Risk Management In Gold Mining Industry

Created by Mbaakoh Longinu, Coventry University

Dedication

To my dear father and mother Mr & Mrs Longinu Fombi

1

Created by Mbaakoh Longinu, Coventry University

Acknowledgement I am extremely grateful to Mr Robert Evans of the Coventry Business School for inspiring and guiding me through this dissertation. His perspective and guidance have been invaluable. Thank you to Mr Azoh-mbi Solomon and Mr Akonde John Best for their words of encouragement and support. This study has also benefited immeasurably from the comments and suggestions of Ijeoma, Karen and Agnese. Finally special thanks to Amanpreet for reading through the first draft. Her suggestions were extremely helpful and insightful.

2

Created by Mbaakoh Longinu, Coventry University

Table of Contents Dedication.............................................................................................................1 Acknowledgement.................................................................................................1 Table of Contents..................................................................................................3 Abstract.................................................................................................................5 List of Tables ........................................................................................................7 List of Figures .......................................................................................................8 1.0 Introduction .....................................................................................................9 1.1 Research Question .................................................................................................. 12 1.2 Research Objectives................................................................................................ 14 2.0 Literature Review ..........................................................................................15 2.1 Risk Management Theories .................................................................................... 16 2.11 Financial Distress Model .................................................................................. 16 2.12 Tax Incentives and Hedging ............................................................................. 17 2.13 Management Incentives and Risk Aversion ..................................................... 19 2.14 The Underinvestment Problem ......................................................................... 20 2.15 Financial Sophistication Hypothesis................................................................. 21 2.2 Empirical Evidence on the Determinants of Corporate Risk Management............ 21 2.21 Empirical Evidence from Surveys ........................................................................ 22 2.22 Empirical Evidence from Cross Sectional Studies ........................................... 22 2.3 Empirical Evidence in the Gold Mining Industry................................................... 25 3.0 Research Methodology .................................................................................26 3.1 Research Perspective .............................................................................................. 27 3.2 Research Approach ................................................................................................. 28 3.3 Research Design...................................................................................................... 29 3.4 Data Collection Methods ........................................................................................ 31 3.5 Sample Description................................................................................................. 32 3.61 Measuring Financial distress............................................................................. 34 3.62 Measuring Investment Opportunity .................................................................. 34 3.63 Measuring Tax Convexity................................................................................. 35 3.64 Alternatives to Risk Management..................................................................... 35 3.7 Construction of the dependent variable .................................................................. 37 4.0 Data Analysis ................................................................................................41 4.1 Univariate Analysis................................................................................................. 42 4.2 Regression Analysis................................................................................................ 44 4.3 Presenting Regression Results ................................................................................ 46 4.31 Results of Financial Distress Variables ............................................................ 47 4.32 Results of Investment Opportunity Variables................................................... 49 4.33 Results of Firm Size.......................................................................................... 49 4.34 Results of Taxation ........................................................................................... 50 5.0 Conclusions ..................................................................................................51 Appendixes 1 Global Positions in OTC derivative market...................................55 Appendix 2 Calculations of Independent Variables.............................................55 Appendix 3 Company year end derivative positions ...........................................57 Appendix 4 Determination of Portfolio delta........................................................58

3

Created by Mbaakoh Longinu, Coventry University Appendix 5 Summary of Pooled Data .................................................................61 List of References and Links...............................................................................62

4

Created by Mbaakoh Longinu, Coventry University

Abstract Purpose – In the last two decades a number of studies have examined the risk management practices within the gold mining industry. For instance some studies report on the use of derivatives in the North American Gold mining companies.

Yet,

another

group

of

researchers

have

investigated

the

determinants of corporate hedging policies. This and other studies of similar focus have made important contributions to the literature. This dissertation uses four mining companies including one based in South African to shed light on some determinants of corporate hedging. The determinants examined include financial distress hypothesis, the Underinvestment Problem and the Tax Incentives to hedging. Furthermore the report investigates the existence of alternatives to risk management. Design/methodology/Approach- This dissertation presents the results of case study of four companies: Barrick Gold, AngloGold Ashanti, Kinross Gold Corporation and Agnico-Eagle Ltd. Using the linear regression model the work focuses on testing for statistical significance some of the theoretical determinants of corporate hedging decisions. Furthermore, it investigates the extent to which the results are consistent or inconsistent with previous empirical works. Findings- The results indicate that companies with high leverage are more likely to hedge consistent with the financial distress model. However the results also indicate an inverse significant relationship between cash costs and hedging. That is over the period examined companies’ reduced hedging activity despite increases in production contrary to popular theory. The results also show that larger firms are likely to hedge than smaller firms. Large firms benefit from scale economies and that information and transaction considerations have more influence on hedging activities than the cost of raising capital. Other effects measured such taxes, investment opportunity and cash balance found little evidence supporting the theoretical models underpinning them. Research Limitations - As with any case study, the small sample size severely limits the power of generalisation. Furthermore, the researcher could not verify if the linear regression model was the most appropriate for data analysis. Further

5

Created by Mbaakoh Longinu, Coventry University research could improve the power of the tests by including more detailed variables, different time spans and larger sample size. Originality/Value – To highlight the determinants of corporate risk management in the gold mining industry using four cases in environment of rising gold prices. Paper type - Dissertation

6

Created by Mbaakoh Longinu, Coventry University

List of Tables Table 1 10 Year gold price history in US $ per ounce……………………………11 Table 2

Firm yearly gold production……………………………………………….32

Table 2

Summary of variables……………………………………………………..35

Table 4

Sample data on Barrick Gold risk management activity……………….37

Table 5

Portfolio delta calculation of Barrick Gold……………………………….38

Table 6

Descriptive Statistics of Pooled data…………………………………….45

Table 7

Determinant of degree to which gold mining firms engage in price risk management using financial derivatives……………………………….46

7

Created by Mbaakoh Longinu, Coventry University

List of Figures Figure 1 Firms facing concave and convex tax schedule………………………17 Figure 2 Firm yearly gold productions…………………………………………….32 Figure 3 Percentage gold production hedged by firms………………………….41 Figure 4 Firm sizes as measured by total assets………………………………..42 Figure 5 Relationship between Leverage, Cash balance and Hedging Factor.43

8

Created by Mbaakoh Longinu, Coventry University

1.0 Introduction The corporate use of derivative products for risk management has grown rapidly over the last two decades. In 2004, the notional value of all over the counter (OTC) derivatives traded in domestic and international markets exceeded US $221 trillion, an increase of more than 1100% on the 1996 figure of US$20 trillion. Corporate risk management is thought to be an important element of the

9

Created by Mbaakoh Longinu, Coventry University overall business strategy both in financial and non financial institutions (El-Masry 2006). However, despite serious managerial and public policy implications, the rationales behind firm hedging decisions have remained unconvincing and mixed. Firms facing the same exposure have adopted different approaches to financial risk management through the use of derivatives. Derivatives have generally been used to manage three financial risks ♦ Commodity Price risks ♦ Interest rate risks ♦ Foreign exchange risk Commodity price risk forms part of business risk. It can be readily defined as risk faced by a business due the possibility of adverse changes in the price of commodities (Stephens 2001). Commodities are divided into three broad categories. The first category is agricultural products, the second category is metals and the third category is energy. Interest rate risk represents the companies’ exposure to fluctuations in interest rates. The debt structure of firms will possess different maturities of debt, different interest structures (such as fixed versus floating) and different currencies of denominations. Interest rates are currency –specific. Hence the multi currency dimension of interest rate risk is of serious concern to firms. Similarly foreign exchange risk is the risk faced by a business due variability in exchange rates. These risks could severely impact a firm’s financial stability. Several financial instruments have been developed over the years to manage these exposures. Hedging is the ‘taking of position, acquiring a cash flow, an asset or a contract that will rise (fall) in value and offset a fall (rise) in the value of an existing position (Moffett, Stonehill & Eiteman 2004:199)’. Hedging therefore protects the owner of an existing asset from loss. However it also eliminates any gain from an increase in value of the asset depending on the instrument employed. A brief review of the more widely instruments are presented below.

10

Created by Mbaakoh Longinu, Coventry University

Forward market (contract) The forward market is an over the counter trade involving contracting today for the future purchase or sale of a commodity or exchange rate. It is an agreement between a buyer and a seller for delivery of specified quantity and quality product at an agreed upon place and date in the future, in return for payment of an agreed upon price. They are not exchange traded and can be tailor-made to suit both parties. This feature distinguishes it from futures contracts, which are standardized contracts traded on an exchange. However commodity forward market does have some disadvantages such as credit risk to both parties. The use of forwards is associated with linear strategy of risk management as this eliminates all exposures the pay off is certain. Options An option is contract giving the owner the right, but not the obligation, to buy or sell a given quantity of an asset against a premium at a specified price (strike price) at some time in the future. There are two basic types of options namely calls and puts. An option to buy the underlying asset is a call, and an option to sell the underlying asset is a put. Because the option holder does not have to exercise the option if it is to his disadvantage, the option has a price, or premium.

Swaps A swap is an agreement between two parties to exchange a periodic stream of benefits payment over a prearranged period. The payments could be based on the market value of an underlying asset. The two parties to the contract are called the counterparties. Swaps are mostly used to manage interest rate exposure. Other derivative products commonly employed in financial risk management include futures, spot deferred contracts and synthetic products such as collars and floors. These products are used differently depending on the industry and type of risk faced. This research will seek to examine some of the

11

Created by Mbaakoh Longinu, Coventry University theoretical rational for hedging in the gold mining industry in the context of an increasing gold price trend.

1.1 Research Question There are several reasons to examine this industry: ♦ There is only one major source of risk- the risk of a fall in the price of gold. Table 1 illustrates the gold price movement over the last decade. Table 1 10 Year gold price history in US $ per ounce

www.goldprice.org

Table 1 shows the movement of gold price over the last decade. The 1990s saw gold prices averaging US$300.

