1990s. At November 2004 the prospective yield on long-dated (~20 years) indexlinked gilts was around 1.8 per cent, and the yields for medium-term (~ten years) and short-term (~five years) index-linked gilts were also below 2 per cent. All of these figures were below their averages for the whole period (about 3.5 per cent) but above the long-run real return on government securities (about 1.3 per cent from 1900 to 2000). The level of government borrowing is generally regarded as a key determinant of gilt rates. There was considerable need for government borrowing in the UK during the 1980s and this could have contributed to the historically high rates during that period. Other relevant factors for the downward trend in yields in recent years include the minimum funding requirement for pension schemes, which increased the demand for both conventional and index-linked government securities and thereby placing upward pressure on their prices. In its 2002 report on Manchester Airport and BAA, the CC used a range of 2.5 to 2.75 per cent for the real risk-free rate, derived from its estimate at that time of medium-term average indexed-linked gilt yields. In its report into Mobile Network Operators (2002), the CC used the same range. Taking all the above into account (and the fact that we are assessing past operating performance rather than forward-looking price caps), we derived an average indexlinked gilt rate for the period 1999 to 2004 based on annual data from the Bank of England on short-, medium- and long-term gilts. Adjusting for inflation to obtain the nominal risk-free rate produces an average of 4.47 per cent for the risk-free rate during the period under consideration. 48. 49. 50. A3(5)-10 TABLE 7 The nominal risk-free rate, 1999 to 2004 % Average annual yield from British Government Securities, 5, 10 and 20 year real zero coupon Average RPIX Risk-free rate* 2.14 2.28 4.47 Source: Bank of England, Office of National Statistics. *Compound rate. The equity risk premium 51. The equity risk premium (ERP) is not directly observable from market data because the future payout from equities, unlike that on bonds held to maturity, is uncertain. In the past the CC and regulators have used two methods to estimate the ERP: historical data showing the difference between the realized return on equities over the risk-free rate; and forward-looking data relating to investors’ current expectations of the ERP. 16 If it is assumed that the ERP is constant over time, then current and future ERPs can be estimated from a historical average of the difference between past equity returns and risk-free rates. Since equity returns tend to be volatile from year to year, it is common practice in the finance literature to consider returns over an extended period. Fama and French’s (2002) estimates of the ERP in the USA for the period 1872 to 2000 using the dividend growth model and average stock returns were 3.5 and 5.6 per cent respectively. They argued that the difference between the two estimates was largely due to the unexpected capital gains during the period 1951 to 2000, and accordingly judged that the ERP estimate using the dividend growth model was closer to the true expected value than an estimate using historical returns. Data obtained from Dimson, Marsh and Staunton (2002) has produced arithmetic averages of ten-year holding returns for UK equities for the period 1900 to 2000 of 6.08 per cent. The arithmetic average of ten-year holding returns on UK bonds over the same period was 1.4 per cent. This suggests a historical risk premium of 4.7 per cent. According to Wright, Mason and Miles (2003), the central estimate of the ERP was between 4 and 5 per cent using an arithmetic mean. This was derived from historical returns (from an international set of data over the 20th century) on equity capital of between 6.5 and 7.5 per cent (arithmetic mean). They added a commonly used estimate of the risk-free rate of the order of 2.5 per cent (based on a sample of data from around 1980). Dimson et al (2002) did not expect stock market investors to enjoy a repeat of the returns of the 1990s. In terms of excess return on equities relative to less risky government bills or bonds, their analysis suggests that investors should expect longrun out-performance by equities closer to 3 per cent a year on average in the future. Dimson, Marsh and Staunton (2005) 17 updated data for the period 1900 to 2004 52. 53. 54. 55. 56. The CAPM ERP is the expected return on equities minus a known risk-free rate. Dimson E, Marsh P, and Staunton M, Global Investment Returns Yearbook 2005, ABN AMRO and London Business School, 2005. 17 16 A3(5)-11 suggested, for the UK, an arithmetic average equity risk premium relative to bonds of 5.2 per cent. 57. Recent market inquiries conducted by the CC into Store cards and Domestic liquefied petroleum gas used a range of 3.0 to 5.0 per cent for ERP. There remains much uncertainty about the ERP and we continue to attach weight to both the historical evidence and the evidence of market expectations. Our view is that the ERP lies in the range 3.0 to 5.0 per cent. 58. The cost of debt 59. The cost of debt for a company is largely determined by its credit rating. A company with a higher credit rating can borrow more cheaply than one with a lower rating. The larger home credit lenders are assigned credit ratings by agencies such as Standard and Poors, Moody or Fitch. Even though the home credit companies may not have recently raised any debt, we can look at the costs of recent debt issues of similarly rated companies. We were able to obtain credit ratings for Provident and Cattles from Fitch. LSB did not appear to be rated, which was unsurprising given that its funding consisted of bank facilities and customer deposits. S&U’s main sources of debt funding were bank loans and preference shares, and Park did not have any debt at a group level for the 2004 financial year, so it was also unrated. Provident is rated BBB+ and Cattles is rated BBB by Fitch. 18 Fitch rates BBB debt as ‘good credit quality’, although it is the lowest investment-grade category. 19 We considered a BBB rating to be a reasonable one for a typical large home credit company. We calculated the average cost of BBB debt using data from Datastream based on yields on Merrill Lynch seven- to ten-year bond indices and on ten-year UK gilts over the period 1999 to 2004. The results are set out in Table 8. per cent Simple average 1999–2004 4.90 2.03 6.92 60. 