An Empirical Study of Derivatives Usage in the Australian Gold Mining Industry.

Tao Jin Victor Fang Monash University

Abstract
This paper investigates corporate hedging activities in the Australian gold mining industry. We find that the Australian companies in 1997 are more actively involved in gold derivative markets than their counterparts in North America in 1993. This may suggest a general risk-averse attitude among the Australian managers when comparing with the North American managers.

We also find that corporate hedging activities are associated with some of firm characteristics, however, in different patterns. In contrast to the linear relationship suggested by previous researches, we find some empirical support for the nonlinear relationship between the possibility of financial distress and hedging levels. High and low cash production costs are found to be associated with low hedging percentage, whilst medium level cash production cost is associated with high hedging percentage. This paper also tests the relationship between hedging level and cash flow since most surveys report that the primary objective of hedging is to manage cash flows. Evidence is found that hedging percentage is associated with internal cash flow constraints. However, we find no evidence to suggest the relationship between hedging activities and cash flow from external funding sources. Surprisingly, hedging is not associated with funding sources from debt market as information asymmetry theory strongly suggested.

1.0 Introduction
Todays financial world is full of uncertainty. The fluctuation of major currencies, oil price and other commodity prices are far out of management control. Facing this unpredictable turbulence, corporations are taking risk management very seriously. Using derivatives as a tool for risk management has been exploding since the last decade. Surveys show that both financial and non-financial organisations are involved in derivative market extensively. As a result of the booming derivatives markets, three industries have blossomed: an exchange industry in derivatives, an OTC industry in synthetic products, and an academic industry in derivative research1 . Scholes (1998) suggests that the fast development of derivative markets could be partly attributed to academic researches that develop risk-managing techniques. Since the breakthrough of Black-Scholes model for option pricing, extensive literatures have covered numerous practical aspects of hedging mechanics, from the computation of hedge ratios to the institutional peculiarities of individual contracts. Unfortunately, having all these theories, the prosperous derivative market still witnessed some extraordinary stories of failures: the bankruptcy of Metallgesellschaft, the loss by Orange County investing in inverse floaters, and the law suit by Proctor & Gamble against Bankers Trust, etc. The 1998 bailout of Long Term Capital Management (LTCM) further deepened the confusion on derivative usage. All these incidents may suggest some gaps between theories and current practice of corporate risk management. In reality, academics know remarkably little about corporate risk management practices. This is not entirely fault of academics. Since derivatives are usually used as off-balance sheet finance, corporations disclose only minimal details of their risk management programs. As a result, most empirical analyses have to rely on surveys and relatively coarse data. This imposes difficulties for academics to identify companies managing more risks and understand how companies engage in dynamic risk management strategies. Surveys report that the way managers use derivative differs from academic suggestion in broad aspects. Given the apparent departure between the meaning of risk management in theories and that in reality, managers have to make tough decisions regarding risk-return
1

Scholes, M. S., (1998), Derivatives in a Dynamic Environment, The American Economic Review, June, Nashville.

trade off without clear theoretical guidelines from time to time. Therefore it is worthwhile conducting researches to capture management risk behaviour pattern, in an attempt to facilitate improvement of theories for risk management. Studying what kind of firm and why they managing risk have been the interest of some recent researches. With many limitations mentioned above, little convincing empirical evidence on the extent of risk taken by companies has been obtained. This is particularly true in Australia despite derivatives and other risk management tools are widely used by Australian companies. A recent study by Tufano (1996) on hedging behaviour of 48 pubic gold companies in North America has been considered as an exception that overcomes many of the limitations in other similar researches (Stulz 1996). This paper, following the thrill of Tufano (1996), tries to describe risk management practices and test some existing theories by analysing derivative usage in gold mining industries in Australia. The remainder of this paper is divided into six sections. Section II discusses some theoretical themes of management hedging activities. Section III presents the whole research design. Hypotheses and methodologies are discussed in Section IV. In Section V, data and discussion on test results are presented. Section VI concludes the paper. Finally, section VII discusses the limitation of this research.