Prices picked up in early 2002 and have

maintained the upward trend. A second reason for studying the gold mining industry is that

12

Created by Mbaakoh Longinu, Coventry University ♦ There exist liquid markets for derivatives based on gold and so there are hedging vehicles available to hedge the risks. ♦ Gold mining firms provide detail information on their hedging activities (more so than most other industries, as they provide details of their hedging activities in their quarterly reports). ♦ Even though many gold producers hedge the price risk, some do not, leading to strongly opposing views among mining firms on the desirability of hedging. ♦ Gold is a unique commodity, and the factors that influence its price make for an interesting analysis of the advantages and disadvantages of hedging A substantial amount of research has been carried to test the various corporate theories on risk management in the gold industry with mixed results. Theorists continue to advance new rationales for corporate risk management while researchers seeking to test these theories have been held back by the lack of reliable data (Tufano 1996). Furthermore the studies carried so far do not cut one way or the other. Several theoretical rationales have been advanced for why companies hedge. They include the financial distress theory which states businesses with high debt levels tend to hedge more. Secondly the underinvestment theory posits that businesses with

investment opportunity

would hedge more so as to secure financing while the tax convexity explanation suggest businesses facing a convex tax shield tend to hedge more to lower the average tax bill. The three hypotheses constitute the shareholder maximisation rational for hedging. Dionne and Garand (2002) and Allayannis and Weston (2001) found evidence shareholder maximization hypothesis. On the contrary, Tufano (1996) found little evidence in the gold mining industry. The second rational for hedging developed by Smith and Stulz (1985) is concerned with managerial incentives and risk aversion. Once again there is empirical evidence is mixed. This dissertation will highlight the aspects of the shareholder 13

Created by Mbaakoh Longinu, Coventry University maximization rational by seeking the answer the following questions using four gold mining firms. 1. What is the extent of hedging in the gold mining firms? 2. What firm characteristics significantly impact on hedging decision? 3. Why do some firms’ hedge and others do not? 4. Are there alternatives to hedging?

1.2 Research Objectives The questions discussed above constitute the basis on which the following objectives will be studied. ♦ Examine for statistical significance the financial distress theory ♦ Testing of the Underinvestment theory in the pooled data ♦ Investigate the significance tax shield on the hedging variable ♦ Examine the use of alternative strategies to risk management This research is divided into five chapters. Chapter 1 introduces the concepts corporate risk management by identifying the different types of exposures faced by firms and the nature of the instruments used for its management. Additionally, discussion on the research questions and objective is examined. Chapter 2 highlights the dominant theories of why companies hedge including an evaluation key empirical studies and literature on hedging. This is followed by examination of the evidence in the gold mining industry. In the light of the discussion in chapter 2, chapter 3 describes the methodology to be applied. This is carried out by reviewing the firm characteristics (variables) that theory would use to explain the cross sectional disparity in risk management choices. Chapter 4 examines using poled data variation financial risk management practices. This is undertaken by testing for significance the

14

Created by Mbaakoh Longinu, Coventry University several firm characteristics when regressed against the extent of hedging as measured by firm delta. The results are evaluated in evaluated in the context of other studies. The last chapter concludes the research with a discussion on the implications of the findings for current theory and subsequent research on risk management. In addition, a detailed appendix depicting on the computations used in this research is provided.

2.0 Literature Review Hedging with financial derivatives is an integral part of most risk management structures. However the debate about its merit has been the subject of numerous academic discussion with both proponents and detractors coming to separate

15

Created by Mbaakoh Longinu, Coventry University conclusions. Two main groups of financial theory on hedging theory exist, most of which arrive at optimal hedging policies by introducing some friction to the classic Modigliani and Miller model. The first group assumes that managers hedge to maximise firm value while the second group predicts managers hedge for personal diversification purposes, or to maximise their personal utility (Stulz 1984 & Tufano 1996). According to Miller and Modigliani paradigm, risk management is irrelevant to the firm. Shareholders can do it on their own, for example, by holding well diversified portfolios. An extension of the shareholder maximisation theory is examined by Bartraun, Brown and Fehle (2004) who suggest that firm’s hedge after acquiring a certain level of financial sophistication. This section looks the theoretical motivations of hedging including the factors that might lead to more or less hedging. This followed by a review of some seminal works on hedging. The chapter ends with a discussion on the empirical evidence on hedging in the gold mining industry.

2.1 Risk Management Theories Corporate risk management is underlined by number theoretical underpinnings such as the financial distress models, Underinvestment theory, Tax incentives, financial sophistication and managerial incentive and risk aversion.

These

theories have been the subject extensive research in both financial and non financial firms.

2.11 Financial Distress Model Volatilities in cash flows can lead firms into situations where available liquidity is insufficient to meet fixed payment objectives such as wages, and interest rate payments especially for firms with huge amount debt. High leverage firms are most likely to face difficulties servicing debt in a falling gold price environment because of the debt covenants (Smith and Stulz 1985). Financial risk management can reduce the probability of such occurrences and thus lower the expected value of costs associated with expected financial distress by lowering cash flow variability (Smith & Stulz 1985). These costs include bankruptcy,

16

Created by Mbaakoh Longinu, Coventry University reorganisation or liquidation and situations in which the firm faces direct legal costs. Warner (1997) finds that these direct costs of financial distress are less proportional to firm size, implying that small firms are more likely to hedge. On the hand Block and Gallagher (1986) and Booth, Smith and Stulz (1984) argue that hedging programs exhibit informational economies of scale and that larger firms are likely to employ managers with specialized information to manage a hedging program. Therefore the relation between firm size and hedging remains an empirical question.

2.12 Tax Incentives and Hedging The structure of the tax code can make it beneficial for companies to hedge and therefore maintain some level of cash flow predictability. If a firm faces convex tax function, then hedging that reduces the volatility of taxable income reduces the firm’s expected tax liability (Smith and Stulz 1985). Graham and Smith (1998) and Mayers and Smith (1982) argue that for firm facing some form of tax progressivity, when taxable income is low, its effective marginal tax rate will be low. But when income is high, its tax rate will be high. If such a firm hedged, the tax increase in circumstances where income would have been low is smaller than the tax reduction in circumstances where income would have been high thus lowering expected taxes. From their analysis of 80000 firm observations, they found in approximately 50% of the case, corporations face convex effective tax functions and thus an incentive to hedge. In approximately 25% of the cases, firms face linear tax functions. The remaining firms face concave tax functions. Firms are most likely to face convex tax functions when: 1. Their expected taxable income is near zero 2. Their income are volatile 3. Their income exhibit negative serial correlation (hence the firm is likely to shift between profits and losses).

17

Created by Mbaakoh Longinu, Coventry University Much of the convexity is induced by the asymmetric treatment of profits and losses in the tax code. That is a zero tax rate on negative income, moderate progressivity and constant rate thereafter. The convex region is extended by tax preference items like investment tax credits and deferred taxation. Figure 1 and 1a illustrates the tax liability for firm facing either a convex or concave tax shield Figure 1 Firm facing concave and convex tax schedule

Adapted from (Smith & Stulz 1985: 293)

Figure 1a illustrates firm facing concave tax shield and figure 1b depicts the firm facing convex tax shield. This simplistic illustration highlights the benefits of hedging when firm is facing a convex tax schedule. The tax liability T1 for convex 18

Created by Mbaakoh Longinu, Coventry University schedule is less than for concave schedule leading tax savings. In fact Graham and Smith found that among firms facing convex functions, average tax savings from five percent reduction in volatility of taxable income were about 5.4 percent of expected tax liabilities. In extreme cases these savings exceeded 40 percent.

2.13 Management Incentives and Risk Aversion Managerial attitude to risk has been found in some studies to be a significant factor in determining the extent of hedging in some firms (Merton 1973 and Tufano 1996). Stulz (1984) and Smith and Stulz (1985) argue that managers are often unable to diversify firm specific risks. Most senior managers derive substantial wealth from the firm and consequently their financial position is highly undiversified. Consequently risk aversion may cause some managers to deviate from acting in the best interests of shareholders by allocating resources to hedge diversifiable risk (Stulz, Mayers & Smith 1985). They argue that unless managers are faced with proper incentives they will not maximise shareholder wealth. When a risk adverse manager owns a large number of the firm shares, his expected wealth is significantly affected by variations in the firms expected profits. Hedging changes the distribution of the firm’s payoffs locking in an expected cash flow, and therefore changes the managers expected utility. These arguments imply that, all else been equal, managers with more wealth in firm’s equity will have a greater incentive to hedge the firm’s risks.(Christopher, Minton & Catherine 1997: 1326).Thus firms that are closely held will be more likely to use derivatives. Consequently Smith and Stulz (1984) predict a positive relation between managerial wealth invested in the firm and the use of derivatives as the managers’ end of period wealth is more a linear function of the value of the firm. In support of the managerial ownership hypothesis Tufano (1996) contends that not only the level of management’s equity ownership, but also the form by which that equity stake is held, is related to firm’s risk management choices. Firms whose managers own more options tend to hedge less than those with equity ownership. This is so because as long as managers hold options, they are sheltered from downside risks. Therefore firms whose managers hold large number of shares of stock may be willing to hedge than those holding options on 19

Created by Mbaakoh Longinu, Coventry University the shares. Hence Smith and Stulz (1985) predict negative correlation between option holdings and derivative usage. This has serious implications for managerial compensation scheme because by increasing equity component of managerial compensation, firms can align managers’ incentives more closely with those of other shareholders. This alignment enables optimal investment decisions to be taken.

2.14 The Underinvestment Problem The investment and financing policies of a firm can be harmonized and integrated by risk management to increase shareholder value (Froot, Scharfstein, and Stein, 1993). They argue without risk management, firms will be forced reject potential investment project; projects with positive Net Present Value. Firms may underinvest because of expensive external cost of capital. When the firms’ cash flow is low, obtaining additional financing is very costly inducing firms to make suboptimal investment decisions. In this case derivatives can be used to lower the cost of capital through the financing and investing decisions (Bartaum, Brown & Fehle 2004). When leverage is high underinvestment problem can occur. Establishing sound risk management policy can limit the underinvestment costs by reducing the volatility of firm cash flow and firm value (Allayannis and Weston 2001). Admittedly firms facing significant growth and investment opportunities are likely to be plagued by the underinvestment problem (Bartraum, Brown & Fehle 2004). Hence Froot, Scharfstein, and Stein’s (1993) theory suggest that firms with key planned investment programs and costly external financing would be inclined to use risk management to avert the need to access costly external financing to continue these programs.

They also argue that

smaller firms are likely to hedge more to avoid the expensive costs of external financing. Various measures such as market to book ratio, research and development expenses to sales ratio, capital expenditure to sales, net assets from acquisitions to size are used for testing the underinvestment hypothesis.