61. TABLE 8 Cost of debt 1999 Ten-year gilts Spread for BBB debt Cost of BBB debt (pre-tax) 5.0 2.5 7.5 2000 5.3 2.3 7.6 2001 4.9 2.2 7.1 2002 4.9 2.2 7.1 2003 4.5 1.6 6.1 2004 4.9 1.3 6.2 Source: Thomson Datastream, Merrill Lynch. We therefore consider an appropriate cost of debt for a large UK home credit lender to be between 6.1 and 7.6 per cent, depending on the year we are examining, with an average of 6.92 per cent for the period under consideration. The weighted average cost of capital 62. Using the above gearing and cost of equity assumptions, we calculated an indicative range for the weighted average nominal cost of capital for the home credit industry. 18 19 Source: www.fitchratings.com on 16 February 2006. Investment grade means that the agency giving the rating thinks the issuer has strong creditworthiness. A3(5)-12 Table 9 shows estimated average pre-tax WACCs for the period 1999 to 2004. The ‘high equity’ and ‘low equity’ WACCs use the upper and lower limits of the chosen gearing range, with observed betas regeared to reflect these limits. TABLE 9 Revised WACC estimate Low equity Lower Risk-free rate (%) ERP (%) Beta (regeared) Tax (%) Cost of equity Cost of debt (%) Gearing (D/D+E) (%) WACC pre-tax (%) Source: CC analysis. 4.47 3 1.32 30 12.0 6.9 70 8.4 Midpoint 4.47 4 1.68 30 16.0 6.9 70 9.6 Upper 4.47 5 2.03 30 20.9 6.9 70 11.1 Lower 4.47 3 1.03 30 10.8 6.9 60 8.5 High equity Midpoint 4.47 4 1.31 30 13.8 6.9 60 9.7 Upper 4.47 5 1.58 30 17.7 6.9 60 11.2 We also estimated a WACC for a typical home credit lender for the single year of 2004. Table 8 shows the cost of debt in 2004 to be 6.2 per cent compared with the average for 1999 to 2004 of 6.9 per cent. Using the 2004 cost of debt figure, but retaining the 1999 to 2004 average cost of equity gives a WACC of 8.0 to 10.8 per cent. A3(5)-13 ANNEX A Formula for regearing betas The relationship between geared and ungeared betas is set out by the following formula: β g = βu + βu D (1 − t ) E Where β g = the geared beta, β u = the ungeared beta, D = the value of debt, E = the value of equity and t = the corporation tax rate. In order to regear a beta to reflect a target gearing, it must first be degeared by rearranging the formula above. The degeared beta can then be regeared to any level of gearing. A3(5)-14 ANNEX B Applying the cost of capital to smaller home credit lenders 1. This appendix sets out to calculate the WACC for a large home credit lender. This is because its primary purpose is to provide a cost of capital figure to be compared with the large lenders’ returns. This annex discusses whether the WACC that we have calculated is also applicable to smaller lenders. Conclusion 2. Although it is reasonable to suggest that the cost of capital may be slightly higher for smaller lenders than larger ones, it is difficult to calculate even an indicative range. However, the quality of the data received from small lenders has generally been poor, particularly with respect to their profits. Given this uncertainty around small lenders’ returns, the importance of an accurately constructed cost of capital to act as a comparator is much diminished. Cost of equity 3. It is considerably more difficult to calculate the cost of capital for smaller, non-listed businesses. In particular, private companies do not have an assigned credit rating or traded shares on which to calculate a beta, and partnerships and sole traders are not easily able to determine the cost of their ‘equity’. Nevertheless, investors in these smaller home credit lenders (be they shareholders, partners or proprietors) should recognize that there is an opportunity cost of investing in the business. This opportunity cost is the returns they would make from investing their money in a venture with similar risk. It is fair to assume that the investment with the most similar risk characteristics to a smaller home credit lender would be other home credit lenders. We can therefore say that the ‘cost of equity’ or opportunity cost of investment for a small home credit lender will be broadly equivalent to the cost of capital of a typical listed lender. Cost of debt 4. We also considered whether it would be reasonable to assume that a smaller UK home credit business would be able to raise debt at the same cost as a BBB-rated company such as Provident or Cattles. Lower-rated (and therefore higher-cost) debt is considered to be speculative or ‘junk’ standard, and therefore is unlikely to be typical of a UK home credit lender. However, a smaller lender might not be of the scale to be able to raise debt on capital markets and would need to use private financing options such as bank loans or mortgages. The cost of these financing options may vary considerably. A bank overdraft or unsecured loan for a small limited company might attract a high interest rate to reflect the credit risk involved. However, if the small home credit lender is a sole trader then they might be able to raise cheaper finance by secured borrowing against their personal assets such as their house. The large differences in costs between these potential types of finance make any estimate of cost of debt very difficult. A3(5)-15 ANNEX C CC calculations of gearing from information supplied in the ‘Financial Information.xls’ spreadsheet 1. As part of their response to the financial questionnaire, the five largest lenders filled in a spreadsheet containing, among other things, balance sheet information for the period 1999 to 2004. From this information we have calculated gearing, expressed as debt as a percentage of debt plus equity. The gearings calculated using this method differ significantly from the gearings used by the parties in their own calculations of cost of capital. We are not sure of the reason for this, although it is possibly due to difficulties in allocating capital between home credit and other group businesses. Park has been excluded because, as discussed in paragraph 29, its capital structure does not permit calculation of a meaningful gearing ratio. £’000 1999 Provident Equity shareholders' funds Debt Debt (%) Cattles Equity shareholders' funds Debt Debt (%) LSB Equity shareholders' funds Debt Debt (%) S&U Equity shareholders' funds Debt Debt (%) Source: Responses in 'Financial Information.xls' spreadsheet. Debt percentage is CC calculation. 2000 2001 2002 2003 2004 2. Gearing calculations for the five largest lenders A3(5)-16