2.0 Theoretical Issues and Literatures

Risk management can be roughly defined as any action taken in an effort to alter risks arising from the primary lines of the business. Cumming et al (1998) identified four types of risk management activities. The first approach is diversification: a firm diverts some of its resources to different areas rather than concentrate on a single business. The second approach is the firm manages its cost pattern to mitigate the risk arising from unfavorable price movement. The third method is to reduce leverage. In that case, the firm faces less pressure from its obligation to repay interest and debt. The last approach is to alter the distribution of cash flows by using derivatives. However, according to modern finance theory, some of these risk management activities may not increase the value of a company. Efficient Market Hypothesis (EMH) and Capital Assets Pricing Model (CAPM) are the two pillars for modern finance theory. EMH implies one cannot gain abnormal profit by trading on information since the market as a whole reflects all information available with sufficient accuracy and speed. CAPM suggests that the required rate of return of an investment is usually commensurate with the underlying risk. A companys cost of capital (required rate of return) depends on the strength of its tendency to move with the broad market (systematic risk) rather than the company's overall volatility. In general, interest rate or commodity price exposure of a particular company would not increase the risk of the whole market. Shareholders can eliminate the risk of a certain exposure, such as gold price fluctuation, with an inexpensive tool -- diversified portfolios. Thus, shareholders would not reward a company for its finance activities without true improvement on productivity. Indeed, Modigliani and Miller (1958 &1961) suggest that operation decisions, motivated by maximizing the value of shareholders, are independent of financial factors. Under the MM framework, hedging activities in a highly liquid market such as heavily traded gold derivative markets would not increase a companys value. Love and Argawa (1997) find no correlation between market values of gold companies in Australia and their levels of hedging activities. This may implies no need for hedging in business.

2.1 Cost Reduction by Derivatives However, the fact that many companies continuously devote resources in derivative markets is conflicting with the "no hedging" suggestion. Academics have tried to append rational explanations to the contradiction without violating EMH. In an efficient market, real costs associated with variations of cash flow may exist for some companies, if they are not engaged in risk management activities. Thus, derivative usage, attempting to reduce cash flow variations, may add value to shareholders. Finance literatures have identified three major costs may associate with high cash flow fluctuation: (1) cost of financial distress, (2) higher expected payment to corporate stakeholders, (3) higher expected tax payment.

2.1.1

Derivatives May Reduce Cost of Financial Distress

Even a well-diversified shareholder will be concerned if cash flow variability materially raises the probability of financial distress. In gold industry, if spot gold price fell dramatically, the company may have difficulties to recover its production costs. Although the discovered gold reserves are still in the mine, the unhedged exposures are simply not worthwhile to be extracted due to the unfavourable market movement. Considering the huge exploration and development costs have occurred, the company is likely to face financial distress. Costs related to financial distress can be both direct and indirect. Direct costs include court and lawyer costs when insolvency actually occurs. Indirect costs from financial distress have adverse impact on companys profitability in the future even when bankruptcy is not the ultimate outcome of the financial distress. When a company becomes weaker financially, it is more difficult and expensive to raise funds. The higher cost of external funding may force managers to give up profitable investment. As continuous exploration is important for ongoing operation of gold companies, underinvestment due to financial distress imposes costs for investors. In this sense, derivatives, which smooth a companys cash flows, can reduce the possible shock of companys insolvency and therefore be viewed as a substitute for equity capital.

Numerous researchers have investigated empirical implications of this theory. Many speculate that derivative usage would be correlated with capital structure of a company. Serving as a substitute for equity, derivatives can mitigate the pressure from a relative high leverage. Mixed results have been found in researches. Mian (1996), Nance, Smith and Smithson (1993), Minton and Schrand (1997) report no evidence to support the relationship between the decision on derivative usage and capital structure. Sinkey and Carter (1994), Gunther and Siems (1995), Cummings, Phillips and Smith (1997) find weak evidence suggesting the correlation. Dolde (1996), Love and Argawa (1997) confirm that high leverage firms are more likely to use derivative.

2.1.2

Derivatives May Reduce Payments to Stakeholders

Non-investor groups such as managers, employees have large stakes in the success of a company. These groups are unable to diversify exposures to their companys financial distress. As a result, they would require additional compensation for greater risk. Stulz (1984) argues that derivative usage is an outgrowth of risk aversion of managers. The empirical research by Tufano (1996) suggests that derivative activities in the gold mining industry of North American are largely associated with management compensation scheme and managerial characteristics. Managers having more options manage less risk while those with more shares manage more. Breeden & Viswanathan (1990) and DeMarzo & Duffie (1992) both argue that managers may undertake hedge in an attempt to influence the labour markets perception. Employees will demand higher wages or reduce their loyalty and work efforts if the company is perceived in financial distress and therefore add costs to the company. By using derivatives to reduce the possibility of financial distress, company can enter into favourable contracts with these stakeholders and therefore increase the value the company. In this sense, derivatives may reduce agency cost for the company.