20

Created by Mbaakoh Longinu, Coventry University

2.15 Financial Sophistication Hypothesis Bartraun, Brown and Fehle (2004) propose this alternative theory after examining empirical evidence on the theoretical motivations of firm use of derivatives. They suggest that because of ambiguous results on theoretical motivation for hedging, firms that hedge do so because of their ability to do so, regardless of other firm characteristics. Financial sophisticated companies described as firms with multiple industry segments, mature treasury and foreign equity listings. It follows then larger firms are more likely to fulfil these characteristics and are expected to hedge more than smaller firms. Their findings are in contrast to Warner (1997) who suggested that from a financing perspective small firms are expected to hedge more.

2.2 Empirical Evidence on the Determinants of Corporate Risk Management Early research on the use of financial derivatives as risk management tool has been inconclusive. The motivations and instruments used vary across industry and geographic presence. This led to hypothesis put forward by Bartraum, Brown and Fehle that firms simply hedge once a certain level of financial sophistication is reached (2004). Most empirical studies have followed the neoclassical work of Modigliani and Miller (1958) where financial risk management at the firm level create shareholder value when in inefficiencies in the capital market give rise to deadweight costs born by the shareholders. In addition early studies test hedging motives of firms on the basis of survey data. For instance, Bodnar et al. (1995); Bodnar et al. (1996); Javlilvand et al (2002) surveyed derivative usage among non financial firms.

21

Created by Mbaakoh Longinu, Coventry University

2.21 Empirical Evidence from Surveys Bodnar et al (1996). Survey 530 US non financial firms about the use of financial derivatives. Their findings indicate that large firms tend to use over the counter (OTC) products, while small firms tend to use a mixture of OTC and exchange traded-products. They also find that 80 percent of firms use derivatives to hedge firm commitments, and 44 percent of firms use derivatives to hedge the balance sheet. One of the key goals of hedging with derivatives is to minimise cash flow fluctuations. Similar survey evidence undertaken by Alkebaeck and Hagelin (1999) on Swedish non financial firms found the use of derivatives to be more common among larger than smaller firms and that the principal use of derivatives is for hedging purposes consistent with Bodnar et al (1996). However Bodnar and Gebhardt (1999) found distinctive differences between German and US non financial firms including the primary goal of hedging firms, firms’ choices of hedging instruments and the influence of market view when taking derivative positions. The choice of instruments varies across industries.

Based on

evidence for a global sample, non financial firms mostly use forwards (36 percent) to manage foreign exchange risk, while swaps (11 percent) and options (10 percent) are less popular (Bartraum, et al, 2003). For interest rate management, swaps are more frequently (29 percent); interest rate options are used as well, but less often (7 percent). Commodity price derivatives are generally used less frequently, and there are few differences across different instruments (3 percent for futures; 2 percent for options) with some variation across industry.

2.22 Empirical Evidence from Cross Sectional Studies The majority of theoretical models of corporate risk management indicate that derivatives use increases with leverage, the existence of tax losses, the proportion of shares held by directors, and the pay out ratio. On the other hand, the extent of hedging decreases with the interest coverage and liquidity (Smith & Stulz, 1985; Froot et al, 1993; Nance et al., 1993).

22

Created by Mbaakoh Longinu, Coventry University However, empirical studies find only weak and at times ambiguous evidence consistent with theory. Mian (1996) finds that there is empirical evidence on the determinants of corporate hedging decisions. Based on non survey data for a sample of 3002 firms the study provides evidence, which emphasize that hedging is desirable because it lowers contracting costs, financial distress costs (Mayers & Smith 1982; Smith & Stulz 1987) and external financing costs associated with capital market imperfections (Froot, Scharfstein & Stein 1993). The evidence is strong with respect to financial distress theory but weak in respect of underinvestment and tax models. However evidence is supportive of the hypothesis that hedging activities exhibit economics of scale. Grezy, Minton and Schrand (1997) analysed a sample of 372 Fortune 500 non financial firms in the United States. They find that firms with greater growth opportunities and tighter financial constraints are more likely to use derivatives to reduce the variation in cash flows or earnings that might otherwise preclude firms from investing in valuable growth opportunities. The evidence is in line with underinvestment theory (Shapiro & Titman, 1986; Froot, Scharfstein and Stein 1993). The underinvestment cost explanation for optimal hedging suggests without hedging firms are likely to pursue suboptimal investment projects. Hence derivatives may provide a valuable benefit to firms that use them rationally (Allayannis and Weston 2001). Graham and Smith (1999) and Graham and Rogers (2002) using simulation model investigate the tax incentive to hedge that a firm facing a convex tax function, hedging that reduces the volatility of the taxable income reduces the firm’s expected tax liability Among firms facing convex tax functions, average tax savings from a five percent reduction in volatility are about 5.4 percent of expected tax liabilities; in extreme cases, these savings exceed 40 percent. However they point out any such program must be compared to the cost of hedging. In addition for firms with convex effective tax functions, the tax savings of hedging are not mutually

23

Created by Mbaakoh Longinu, Coventry University exclusive from hedging the benefits of controlling the underinvestment problems, increased debt capacity. In contrast Haushalter (2004) in a study of oil and gas producers using the Tobit regression model found no conclusive evidence between firm’s risk management policy and tax function. These results highlight the difficulty in correctly capturing tax save due to hedging.

Both the

measurement and analysis of the variable involve substantial statistical challenges. Furthermore the complexities of certain tax code present an added dimension to the problem.

Some studies focused on specific industries or individual firms benefit from the availability of detailed data on exposure and corporate hedging activities. Admittedly these data ensure the calculation of more precise measures of the extent of hedging. In a study of a sample of 100 oil and producers in the US Haushalter (2000) finds evidence of a positive correlation between the extent of hedging and financial leverage supporting the theory that corporate risk management is used to alleviate financing costs. Secondly a positive correlation was observed between the decision to hedge and the total asset. This is consistent with the notion that companies can face significant economies of scale in hedging, particularly in setting up a hedging program and therefore increases firm value. Contrasting these findings Jin and Jorion (2006) found no relationship between derivative activities and firm value in the US oil and gas industry. Similarly, Brown (2001) undertakes a clinical study of a US based manufacturers’ use of FX derivatives and finds little support for the financial distress or other primary theories of risk management and instead proposes that earnings management, competitive factors in the product market, or contracting efficiency gains motivate hedging. Clearly statistical support for popular theories of derivative use is mixed. However Bartraum, Brown and Fehle found evidence supporting a ‘naïve’ hypothesis that firms simply hedge once a certain level of financial sophistication is reached. Their study examines the use of derivatives by 7319 firms in 50 24

Created by Mbaakoh Longinu, Coventry University countries that together comprise 80 percent of the global market capitalisation on non financial companies. Their results are supportive of the theory that derivative use increases firm value especially for firms using interest rate derivatives.

2.3 Empirical Evidence in the Gold Mining Industry In this industry commodity price exposure is transparent and easy to hedge by investors. Theory might predict that no firms manage gold price risk since investors can diversify the way the risks. On the contrary risk management is practised by over 85 percent of the firm in the industry (Tufano, 1996). Though faced with identical price exposure gold mining firms have adopted very different approaches to risk management. Hedging policy has been extensive studied in the gold mining industry but the results have been at best weak and inconclusive. Tufano, (1996) examined 48 North American firms and finds risk management practices are consistent with some extant theory. He finds virtually no relationship between risk management firm characteristics that value maximising risk management theories would predict. In contrast managerial risk aversion seems particularly relevant bearing out Smith and Stulz (1985) prediction that firms whose managers own more stock options manage less gold price risk, and those whose managers have wealth invested in common stock manage more gold price risk. Another study of 44 North American gold firms from 1991 to 2000 Jin and Jorion (2006) found no relationship between hedging activities and firm values as measured by Tobin’s Q. The Tobin Q is defined as the ratio of the market value of the firm to the replacement cost of the assets, evaluated at the end of the fiscal year. However, the findings of Dionne and Garand show that seven variables (deferred taxation, tax save, production cost, dividend pay out ratio, preferred shares, and firm size) related to maximizing the firm’s value significantly affect the decision to hedge the price gold (2000). They considered hedging decisions based on quarterly data and extended analysis over longer period.

25

Created by Mbaakoh Longinu, Coventry University Arguably empirical tests on the determinants of corporate hedging policy have yield mixed results. The results are significantly influenced in some cases by the sample size, complexity of variables measured and analytical model applied. In the light of the mixed results it is important for further research to be conducted especially in environment of higher gold prices. This research will be limited to testing the financial distress, underinvestment and tax incentive theories.

3.0 Research Methodology

Research methodology involves a description of the process, variables to be measured and analytical tools. It examines the relevance and appropriateness 26

Created by Mbaakoh Longinu, Coventry University research design to objective. This section examines the perspective, approach and research design adopted. Issues about sampling and data collection methods are also explored. The latter part looks at different firm characteristics and their relevance to the objective. Firm characteristics will be measured by the construction of six explanatory variables and one dependent variable. These explanatory variables will then be pooled and regressed against the extent of hedging given by hedge ratio.

3.1 Research Perspective Two research perspective; positivist and interpretivist are widely associated with management research (Collis and Hussey 2005). The positivist approach seeks the facts or causes of social phenomena, with little regard to the subjective state of the individual. Furthermore the researcher assumes the role of an objective analyst, making interpretations about data that have been collected in justifiable manner (Saunders, Lewis & Thornhill 2003). As a result the positivist perspective emphasises on a highly structured methodology to facilitate replication and quantifiable observations that lend themselves to statistical analysis. Critics of positivism argue that the social world of business and management is far too complex to be defined by laws in the same way as the physical sciences (Saunders et al 2003). They are argue that important insights into this complex world is lost if such complexity is reduced to a series of law-like generalisations. Interpretivism stresses the importance of complexity and uniqueness of business situations (Bryman and Bell 2003). The approach emphasises the importance of making sense of the world through our own experiences. It argues that if businesses are unique and the business environment is always changing then there is little value in law-like generalisations (Saunders et al 2003). The positivist perspective will dominate the research as much of the variables under investigation are easily quantifiable such as leverage ratio, quick ratio, and

27

Created by Mbaakoh Longinu, Coventry University portfolio deltas. However the use of simple linear regression model to evaluate the relation severely limits the generaliasability of the results. On the other hand the research will involve four companies in a particular context involving variables influenced by people, economic and social factors. This can be seen as interpretivist perspective. Hence both perspectives will guide the research approach.