2.1.3

Derivatives May Reduce Tax

Smith and Stulz (1985) argue that if taxes are a convex function of earnings, it will be worthwhile to hedge. Given the time value of money, hedging activities may enable companies to exploit its tax deduction when income is low or negative. A more volatile earnings stream would lead to higher expect tax than a less volatile earnings stream.

2.2 Objective of Using Derivative Theoretically, the primary goal of derivatives is to reduce the variability of corporate cash flows to maximize market value of the company. Risk, arising from uncertainty of the future, is usually quantified as variances from the expected value. The upside and downside variances appear to be indifference in these theories. Theoretical recommendation for designing a good hedge too often focus exclusively on reducing variance in the total value of the firms projects. The actual corporate use of derivatives, however, does not seem to correspond closely to these theories. In Tufano (1996), gold mining companies in North American have adopted different risk management levels. While some companies do not hedge at all, some others have very high percentages of derivative sales in their production volumes. The wide range of hedging levels imposes some doubt on theoretical definition of risk management as minimizing variance when companies are in the same environment. Surveys provide further evidence on this doubt. Many corporations engaged in risk management seem to hedge downside exposures while pursuing opportunities to take advantage of upside volatility. Wharton school of Pennsylvania conducted a survey2 on derivative usage by US non-financial firms in 1995 with about 350 firms responded. The survey finds that a large percentage of firms sometimes take account of market outlooks before choosing a strategy. About 12% of those firms report frequently altering timing or size of hedges. Further 61% of those firms report sometime altering the timing of their hedges and 48% sometimes altering the size of their hedges. When asked to indicate the frequency of taking a position in derivative usage basing on their market views, 6% of those firms report frequently, about 33% respond sometime. Dolde (1993) finds similar result. Almost 90% of firm in his survey sometimes hold a view when choosing hedging strategies and a proportion of firms use selective hedging rather than fully hedge. These results show that corporations do not systematically hedge their exposures. Dolde (1993) suggests that most firms in the survey employed a range of blended strategies to trade off cash flow stabilisation provided by hedging against speculative gains 3 . Stulz (1996) summarises a more realistic rationale for corporate risk management behaviour:

Bonday, G. M., G. S. Hayt & Richard, C., (1996), 1995 Wharton Survey of Derivatives Usage by US Non-financial Firms, Financial Management, Tampa. 3 Dolde, W., (1993), The Trajectory of Corporate Financial Risk Management, Journal of Applied Corporate Finace, Fall, Pp33-41.

Managements view of future price movements was an important determinant of how or whether risk was managed. Risk management did not mean minimizing risk by putting on a minimum-variance hedge. Rather, it meant choosing to bear certain risks based on number of different considerations, including the belief that a particular position would allow the firm to earn abnormal return (p11). Because of the choice on bearing certain risk in derivative usage, one may have doubt whether a company involved in derivative markets is a real hedger or a speculator. These speculations, by changing hedge level according to market view over times, simply show management derivative activities are apart from believing in efficient market hypothesis. Two reasons may contribute to the departure: information asymmetry and risk attitude.

2.3 Information asymmetry Contrary to MM framework, some argue that there is a link between a companys financial structure and its investment decision. This is because internal finance is not a perfect substitute for external finance. While internal finance refers to cash flows generated from business operation, external finance includes borrowing from outsiders and issuing new equity. Fazzari, Hutbbard & Petersen (1988) argue that a financing hierarchy exists for the three types of financing due to information asymmetry in markets. Asymmetric information can lead to cost disadvantages for external finance. This argument draws from the lemon problem first considered by George Akerlof 4 . The lemon refers to a good quality company that fails to be distinguished from normal companies because of information asymmetry. They found empirical evidence to support their argument. When information is asymmetric in share market, normal investors without inside information are in a disadvantage position discerning the quality of companies. They will value all similar companies at a population average. Consequently, investors without inside information implicitly demand a premium for shares of a relative good company. For those investors with inside information, they are unwilling to accept the price offered by a less informed buyer. The premiums required by less-informed investors add cost for high-quality companies.
4

Akerlog, G.A., (1970), The Market for Lemons: Quality and Certainty and the Market Mechanism, Quarterly Journal of Economics, Vol. 84, August, Pp488-500.