3.2 Research Approach Research projects also involve the use of theory and the extent to which theory is explicit in the design of the project raises important questions about the approach being adopted. Two approaches have identified in literature: deductive and inductive. The deductive approach entails development of a theory or hypothesis, and designing a research strategy to test the hypothesis. On the other in inductive approach theory is developed out of the data analysis. The deductive approach involves the development of a theory that is subjected to rigorous test. It owes much to the thinking of scientific research and positivism (Saunders et al 2003). There are several important characteristics of deductive approach. First, there is the search to explain the causal relationships between variables. These variables must be quantifiable and hence quantitative data is paramount to any analysis. In order to ascertain causality controls are introduced to allow testing of hypothesis (Bryman and Bell). The controls would ensure the direction of causality is ascertained. The final characteristic of deductive approach is generalisation. However in order to generalise observations it is necessary to select sample of sufficient numerical size. Induction or theory building is the alternative approach to deduction. This involves the data collection, analysis and as end result the formulation of theory. One of the criticisms of deductive approach is that it enabled cause-effect link to make between particular variable without an understanding of the way in which humans interpret their social world (Saunders et al 2003). Developing such an 28

Created by Mbaakoh Longinu, Coventry University understanding is one of the strengths of deductive research. Research using deductive approach would be particularly concerned with the context in which events are taking place and involves the collection of qualitative data. This approach owes much to interpretivism. The deductive approach has been the dominant approach in research on the corporate risk management. This can justified in the sense that the data is readily available and empirical evidence on cause-link between variables are easier to measure (Bryman and Bell 2003). However the mixed nature of evidence suggests lack of detail understanding of the context and interaction among variables which an inductive approach might shed light on. For the purpose of this dissertation, the deductive approach will be applied for number of reasons. First, time constraints does not permit elaborate data collection needed to conduct inductive research. Secondly the data to be used is readily available for analysis making it less risky than otherwise would be with questionnaires and interviews associated with inductive approach.

3.3 Research Design Research design can take several forms (Saunders et al 2003) •

Experimental design

•

Cross sectional

•

Longitudinal design

•

Case study design

•

Comparative design

•

Survey

Early studies on hedging in the gold industry were based on survey literature (Bodnar et al 1995; Alkebaeck and Hagelin 1999 Bailey, N. 1985). Tufano 1996 works on the practices of risk management in the gold industry focused on cross sectional data among 48 North American Gold mining firms. Other cross

29

Created by Mbaakoh Longinu, Coventry University sectional studies includes Adam, Fernando & Salas 2007; Dionne, & Garand 2000). This research will bear elements of case study design and cross sectional design. Collis and Hussey (2003) define a case study as ‘an extensive examination of a single instance of a phenomenon of interest.’ The importance of context is essential as it focuses on understanding the dynamics present within a particular setting. Yin (1994) identifies the following characteristics of case study research:
The research aims not only to explore certain phenomena, but to understand them within a particular context
The research does not commence with a set a of questions and notions about the limits within which the case study will take place
The research uses multiple methods for collecting data which may be quantitative and qualitative. However these characteristics are open to debate (Collis and Hussey 2005). They argue that if one is taking a more positivist approach one might wish to commence with strong theoretical foundation and specific research questions as is the case with this dissertation. Saunders contends that case study design often uses multi- cases to explore phenomena and is for a particular purpose. Some elements of cross sectional design will be introduced into this research. Cross sectional design ‘entails the collection of data on more than one case at a point in time in order to collect a body of quantitative or quantifiable data in connection with two or more variables, which are then examined to detect patterns of association’ (Bryman & Bell 2003: 48). Cross sectional design permits the examination of the relation between variables. Though establishing the directional of causality is problematic, useful inferences can deduce using appropriate statistical package.

30

Created by Mbaakoh Longinu, Coventry University

One major criticism of the case study and cross sectional design is that generalisation is difficult. This research seeks to test some of the theories on hedging within among four selected companies. Clearly the results might not generaliasable but important conclusions could be drawn.

3.4 Data Collection Methods There are two main approaches to data collection: quantitative and qualitative and each present a mixture of advantages and disadvantages. One of the main advantages of quantitative approach to data collection is the relative ease and speed with which collection can occur. However the analytical and predictive power which can be gained from statistical analysis must be set against the issues of sample representativeness, errors in measurement and quantification Collins and Hussey (2005). Qualitative data collection methods can be extensive and time consuming although it can be argued that qualitative data in business research provides a more ‘real’ basis for analysis and interpretation (Bryman and Bell 2003). Moreover qualitative approach presents problems relating to rigour and subjectivity. Data collected for this dissertation is mostly quantitative. There exist several ways to collect data for research purposes. These include:
Using secondary data
Through observation
Using interviews
Using questionnaires (Saunders, M 2003) This research involves analysis of pooled data over a five year period for four companies. Consequently secondary data has been extensive used. The main sources of data are company annual reports Form 10k disclosures and Form

31

Created by Mbaakoh Longinu, Coventry University 20F. Other sources include yahoo finance, financial times and Edgar-online database.

3.5 Sample Description Selecting a sample is a fundamental element of a positivist study. The sample size can determine the extent to which results are representative of the population. Several sampling methods exist some of which include:
Random sampling
Systematic sampling
Stratified sampling
Quota sampling
Cluster sampling etc (Collis & Hussey 2005) Among the methods quota sampling will be employed in this research. The aim of quota sampling is to produce a sample that reflects the industry in terms size. Most of theories on hedging in the gold industry have an element of firm size. Consequently size will be the key factor in the selection of the four companies. Size will measured in terms of average production per year over the last five years as well as the total assets. Large firms are considered as producing more than two million ounces per year. Average production per year has been used as close substitute for market capitalisation; the industry measure of firm size (Tufano1996). Table 2 and figure 2 illustrate the company’s gold production in million of ounces over five years. Table 2 Firm yearly gold production 2002/m

2003/m

2004/m

2005/m

2006/m

ounces

ounces

ounces

ounces

ounces

32

Created by Mbaakoh Longinu, Coventry University Barrick

5.695

5.51

4.958

6.161

6.639

5.46

8.643

Gold AngloGold Ashanti

6.516

Kinross

0.888634

Agnico-

0.26

1.62041 0.236653

6.765

6.182

1.653784

1.608805

1.476329

0.271567

0.241807

0.245826

Eagles ltd Data collected from annual reports

Figure 2

Gold Production 10 8 Production 6 (million/Oz) 4 2 0 Barrick

Kinross

Agnico- AngloGold Eagles Anshanti

2002 2003 2004 2005 2006

From Table 2 and figure 2 Barrick gold and AngloGold Anshanti are considered large firms while Kinross and Agnico- Eagles are termed small firms. All firms considered in the sample use derivatives as part of risk management strategy.

3.6

Construction of independent variables

This research is concerned with examining the significance of three of the major determinants often cited to justify risk management activities all of which were reviewed in chapter 2. It also seeks to test the existence of alternatives to risk management. The determinants include ♦ Reduction in expected costs of financial distress;

33

Created by Mbaakoh Longinu, Coventry University ♦ Increase investment opportunities or the underinvestment problem ♦ Reduction in expected tax payments These outcomes are measured by constructing variables to capture their likelihood. A summary description of the variables is shown in Table 3.

3.61 Measuring Financial distress Gold mining firms face financial distress if the price of gold falls below their costs of production often termed cash-costs. Cash costs is used on the basis that firms with high production costs are less efficient and more prone to financial failure. Additionally they are more likely to pay higher premiums to their partners. To measure the relative likelihood of financial distress data is collected on the firm cash costs. Another variable used to measure costs of financial distress is long term debt weighted according to market value (Tufano 1996; Dionne & Garand 2000). Measuring cash costs is more closely related to the probability of financial distress, whereas leverage has more to do with costs resulting from financial distress, supposing such costs are proportional to the face value (Dionne & Garand 2002). In this research long term debt is weighted according to total assets because this value was readily available. Theory predicts positive relationship between delta percentage and both cash costs and leverage. The data on cash cost and leverage for the companies is presented in Appendix 2.

3.62 Measuring Investment Opportunity Scharfstein and Stein (1987) theory predicts that firms with key planned investment programs and costly external financing would be inclined to use risk management to mitigate the need to access costly external financing to continue these programs. A decline in the price of gold could severely obstruct the investment programs of mining firms: exploration and acquisition. To measure the significance and importance of these activities, information is collected on the

34

Created by Mbaakoh Longinu, Coventry University firms’ annual exploration expenditure and net cash acquisition activities from both income and cash flow statements. Here again it scaled by total assets instead of market value. If risk management is used to protect the continued funding of these programs, theory predicts a positive relationship between these measures and the delta percentages. Similarly it is reasonable to suggest that transaction costs and information asymmetries for smaller firms are greater than for larger firms; at least for financing activities (Tufano 1996). Hence theory suggests from a financing perspective an inverse relationship between firm size and delta percentage. That is smaller firms might actively adopt risk management so as to avoid to seek costly external financing. For this research firm size is measured by total assets as shown in appendix 2. However reserves are common measure of firm size in the gold industry.

3.63 Measuring Tax Convexity Firms facing convex tax structure may lower average taxes through reducing fluctuations in earnings. The complex nature of tax structure has meant that no obvious variables have been agreed by researchers as appropriate for measuring the convexity of the tax structure (Dionne & Garand 2000). Graham and Smith (1999) formulated an equation allowing the computation of taxes saved as a result of risk management. This variable should have a positive effect on hedging. However for this research that used by Dionne and Garand (deferred income tax) will be employed. Tax credits for losses reduce deferred income thus the ratio measures the inverse of the tax function’s convexity. So a negative sign is predicted. The data is shown in Appendix 2

3.64 Alternatives to Risk Management Some firms pursue alternative activities that substitute for financial risk management strategies. Diversification could be undertaken instead of hedging

35

Created by Mbaakoh Longinu, Coventry University or they could adopt conservative financial policies such as carrying a large cash balance or maintaining a low leverage. Consequently firms that pursue this strategy should be associated with less financial risk management and lower delta percentage. However these do not represent explanations for financial risk management, but rather controls for substitute forms of risk management. To measure the existence of these alternatives information is collected on firm cash balances and firm leverage as shown in appendix 2. The quick ratio represents the degree of available cash balance in excess of current needs. This ratio is given determined by (cash and cash equivalents + receivable) dividend by current liabilities (McKenzie 2003). Table 3 Summary of Variables Delta % The delta is the variation in the value of the portfolio of the derivative products for every $1 variation in the price of gold. The aggregate value of the portfolio, calculated yearly is then divided by the firm’s gold production over the same period. The delta measures the level of derivatives used, that is the degree of risk management Cash Cost Average Cost to produce an ounce of gold. The cash cost is used to capture the likelihood of financial distress. When the price of gold decreases less efficient firms will be may be unable to pay current expenses. A positive sign is predicted. The annual value is used in this research. Total Assets Book Value of total assets is used as substitute for market value Long term Debt/ Total assets The book value of long term debt divided by the book value of the firm’s total assets. Debt generates obligatory interest payments. If the firm is unable to make its interests payments, it will get into financial hardship. This ratio therefore captures financial distress factor. A positive sign is predicted

36

Created by Mbaakoh Longinu, Coventry University Exploration Expenditures/ Total Assets The net cash payments for acquisitions divided total assets. This variable is collected on an annual basis. This variable captures the growth opportunities factor to the extent that exploration efforts are profitable. A non significant relationship is expected, since the correlation between investment opportunities and cash flows is positive in the gold mining industry. That is gold mining companies benefit from natural hedging. Acquisitions/ Total Assets The net cash payments due to acquisition activities divided by total assets. A non significant relationship is presumed, since correlation between investment opportunities and cash flow is positive in the gold industry. Deferred Income Tax/Total Assets The deferred Income item of the statement divided by total assets. Cash balance (Quick Ratio) Cash plus cash equivalent divided by current liabilities (McKenzie 2003). Liquidity can act as a cushion for financial disasters. They are thus substitute for risk management and a negative sign is predicted

3.7 Construction of the dependent variable The extent of risk management for each firm is determined by calculating the effective amount of ounces of gold that each firm has hedged, or sold forward denoted by delta. Rather than analyse each financial contract separately, the portfolio delta is calculated. Portfolio delta gives a measure of reported financial risk management activity and is regarded as the industry measure of investment portfolio. The delta represents the change in the price of the portfolio with respect to a small change in the price of the underlying asset (Hull 2003). Table 4 illustrates the risk management activities of Barrick Gold as reported in the annual report. As of 2002, the firm had committed to sell 365000 ounces of gold under forward sales commitments at an average price of $365. It had purchased put options expiring before the end of 2002 with an average strike price of $297/ounce on 160,000 ounces. Finally, it wrote call options on gold at a price of 37

Created by Mbaakoh Longinu, Coventry University $330/ounce on1330, 000 ounce, expiring before the end of 1991.The data for the three others is shown in appendix 3. Table 5 displays the determination of portfolio delta for Barrick Gold. For forward sales or spot deferred contracts, the delta is equal to -1 because there is no uncertainty that the transaction will occur; but for firms that hold options an effective portfolio delta must be calculated using the Black and Scholes formula (see Table 5) which takes into the account that the option will be exercised. Finally the firm’s total is calculated by dividing the sum of ounces whose price is covered over the same period by the total production for that year.

Hence delta expresses the equivalent number of

ounces (3564880 in 2002) that the firm would need to hold in a replicating portfolio to their hedged positions. In other words the firm had a gross short position in gold equal to 3564880 ounces of gold sold. In aggregate for 2002, for a $1 drop in the gold price, the market value of the firm’s gold portfolio should rise by $3564880. Although quarterly data for instruments would be more appropriate yearly data is collected because of time constraint. The portfolio delta calculations for rest of the companies are shown in Appendix 4. Table 4 Sample Data on Barrick Gold Risk Management Activity 2002

2003

Ounces

Price/US$

Ounces

Forward

2800

365

2800

Put sold

1600

297

Call options sold

1330

303

2004 Price

2005

Ounces

Price

Ounces

340

1350

345

1550

250

344

300

310

425

363

570

328

2006 Price

Ounces

Price

335

1540

338

300

317

250

332

550

336

1460

362

Table 4 shows the risk management activity of Barrick gold. All prices are in US$ and contracts are in thousands of ounces. Table 5 Delta of Barrick Gold Derivative Portfolio

38

Created by Mbaakoh Longinu, Coventry University Calculation of Portfolio Year

Delta

2002

2003

2004

2005

2006

Ounces/

Delta

Equivalent Ounces

Forward

2800000

-1

-2800000

Put sold

160000

-0.325

-52000

call sold

1330000

-0.536

-712880

Equivalent ounces

3564880

production

5695000

Delta percentage

62.59666

Forward

2800000

-1

-2800000

Put sold

250000

-0.309

-77250

call sold

425000

-0.622

-264350

Equivalent ounces

3141600

production

5510000

Delta percentage

57.01633

Forward

1350000

-1

-1350000

Put sold

300000

-0.108

-32400

call sold

570000

-0.849

-483930

Equivalent ounces

1866330

production

4958000

Delta percentage

37.6428

Forward

1550000

-1

-1550000

Put sold

300000

-0.076

-22800

call sold

550000

-.889

488950

Equivalent ounces

1083850

production

5460000

Delta percentage

19.85073

Forward

1540000

-1

-1540000

Put sold

250000

-0.11

-27500

call sold

1460000

-0.974

-1422040

Equivalent ounces

2989540

production

8643000

39

Created by Mbaakoh Longinu, Coventry University Delta percentage

34.58915

Black Scholes formula Call delta = ∆c = N ( d 1) e −δT where d 1 =

ln( S / K ) + ( r − δ + σ 2 / 2)T σ T

Put delta =∆p = - N ( − d 1 ) S = Stock price K = Strike price T = time to maturity assumed to be a year r = risk free rate ( average 10 year US Treasury note rate 5.1%

σ = Volatility of gold (average return standard deviation of annual gold returns over the past 30 years; 30.13%

δ = Annual Gold lease rate of 0.39% (Nitzsche & Cuthberston 2003: 270) Table 5 is a sample illustration of the portfolio delta of Barrick Gold at the end of year. The calculations assume that the options expire at the end of the year. The average spot prices for 2002, 2003, 2004, 2005 and 2006 are taken to be 310, 364, 410, 445 and 604 US$ respectively. The same methodology is used for all four firms to enable consistent results. The consistency of approach ensures the data collected is reliable and valid conclusions can be drawn. The variables measured are recognised industry benchmarks and have been used by Tufano (1996). However it might difficult to generalise conclusions from a case study especially given the fact that only four companies are studied. To attempt to overcome this factor the data will be pooled to obtain 20 observations as shown in Appendix 5. The calculated variables will be used in the next section to undertake univariate analysis with further examination using the linear regression model. Regression analysis will allow for the study of relation among variables including their strength and significance.

40

Created by Mbaakoh Longinu, Coventry University

4.0 Data Analysis This section involves the analysis and presentation of data. The first part presents a summary of the key findings univariate results including descriptive statistics. Any observable trend will be highlighted and further examined using 41

Created by Mbaakoh Longinu, Coventry University

the simple linear regression model.

The linear regression model allows

statistically significant observations to identify through the use of P value test. The results of the regression will then be evaluated against the theoretical models and previous empirical studies in the Gold industry.

4.1 Univariate Analysis The univariate analysis uses five year averages of firm characteristics for all four companies. Figure 3 below displays the extent of hedging among the companies as measured by the percentage of production hedged. Figure 3

Percentage production Hedged 50.00

42.37

40.00 30.00

28.05 Delta Percentage

20.00

14.54

10.00 0.00

1.07 Barrick AngloGoldKinross Agnico

Figure 3 illustrates the five year averages of percentage of total production hedged as measured by the hedging factor delta. The Y axis shows the values of the delta values while the companies are shown on X axis. The results show that Barrick Gold and AngloGold Ashanti are more active users of derivative as corporate risk management strategy while the remaining two companies have been less reliant on derivatives with Agnico- Eagles selling almost all of its entire production on the spot market. Secondly it could be argued from larger firms’ hedge more than smaller firms. Figure 4 displays the firm size values as measured by total assets.

42

Created by Mbaakoh Longinu, Coventry University

Figure 4

Firm Size as measured by Total Assets A quick look at figure 2 and 3 appear to show larger firms end to hedge more than 10000 8000 U$ thousands

6000 Total Assets

4000 2000 0 Barrick AngloGold Kinross

Agnico

It appears to show larger firms hedge more than smaller firms. This observation will need further examination using the simple linear regression model. Another variable that merits discussion from univariate analysis is the apparent relationship between hedging and the company long debt. Leverage has been used to measure the company debt position over the years. Figure 5 presents five yearly averages for leverage and cash balances measured against the hedging factor given by the delta percentage.

Figure 5

43

Created by Mbaakoh Longinu, Coventry University

Relationship between Leverage, Cash balance And hedging factor Agnico Quick Ratio Leverage Delta Percentage

Kinross AngloGold Barrick 0.00 10.00 20.00 30.00 40.00 50.00 Percentage

The figure above show that firms employing little risk management are barely distinguishable from those employing moderate to high levels of risk management, apart from carrying higher cash balances as predicted by theory. Analyses of the other variables do not yield substantial variations. Given the correlations among the different firm characteristics, these univariate tests do not reveal significant differences in firm traits, holding other attributes firm attributes constant. Thus multivariate tests would be appropriate. However this work the linear regression model will be employed.