The same problem is also applicable in debt markets, however to a less extent. As Calomiris and Hubbard (1988) point out, creditors in a full-information market (bond or commercial paper) coexist with those in bank loan market. Since the first type market is usually accessible within big size borrower and big size creditors, inside information could be easily communicated. In the bank loan market, although borrowers are smaller in size, banks specialize in monitoring financing projects. Their expertise helps to overcome the problem of information asymmetry to some extent. Further, as corporate managers can be assumed to have full information about the value of the firms existing assets and new investment project, internal finance does not have the lemon problem. As a result, while the internal finance is cheapest for investment, debt markets offer cheaper finance than equity markets. In addition to risk-return trade-off in light with CAPM, information asymmetry also contribute to the financing hierarchy or the so-called pecking order of finance. Being aware of the financing hierarchy, companies tend to have preferences on different funding strategies. While internal finance should always be in priority, choice between the two types of external finance depends on the level of information asymmetry for the company. For companies that are likely to have the lemon problem, the gap between required rate of return from a normal investor and borrowing rate would be bigger than the gap of a normal quality company. This is because information asymmetry is less significant for creditors than normal shareholders. Thus, "lemon" companies should always consider using debt finance prior to equity. On the other hand, creditors are usually concerned about agency cost associated with debt, particularly for long-term debt. According to agency theory, managers may forgo profitable investment and accept other projects with negative present value for the benefit of shareholders; they also have incentives to lower the value of existing debt by offering repayment priority for new debt. The greater the debt-equity ratio, the more likely is this agency problem. Creditors are aware of this problem and tend to impose debt covenants that restrict the behaviour of managers. However, these restrictions on financial flexibility may limit managements choice on investment. As a result, a company may use equity market if restrictions by debt covenant impose higher costs especially when it has borrowed up to full capacity. This is more likely for a normal company without lemon problem.

2.4 Speculation and Hedging Level Strange(1986) in her book, Casino Capital, suggests that many people works cheaper for risk with the hope to beat market. In real business world, managers are not totally risk averse as assumed by many finance theories. To honor their fiduciary duty, managers are expected to achieve higher return for investors and also maintain solvency of the company. This may explain why many companies adopt the blended strategy combining speculation and hedging at the same time. Stulz (1996) suggests that companies should manage risk in a way that makes financial distress highly unlikely, in so doing they preserve financing flexibility necessary to carry out their investment strategies. A companys ability to withstand a certain level of trading loss from speculative depends on its general financial health and capital structure. Therefore, a company that is very unlikely to fall into financial distress can take bet on managers market view and choose a low level hedging strategy. If a bet turns out badly, the loss will not affect the companys ability to carry out its business. For companies already in financial distress, reducing risk is not in the interest of stakeholders. In this circumstance, a manager intends to maximize firm value by taking speculations for abnormal profit to rescue the company from the distress. These companies will also choose a low level hedging strategy. In contrast, companies somewhere in between, unfavourable cash flow variations may push them into financial distress. Consequently, these companies should take hedging more seriously comparing with companies in the extreme positions.

10

3.0 Research Design

Many researches have looked at different aspects in attempting to explain risk management activity. This research is designed to test some of these theories using data collected from Australian gold industry. Tufano (1996) pointed out that gold mining industry is almost tailor-made for academic investigation of risk management: these companies tend to be single-industry firms sharing a common and clear exposure to fluctuating gold prices; gold prices have experienced substantial volatility; capital market have developed a wide variety of techniques to manage this volatility; firms have adopted a rich variety of policies with respect to gold price management. Traditionally, mining companies are intensively involved in derivative market. Most importantly, firms in the gold mining industry disclose their risk management activities in detail in their annual reports. This enables meaningful measurements to be calculated on the degree of risk management undertaken.

3.1 Sample Construction Listed companies in Australian Stock Exchange classified as gold mining industry form the sample of this research. 1997 annual reports of these companies are used. This is because, in 1997, Australian Accounting Standard Board (AASB) published a new disclosure standard for financial instruments. This standard requires detailed information to be disclosed and significantly improves comparability among companies. All annual reports used in this research are from the CD-ROM published by Connect4. 64 annual reports are available. Information related to derivative usage is collected from these annual reports. Among those 64 firms, 20 are mainly in exploration stage in 1997 financial period and no production was carried out. There were 4 companies suspended its gold production due to the falling gold price that makes their reserves not economically viable any more. This means 24 companies did not have proved or probable reserve economically viable at the end of the reporting period. None of these companies reported derivative usage in its 1997 annual report. This left 40 companies that were engaging in gold production as at the end of the reporting period. 5 of the 40 companies reported derivative usage, however, with no quantitative disclosure. At last, 35 companies make up the database in this research. Attempts are also made to have 1998

11

reports. Till the end of March 1999, only 28 companies have published their annual reports. Among these 28 companies, 18 have both reserve and disclosure of derivatives. Because of the small sample size of 1998, it is unable to use them as separate year data. All the companies are listed in appendix I.

3.2 Measurement for the Extent of Risk Management Activity The rich menu of derivative instruments enables gold companies to customise their gold price exposure. Similar to Tufano (1996), this article primarily focuses on the level of risk management activity chosen by the companies. He points out that in order to measure a mining companys risk management on gold price, one would like to know the economic magnitude of risk modification activities across all types of transactions, scaled by the size of the underlying exposure of the company. He defined the risk management level as delta-percentage. The percentage equals to total equivalent ounces of gold that the firm has effectively sold short through derivatives within the research period divided by the amount of gold expected to be produced over the same period. One limitation of this measure is that gold sales through derivatives beyond three years are not considered. Because of different information available in annual reports, this paper will use a slightly different measurement. In the annual reports, many companies disclose all their outstanding gold derivative contracts at the balance sheet date. Thus, risk management can be measured by using total equivalent volumes of gold sold short in outstanding derivative contracts divided by total reserves as at the balance sheet date. This deltapercentage represents the percentage of future production (proved and probable reserve) known as at the balance sheet date that has been sold forward. Table I shows the calculation. Delta is the probability that the hedge contract will be exercised. For forward and futures contracts, the probability equals to 1 as both parties are bound to honour the contracts. For options, as option buyer has the choice to exercise the contract depending on market movement, the possibility is measured by N(d2) from Black-Scholes option pricing model for call options and N(-d2 ) for put options. Here all options are assumed as European options.

12

Table I
Sample Data on One Gold Mining Firms Risk Management Activities

Panel A: data from the annual report of DGD1997 Time Forward: < 1 year 1 ~ 2 year 2 ~ 3 ear 3 ~ 4 year 4 ~ 5 year Put option purchased: < 1 year Call option sold: 1 ~ 2 year 3 ~ 4 year 4 ~ 5 year Ounces Price /ounce 649 A$ 673 A$ 697 A$ 702 A$ 828 A$ 441A$ 410 US$ 533 A$ 550 A$

58,000 30,000 51,000 50,000 23,000 72,000 9,000 12,000 48,000

Panel B: Delta of the firms gold derivative portfolio Position Ounces Forward 212,000 Put option 72,000 Call option (2.5 year) 9,000 Call option (3.5 year) 12,000 Call option (4.5 year) 48,000 Total ounces 353000 Aggregate equivalent ounces (a) Total reserve (b) Delta percentage (a/b) Delta 1 0.095989 0.248452 0.599445 0.628507 Delta-ounce 212,000 6,911.214 2,236.064 7,193.339 30,168.34 258,509 2.05 million ounces 0.125722

In our paper, the spot gold price was A$465 or US$330 per ounce, which was taken as at the end of June in 1997. Risk free rate was 5.4 percent using the Australian government bond rate. Volatility of gold price at 0.084757 is the implied volatility by using gold price fluctuation in 1997.Gold lease rate is ignored here for two reasons: (1) gold lease rate is very low. Historically, it is around 1 percent per annum, which is far below cash market interest rate5 . (2) Storage costs is needed for physical gold that can be regarded as negative incomes that may offset the gold lease income.
5

See Love & Argawa (1997) for the reason why gold lease rate is consistently lower than other market interest rate.