4.2 Regression Analysis Regression analysis is used primarily for the purpose of prediction. The goal in regression analysis is to develop a statistical model that can be used to predict the values of dependent variable or response variable based on the values of at least on explanatory variable or independent variable (Berenson, Livine & Krehbiel 2002). The nature of the relationship between variables can take many forms ranging from simple to extremely complicated mathematical functions. In this research the linear model will be applied. The relationship between the variables could be positive linear in which case as the independent variable increases the dependent also increases while a negative linear relationship will

44

Created by Mbaakoh Longinu, Coventry University

involve one variable increasing while the other decreases. The sign coefficient of regression given by the slope of the regression line is indicative of the nature of the relationship. A positive coefficient relates to positive linear relationship and vice versa. To examine the ability of the independent variable to predict the dependent variable in the statistical model, several measures of variation have been developed. One of the measures is the coefficient of determination. It measures the proportion of variation of the dependent variable that is explained by the independent variable in the regression model. For example of 91% implies 91% of the dependent variable can be explained by the variability independent variable. This is an example of a strong positive linear between the two variables. To test for the significance of the relationship the t Test for the slope is employed. By setting a level of significance of 0.05, any p value < 0.05 is regarded as significant. However for regression analysis to hold three assumptions have to be satisfied ♦ Normality of Error ♦ Homoscedasticity ♦ Independence of Errors

The first assumption, normality, requires that the error around the line of regression be normally distributed at each value of the independent variable (X). The second assumption homoscedasticity requires that the variation around the line of regression be constant for all values of X. This means that the errors vary the same amount when X is a low value as when X is a high value. The third assumption, independence of error, requires that the errors around the regression line be independent of each value of X. This assumption is particularly important when data is collected over a period of time. In such situations the errors for a specific time period are often correlated with those of the previous time period. With these assumptions in mind the hedging factor is regressed against the firm characteristics in order to create a regression equation that can used to generate

45

Created by Mbaakoh Longinu, Coventry University

the strength of relationships among variables. The signs of coefficient indicate nature of the relationship while the p value determines the significance. For this research p = 0.05.This implies that all p < 0.05 is regarded as significant.

4.3 Presenting Regression Results In order to regress the variables the data five year observations for each company was pooled to produce a sample of 20 observations. Table6 Descriptive Statistics of Pooled Data

N Delta percentage 20 Cash cost ($US/oz 20 Leverage % 20 Exploration 20 activities( % Acquisition 20 activities (%) Deferred Taxation 20 (%) Quick Ratio 20 Total assets 20 Valid N (listwise )

Minimu m .0 177 9.17

Maximu m 62.6 690 43.89

Mean 21.401 274.05 31.1283

Std. Deviation 18.9741 112.659 10.89632

.45

2.56

1.2549

.69873

.02

222.85

24.8427 48.47892

.55

14.77

6.2420

.53

8.51 2.7238 2.25439 21373.0 4763.39 4993.28998 0 65

593.81

4.34859

20

Table 6 illustrates the descriptive statistics of the pooled data. Average the firms hedged 21.4% of production over the period under observation. The standard deviation of 18.97% is indicative of high degree of dispersion among the firms into active and moderate hedgers. The results of regressing annual percentage delta against the firm characteristics described in table 3 are shown in the table below.

46

Created by Mbaakoh Longinu, Coventry University

Table 7 Determinants of the degree to which Gold Mining Firms Engage in Price Risk Management Using Financial Derivatives.

INSTRUMENT VARIABLE Cash Costs ($/Oz) Leverage (%) Exploration Activities Acquisition Activities Deferred Taxation Quick Ratio Firm Size

COEFFICIENT

R2

P VALUE

-0.077 0.8334 7.520

21.2% 22.9% 7.6%

0.0406 0.03278 0.2371

-.0.03125

0.6%

0.7379

0.3025 -2.855 0.0016

0.48% 11.5% 19.9%

0.7714 0.1433 0.0481

The dependent variable for each firm year observation is the delta percentage, the percentage of estimated production that has effectively been sold short through financial contracts. The independent variables are defined in Table 3. The second column gives the regression coefficient while R 2 represents the coefficient of determination. The P value indicates the desired level of significance. P values less than 0.05 are shown in bold face type.

4.31 Results of Financial Distress Variables Table four suggests that the notions of corporate risk management on tendency of financial distress have some predictive power among firms in the gold mining industry. There negative sign for coefficient for cash suggest an inverse relationship between cost

of production as measured using cash costs and

hedging factor. This is contrary dominant theory dominant theory of a positive relationship. A p-value of 0.0406 suggests this relationship is statistically significant. In other words increases in production cost over the period of observation have been followed by decrease in hedging activity not increase. This result is contrary to Tufano (1996) who found no significant relationship

47

Created by Mbaakoh Longinu, Coventry University

between cash costs and Dionne and Garand (2000) who found significant positive relationship. Many factors can be attributed for the inverse relationship ship among which is the consolidation in the gold mining industry. Previous research was carried over periods when the gold mining industry was fragmented and the disparity between efficient and non efficient firms was great. Cash costs as measure probability of financial distress is based on the premise that less efficient firms are more likely to face financial difficulties. But the spate of mergers and acquisitions activities in the industry coupled with increasing gold price trend could account for the inverse relationship. That is the companies have tended to less hedge less though cash costs have been increasing. Similarly it could be argued that because gold prices have been rising over the last five years less efficient mines have been brought on board raising the average cost of production sustained by the increased revenue from sale on the spot market. In terms the of likelihood of financial distress as measured by the leverage scaled by total assets the results as shown in figure 5 are consistent with theory . A positive relationship is predicted and obtained in the results. This result is statistically significant with a p value of 0.03278. This result is in line Dionne & Garand (2000) and Haushalter (2000) who found a significant positive relationship between hedging and leverage. That is higher levered firms tend to hedge more than low levered firms. On the hand the results are contrary (Tufano 1996) Jin & Jorion (2006). Tufano (1996) argues that financial distress may be less of a rational of risk management in the gold mining industry because deadweight costs of bankruptcy may be small. As opposed to many other companies gold mines own tangible assets the produce an ‘unbranded’ commodity product with no after market price, leading to little loss of franchise value in terms of financial distress (Tufano 1996).

48

Created by Mbaakoh Longinu, Coventry University

4.32 Results of Investment Opportunity Variables The investment opportunities as measured by exploration and acquisition activities show no statistical significant observation. This result is consistent with Dionne and Garand 2002 but contrary to Allayannis and Weston (2001). Theory predicts positive sign for both exploration and investment activities. A positive relationship emerges for exploration activities and negative for acquisition activities. The non significance of the results is therefore contrary to the notion to that firms set up risk management programs to protect large on going investment programs. However these results might have been influenced the values of the acquisition figures used which represented the net cash figure shown in the cash flow statements. Tufano (1996) used the dollar value of attempted acquisitions from the acquisition and mergers database but found no significant relation between hedging and acquisition activities.

4.33 Results of Firm Size Total assets has been used a close substitute for firm size for the research. Theory predicts an inverse relationship between firm size and hedging at least from a financing perspective. As a rule, the largest firms have greater negotiation power and thus low financing costs, which reduces the need to hedge. However, most empirical research shows that larger firms tend to hedge more than smaller firms in support of the financial sophistication hypothesis. From Table 6 a positive significant relationship exists between firm size and use of financial derivatives. This result strengthens the univariate observation that Barrick tended to hedge more than Agnico- Eagle ltd. This positive association between firm size and hedging suggests that the relationship between size and hedging is more strongly influenced by economies of scale in risk management activities rather than financial distress models or costs associated with raising capital (Allayannis and Weston 2001)

49

Created by Mbaakoh Longinu, Coventry University

4.34 Results of Taxation The complex nature of the tax structure implies no consistent variables and results have been maintained for the extent of taxes on corporate hedging activities. Generally theory predicts an inverse relationship between firm’s deferred taxes and hedging. Deferred taxes measure the inverse of convexity. That is firms facing convex tax structure may lower average taxes by reducing fluctuations in earnings. The results from table 5 show no significant relationship between amount of deferred taxes and the extent of hedging. This implies the value taxation cannot be used to predict the extent of hedging by firms. Previous empirical research has found no consistent relationship between measures of tax – schedules and degree of derivative use. Nance, Smith, and Smithson (1993) find a positive relationship, but Grezy, Minton and Schrand (1995) do not. Generally the predictive power of the tax save function has been difficult to quantify with accuracy because of the complex nature of the tax system across countries. This has serious limited the ability of research work into tax incentive hypothesis to hedging. 4.35 Results of Cash balances (Quick Ratio) Some firms pursue alternative activities as a substitute for financial risk management. High cash balance can be used as buffer against adverse movement in prices. Univariate analysis showed firms with higher cash balances engaged less in risk management. This results appears be borne out after regressing the quick ratio against the hedging variable. A negative sign emerges as predicted that as firms accumulated high cash reserves they tend to hedge less. This appears to be the case for Barrick Gold which has substantial reduced its level of hedging with an increasingly high cash balance. That notwithstanding a p value > 0.05 makes the not statistically significant.

50

Created by Mbaakoh Longinu, Coventry University

5.0 Conclusions This paper studies the hedging activities of 4 gold mining companies between 2002 and 2006 and examines the relationship between gold hedging and three of shareholder maximization theories which include the financial distress models, underinvestment problem and tax save incentive. Pooling the data over the period generated 20 observations on which regression was done. The unique aspects of the study are: ♦ Use of financial statement footnotes to derive information on corporate

hedging decisions, instead of survey data as is typical of most previous work on hedging ♦ Use of case study approach focuses on just four companies and unlike

previous works focusing on mostly North American Mining firms this paper includes AngloGold Ashanti a South African based gold mining firm. ♦ This dissertation is among the few studies that have been carried out in

environment of rising gold prices and should shed considerable light on the light on the validity of theoretical underpinnings of hedging in the gold mining industry. Out of the four companies two are classified as large firm and the remaining two small firms based on their totals assets and yearly gold production. As far as the empirical tests of the determinants of hedging are concern, the relevant question is whether there is any statistical significance between firm characteristics (cash costs, leverage, investment activities, exploration activities, taxes, size, and cash balance) defined in this research as independent variable and the extent of hedging as defined by the percentage of yearly production that has effectively been shorted, delta. The evidence is mixed is with respect to models of hedging emphasising the likelihood of financial distress and determinant of hedging. Financial distress as

51

Created by Mbaakoh Longinu, Coventry University

measured by the cash costs has a significant inverse relationship with extent of hedging. Cash cost is used as measure of likelihood of financial distress with the assumption that less efficient firms (high production cost) are more likely to encounter financial difficulties and hence hedge more. However on the evidence of this work as cash costs are rising hedging has reducing. This evidence is contrary most works done in the gold mining industry including those of (Tufano 1996) who found positive but insignificant relations between cash cost and hedging among North American firms and Dionne and Garand who revealed a positive significant relationship between the two variables. Possible explanations for this result could be the correlation between the price of gold and hedging. In this sample the price of gold is inversely proportional to changes in delta hedging. In other words as the companies have tended to hedge less as prices have increased. Another explanation still related to price of gold could be as prices of gold have risen less efficient mines have been brought on stream leading to high average production cost of gold to rise. The second variable used to measure financial distress leverage with a p value of was found to be significantly positively correlated with hedging variable. The result show firms with high debt are more likely to hedge consistent with the works Stulz and Smith (1985) but at variance with Tufano who found no significant relation. The mixed nature of the findings highlight the issues associated with measuring an effect such financial distress. Another important observation is the strong positive correlation between firm size and hedging. This report supports the theory that larger firms are more likely to hedge than smaller firms. Large firms benefit from scale economies and that information and transaction considerations have more influence on hedging activities than the cost of raising capital. The result contradicts Warner (1997) external finance hypothesis that smaller firms are likely to hedge to avoid costly external finance.

52

Created by Mbaakoh Longinu, Coventry University

The underinvestment hypothesis is not supported in this dissertation as the report found little evidence in its favour. These findings could have been influenced however by the fact the value of acquisitions and exploration activities were net cash positions report on the cash flow statements and not the gross values. This report also finds little evidence that hedging strategies are motivated by tax saving strategies. Deferred Taxation which measures the inverse of tax convexity yielded no significant result. However as discussed earlier the documented problems associated with selecting a variable and different tax structures among the companies impacted the results. The evidence in this study suggest that not all aspects of the shareholder maximisation theory are valid hence the mixed results. Clearly the dissertation being a case study limits the generalization of results but presents a fresh perspective

on the debate on the merits and rational for risk

management. One major shortcoming of the project was the linear regression method used. A more appropriate would have been multivariate analysis which permits the interaction between variables to be isolated. There exist correlations between the variables and some of were strong enough to have influenced the results. Formal interpretations of these correlations require specifications of a simultaneous equations framework. This report did not examine the managerial aversion incentive to hedging which was found by Tufano (1996) to be more significant influence on hedging decisions than the concept of shareholder maximization. Further investigation of these issues is suggested as a line for future research. Additionally most studies in the gold mining industry have

concentrated on north American mining firms more

research should carried including mines from other parts of the globe to improve generalization of the results. This report makes that attempt by including South

53

Created by Mbaakoh Longinu, Coventry University

African based mining company. Subsequent work might increase the power of the tests used in this dissertation: ♦ Use of more data ♦ Use of continuous measure of hedging activity ♦ More effective separation and description of variables.

54

Created by Mbaakoh Longinu, Coventry University

Appendixes 1 Global Positions in OTC derivative market

Appendix 1

http://www.bis.org/triennial.htm

Appendix 2 Calculations of Independent Variables Barrick Gold Production/million (m) Financial Distress Cash costs/US $ long-term debt/ ‘000’ Total assets/’000’ Leverage Investment opportunity Exploration/US $ million Exploration/total asset Acquisition/US$ million

2002 5.695

2003 5.51

2004 4.958

2005 5.46

2006 8.643

177 1927 5261 36.628017

189 1864 5358 34.7891

214 2711 6274 43.21007332

227 3012 6862 43.89391

282 7173 21373 33.56103495

104 1.9768105 58

137 2.556924 334

141 2.247370099 821

141 2.054795 1180

171 0.800074861 1593

55

Created by Mbaakoh Longinu, Coventry University Acquisition/total asset Taxation Deferred Income tax/US$m Deferred/Tax

1.102452

6.233669

13.08575072

17.19615

7.453328966

155 2.9462079

230 4.292647

139 2.215492509

114 1.661323

798 3.733682684

2.373

3.308

4.18

2.403

2.087

Quick Ratio

AngloGold Ashanti Financial Distress Cash costs/US$ long-term debt/’000’ Leverage Investment opportunity Exploration/$ Exploration/total asset Acquisition/US$ million Acquisition/total asset Taxation Deferred Income tax/m$ Total assets/’000’ Deferred/total asset Production/kg Quick Ratio

2002

2003

2004

2005

2006

213 1886 43.35632184

225 2009 37.60059891

268 3516 37.42018

277 3815 41.86327225

308 3397 35.70903

28 0.643678161 305 7.011494253

40 0.748643084 11907 222.8523302

44 0.468284 4640 49.38272

45 0.493800066 4355 47.78887304

61 0.641228 80 0.840954

505 4350 11.6091954 184722 1.738

789 5343 14.76698484 174668 0.951

1158 9396 12.32439 188223 0.802

1152 9113 12.64128169 191783 0.639

1275 9513 13.40271 175263 0.544

Kinross Financial Distress Cash costs/US$ long-term debt/m Total assets/m Leverage Investment opportunity Exploration/m Exploration/total asset Acquisition/m Acquisition/total asset Production/m ounce Taxation Deferred Income tax/million Deferred/total asset

2002

2003

2004

2005

2006

201 92.4 598 15.45151

222 164.5 1794.5 9.166899

243 533.4 1834.2 29.0808

275 607.6 1698.1 35.78117

319 570.6 2053.5 27.78670562

2.7 0.451505 0.1 0.016722 0.888634

24.3 1.354138 81.9 4.563945 1.62041

25.8 1.406608 442.3 24.11406 1.653784

26.6 1.566457 257 15.13456 1.608805

39.4 1.918675432 257 12.51521792 1.476329

3.3 0.551839

54.1 3.014767

119.9 6.53691

129.6 7.632059

114.4 5.570976382

56

Created by Mbaakoh Longinu, Coventry University Quick Ratio

2.81

1.954

0.526

0.712

0.932

Agnico-Eagle Mines Ltd 2002 Financial Distress Cash costs/US$ long-term debt/’000’ Total Assets/’000’ Leverage Production/ounce Investment opportunity Exploration/m Acquisition/million Exploration/total asset Acquisition/total asset Deferred Income tax Deferred/total asset Quick Ratio

2003

2004

2005

2006

182 175.02 593.807 29.47422 260000 0.26

269 199.975 637.101 31.38827 236653 0.236653

56 213.446 718.164 29.72107 271567 0.271567

43 267.45 976.069 27.40073 241807 0.241807

690 197.148 1521.488 12.95758 245826 0.245826

3.766 66.609 0.634213 11.21728 20.899 3.519494 8.511

5.975 105.907 0.937842 16.62327 29.378 4.6112 4.19

3.584 94.832 0.49905 13.20478 17.684 2.46239 4.8

16.581 66.539 1.698753 6.817039 52.5 5.378718 3.455

30.414 299.723 1.998964 19.69933 90.793 5.967382 7.56

Appendix 3 Company year end derivative positions Barrick Gold Forward Put sold call options sold

2002 Ounces/’000’ 2800 1600 1330

Price/$ 365 297 303

2003 Ounces 2800 250 425

Price 340 344 363

2004 Ounces 1350 300 570

Price 345 310 328

2005 Ounces 1550 300 550

Price 335 317 336

2006 Ounces 1540 250 1460

Price 338 332 362

AngloGold Ashanti Derivative Position 2002

Forward Put purchased put sold call purchased call sold

kg 61.727 10.238 3.732 24.535 24.584

2003 price 299 312 273 338 340

ounces 15.289 5.808 12.752 4.555 18.83

2004 price 307 352 307 351 332

ounces 18.056 796 7.466 0.572 5.829

2005 price 313 291 317 360 330

ounces 34.021 757 6.21 9.88 29.49

2006 price 315 291 397 340 322

57

ounces 30.428 563 4.354 3.03 18.017

price 333 291 339 351 329

Created by Mbaakoh Longinu, Coventry University

Kinross Gold Corporation 2002

Forward Calls Put options bought

Ounces 113 50

2003 price 271 340

2004

Ounces 137.5 100

price 278 320

2005

Ounces 137.5 50

price 278 340

Ounces 200

150

250

150

2006 price 452

250

Ounces

Price

225

522

150

250

Agnico Eagle 2002 ounces

2003 price

ounces

2004 price

Put options bought

2005

2006

ounces

price

ounces

price

ounces

price

136.644

260

190.2

260

152

260

Appendix 4 Determination of Portfolio delta AngloGold Ashanti 2002 Forward Put purchased Put sold Call purchased Call sold Equivalent ounces Production Delta percentage

kg 61727 10238 3732 24532 24584

Delta -1 -0.386 -0.232 -0.506 -0.499

Equivalent Ounces -61727 -3951.868 -865.824 -12413.192 -12267.416 -91225.3 184722 49.38518

2003

Forward Put purchased Put sold Call purchased Call sold Equivalent ounces Production Delta percentage

15289 5808 12752 4555 18830

-1 -0.337 -0.19 -0.663 -0.727

-15289 -1957.296 -2422.88 -3019.965 -13689.41 -36378.551 174668 20.82726

2004

Forward Put purchased Put sold Call purchased Call sold

18056 796 7466 572 5829

-1 -0.074 -0.122 -0.874 -0.845

-18056 -58.904 -910.852 -499.928 -4925.505

58

Created by Mbaakoh Longinu, Coventry University Equivalent ounces Production Delta percentage

-24451.189 188223 12.99054

Kinross Gold Corporation Calculation of Portfolio Delta Kinross Ounces Delta Forward 113000 -1 calls 50000 -0.498 Aggregate equivalent portfolio (ounces) Production 2002 (ounces) Delta percentage

Equivalent Ounces -113000 -24900 -137900 888634 15.51819984

2003

Forward 273000 -1 Calls 100000 0.766 Aggregate equivalent portfolio (ounces) Production 2003 (ounces) delta percentage

-273000 -76600 -349600 1620410 21.57478663

2004

Forward 137500 -1 Calls 50000 0.82 Put bought 150000 -0.0257 Aggregate equivalent portfolio (ounces) Production 2004 (ounces) delta percentage

-137500 -41000 -3855 -182355 1653784 11.02653067

2005

Forward 200000 -1 Put bought 150000 0.01314 Production Aggregate equivalent portfolio (ounces) delta percentage

-200000 -1971 1608805 -201971 12.55410071

2006

calls 225000 0.78257 puts bought 150000 0.010876803 Aggregate equivalent portfolio (ounces) Production delta percentage

-176078.25 -1631.52044 -177709.7704 1476329 12.03727424

2002

Calculation of Portfolio Delta Agnico-Eagles Mines Ltd 2004

Ounces puts 136644 Aggregate equivalent portfolio (ounces) production Delta

2005

puts

190200

Delta 0.0344

Equivalent Ounces -4700.5536 -4700.5536 271567 1.7309

0.0182

-3461.64

59

Created by Mbaakoh Longinu, Coventry University Aggregate equivalent portfolio (ounces) production Delta 2006

puts 152340 Aggregate equivalent portfolio (ounces) production Delta

-3461.64 241807 1.431571 0.00094

-143.1996 -143.1996 245826 0.058252

Black Scholes formula Call delta = ∆c = N (d 1)e −δT where d 1 =

ln( S / K ) + ( r − δ + σ 2 / 2)T σ T

Put delta =∆p = - N ( − d1 ) S = Stock price K = Strike price T = time to maturity assumed to be a year r = risk free rate ( average 10 year US Treasury note rate 5.1%

σ = Volatility of gold (average return standard deviation of annual gold returns over the past 30 years; 30.13%

δ = Annual Gold lease rate of 0.39%

60

Created by Mbaakoh Longinu, Coventry University

Appendix 5 Summary of Pooled Data Stock Ticker ABX ABX ABX ABX ABX AU AU AU AU AU KGC KGC KGC KGC KGC AEM AEM AEM AEM AEM

Year 2002 2003 2004 2005 2006 2002 2003 2004 2005 2006 2002 2003 2004 2005 2006 2002 2003 2004 2005 2006

Delta percentage 62.60 57.20 37.64 19.85 34.57 49.39 20.83 12.99 27.79 29.24 15.52 21.57 11.03 12.55 12.04 0.00 0.00 1.73 1.43 0.06

Cash cost ($US/oz 177.00 189.00 214.00 227.00 282.00 213.00 225.00 268.00 277.00 308.00 201.00 222.00 243.00 275.00 319.00 182.00 269.00 290.00 410.00 690.00

Leverage % 36.63 34.79 43.21 43.89 33.56 43.36 37.60 37.42 41.86 35.71 15.45 9.17 29.08 35.78 27.79 15.45 9.17 29.08 35.78 27.79

Exploration activities( % 1.98 2.56 2.25 2.05 0.80 0.64 0.75 0.47 0.49 0.64 0.45 1.35 1.41 1.57 1.92 0.63 0.94 0.50 1.70 2.00

Acquisition activities (%) 1.10 6.23 13.09 17.20 7.45 7.01 222.85 49.38 47.79 0.84 0.02 4.56 24.11 15.13 12.52 11.22 16.62 13.20 6.82 19.70

Deferred Taxation (%) 2.95 4.29 2.22 1.66 3.73 11.61 14.77 12.32 12.64 13.40 0.55 3.01 6.54 7.63 5.57 3.52 4.61 2.46 5.38 5.97

Quick Ratio 2.37 3.31 4.18 2.40 2.09 1.74 0.95 0.80 0.64 0.54 2.81 1.95 0.53 0.71 0.93 8.51 4.19 4.80 3.46 7.56

61

Total assets 5,261.00 5,358.00 6,274.00 6,862.00 21,373.00 4,350.00 5,343.00 9,396.00 9,113.00 9,513.00 598.00 1,794.50 1,834.20 1,698.10 2,053.50 593.81 637.10 718.16 976.07 1,521.49

Created by Mbaakoh Longinu, Coventry University

List of References and Links Adam, T.R., Fernando, C. S. & Salas, J. M. (2007). Why do firms speculate? Evidence from the gold mining industry. Journal of Financial Economics 81, 283-309 Alkebaeck, P. & Hagelin, N. (1999). Derivatives usage by non financial firms in Sweden with international comparison. Journal of International Financial Management and Accounting 10 (2) 105-20 Allayannis, G. & Weston, P. (2001). The Use of Foreign Currency Derivatives and Firm Market Value. The Review of Financial Studies 14 (1) 243-276 Bailey, N., Browne, D., Hicks, E. & Skerrat, L. (2003). UK Corporate use of derivatives. The European Journal of finance 9, 169-193 Bartram, Soehnke, Brown, G.W. & Fehle, F. R. (2004). International Evidence on financial derivatives usage. Journal of Finance 4 (2) 342- 276 Berenson, M., Levine, D., & Krehbiel, T. (2002). 8th edn Basics Business Statistics: Concepts and Applications. New Jersey: Prentice Hall 488-490 Bessembinder, H. (1991). Forward Contract and firm value: Investment Incentives and Contracting Effects. European Economic Review 47 (1) 165-195 Block, S. B., & Gallagher, T. J. (1986).The Use of Interest rate futures and options by corporate financial managers. Financial Management 15, 73-78 Bodnar, G. M., Hayt, G., Marston, R.C., & Smithson, N. (1995). How corporations use derivatives. Financial Management 24 (2) 104-25 Bodnar, G. M, Hayt, G.M, Marston, R.C. & Smithson, N. (1998).Wharton survey of financial of derivatives usage by US non financial firms. Financial Management 25 (4) 113-33 Bodnar, G.M., & Gebhardt, G. (1999). Derivatives usage in risk management by US and German non financial firms: a comparative study. Journal of international financial management and accounting 10 (3) 153-87 Boyle, P. & Boyle, F. (2001). Derivatives: The Tools that Changed Finance. London: Risk Books 103-107 Brown, G.(2001). Managing Foreign Exchange Risk with derivatives. Journal of financial economics 60 (2/3) 401-449

62

Created by Mbaakoh Longinu, Coventry University

Bryman, A. & Bell, E. (2003). Business Research Methods. NewYork: Oxford University Press: 45-67 Collis, J., & Hussey, R. (2005). 3rd end Business Research. New York: Palgrave Macmillan: 70-78 Geczy, C., Minton, B. & Schrand, C. (1997). Why firms use currency derivatives. Journal of finance LII (4) 1326 Dionne, G. & Garand, M. (2000). Risk Management Determinant Affecting Firms’ Values in the Gold Mining Industry: New Empirical results. Ecole des HEC-Montreal and Thema 2000 (48) 4-13 El-Masry, A. (2006). Derivatives use and risk management practices by UK non financial companies. Managerial Finance 32 (2) 137-169 Froot, K. A., Scharfstein, D. A. & Stein, J. C. (1993). Risk Management: Coordinating Corporate Investment and Financing Policies. Journal of finance 48:5: 1629-1658 Graham, J.R. & Smith, Jr. (1999). Tax Incentives to hedge. Journal of finance 54:6 2241-2263 Hull, J. C. (2003). 5th edn. Options, Futures and Other Derivatives. New Jersey: Prentice Hall: 30-49 Javlilvand, A., Switer, J., Tang, C. (2002). A global Perspective on the use of derivatives for corporate risk management decisions. Managerial finance 26 (3) 29-39 Jin, Y. & Jorion, P (2006). Firm Value and Hedging: Evidence from the U.S Oil and Gas Producers. Journal of Finance 61 (2) 893-919 Jin, Y. & Jorion, P. (2006). Does Hedging increase firm value: Evidence from the Gold Mining Industry? Journal of Finance 2-29 Haushalter, G.D. (2000). Financing Policy, Basis Risk, and Corporate Hedging: Evidence from oil and gas producers. Journal of finance LV (1) 23-45 Lealand, H. (1998). Agency costs, risk management, and capital structure Journal of finance 53:4 1213-1243 Mayers, D. & Smith, C.W. (1982). On the Corporate demand for Insurance. Journal of business, 55 281-296

63

Created by Mbaakoh Longinu, Coventry University

McKenzie, W. (2003).3rd edn. The Financial times guide to using and interpreting company accounts. London: Prentice Hall: 204-210 Mian, S.L. (1996). Evidence on Corporate Hedging Policy. Journal of Financial and Quantitative Analysis 31 (3) 419-439 Moffett, M,. H., Stonehill, A., & Eiteman,. D. (2004). 10th edn Multinational Business Finance. Boston: Pearson Addison Wesley: 199 Moffett, M. H. (2001). Cases in International Finance. Boston: Addison Wesley Longman: 10-27 Myers, S.C. (1977). Determinants of Corporate borrowing. Journal of financial economics5: 147-175 Myers, S.C.(1984). The capital structure puzzle. Journal of finance 39:3 357-592 Myers, S.C (1993). Still searching for optimal capital structure. Journal of finance.203-232 Nitzsche, D. & Cuthberston, K. (2003). Financial Engineering: Derivatives and Risk Management. New York: John Wiley & Sons Ltd: 220-270 Quinton, S. & Smallbone. T. (2006). Post Graduate Research in Business: a Critical Guide. London: Sage Publications: 24-46 Redhead, K. (199). Risk Management with Futures and Options. London: Prentice Hall Europe: 30-67 Saunders, M. et al (2003) 3rd edn. Research methods for business students. Essex: Prentice Hall: 42-57 Shapiro, A. & Titman, S (1986). An integrated approach to corporate risk management, in Joel Stern and Donald chew, eds: The revolution in corporate Finance: Basil Blackwell ltd England. Journal of Financial and Quantitative Analysis 3 (3) 23-56 Smith, C.W & Stulz, R.M (1985). The Determinants of firm Hedging Policy. Journal of financial and quantitative Analysis 20:4: 391-405 Stephens, J. J. (2001). Managing Commodity Risk. New York: John Wiley & Sons Ltd: 34-40 Stulz, R.M. (1984).Optimal hedging policies. Journal of financial and quantitative Analysis, 19 (2) 127-40

64

Created by Mbaakoh Longinu, Coventry University

Tompkins, R.G. (1994). Options Analysis. Chicago: Irwin Professional Publishing: 45 Tufano, P. (1996). Who Manages Risk? An Empirical Examination of Risk Management Practices in the Gold Mining Industry. Journal of Finance 51 (4) 1097-137 Warner, J.B. (1977). Bankruptcy Costs: Some evidence Journal of finance 32: 2; 337-347 Yin, R. (1994). Case Study Research: Design and Methods. Beverly Hills: Sage: 45-47 http://www.anglogold.com http://www.barrick.com/ http://www.digitallook.com/ http://www.edgar-online.com http://www.gold.org/ http://www.kinross.com/ http://www.ft.com

65