13

For example, the delta-ounce of a call option with the exercise price at A$533 and maturity period of 3.5 years in GDG (table I) is calculated using Black and Scholes(1973) option model ( see Appendix II):

d2 =

ln( 465 / 533) + (0.054 0.084757 2 / 2) 3.5 0.0845757 3.5

=0.251911

Delta = N(d2 ) =N(0.251911) =0.599445 Delta-ounce = 12,000 0.599445 = 7193.339 ounces

3.3 Data Description 3.3.1 Risk Management of Gold Mining Companies in Australia 1997 was regarded as a bad year for many gold mining companies. Gold price dropped continually during the year. Supplies of gold increased partly because of the announcement that Australian Reserve Bank would sell a significant amount of gold reserves. A soft market demand for gold was predicted due to the Asian financial crisis. At the same time, a scandal involved in a huge gold project in Indonesia made investments in gold industry not a compelling decision at that moment. Similar to those companies in Tufano (1996), gold mining companies in Australia are generally not well diversified. In our samples, one company diversified into construction industry. Some of the companies also produce small amount of other metal complement to gold mining. Thus, possible risk management activities usually fall into last three methods classified by Cumming et al (1997). The simple structure of these gold mining companies improves comparability of derivative usages between them.

14

Cash production cost is a critical measure on whether gold reserve is economically viable. On one hand, it is determined by the geological nature of the mine. Companies with mines easy to extract (open pit or short digging distance) have cost advantage over others. These geological characteristics of mines are difficult to be manipulated. On the other hand, companies can improve its extraction technique to reduce the site costs. In annual report, many companies reported efforts to reduce cash production cost to improve profitability and reduce the risk of financial distress. Derivative has been the fastest growing risk management method in the recent two decades. Tufano (1996) differentiate two types of derivative activity: hedging and insurance. Hedging sheds all exposure through the sale of gold at fixed prices and therefore eliminates any future variance of cash flow. Hedging instruments include overthe-counter forward sales contract, exchange-traded future contracts, gold or bullion loans, gold swaps and spot deferred contracts. Insurance protects firms against downside exposure only. For gold mining companies, insurance instruments usually include purchased put options. Besides these straightforward instruments, some companies use complex instruments including: Convertible put options, convertible forwards, floating forwards, knock-in and knock out options 6 . Some companies used purchased call options and write options to adjust risk management level.

Floating forward: if spot price is higher or lower than a certain level during the contract period, the forward will be cancelled. Convertible puts are put options that become forward sales if gold reaches defined gold prices. Variable priced forwards have a realisable value that is dependent upon the spot price at maturity. Knock-out put options represent put options that are extinguished if spot gold prices reach defined prices. Convertible forwards are forward sales that convert to put options if gold falls to defined low prices, or trigger additional call options if gold reaches defined high prices.

15

3.3.2

Hedging Level

Table II reports the level of risk management activity entered by the sample firms in the financial year of 1997.

Table II
Distribution of Risk Management Activity in the Australian Gold Mining Industry Firm Hedging level Exactly 0 0.1 ~10 10 ~ 20 20 ~ 30 30 ~ 40 40~ 50 50 ~ 60 60 ~ 70 70 ~80 80 ~ 90 90 ~ 100 > 100 Total Mean Median Standard deviation Minimum Maximum No of Firms 1 4 5 6 6 2 2 3 2 1 2 1 35 Industry reserve 0.01 17.13 46.79 5.28 10.78 9.29 1.21 4.66 1.1 0.09 3.57 0.09 100 Hedging Percentage 2.86 11.43 14.29 17.14 17.14 5.71 5.71 8.58 5.71 2.86 5.71 2.86 100 38.86 33 28.21 0 101 25.6 22.9 22.4 0 85.9 Percentage in Tufano* 14.6 14.6 14.6 14.6 25.0 2.1 4.2 4.2 4.2 2.1 -

* Tufano (1996) Comparing with the hedging ratios in Tufano (1996), it is clear that Australian gold mining companies are more actively involved in gold derivative markets in 1997 than their competitors in North America did in 1993. Australian companies have higher value on both mean and median of hedging level. While American companies deploy higher percentage in the low hedging level groups (percentage is less than 40), Australian companies consistently have higher percentage in the high hedging level groups. Further, Tufano (1996) only takes account of the three-year derivative activities scaled by the three-year production. The hedging level in this research considers all the hedging activities of the company scaled by reserves, which is the total future productions known at present. As companies are more concerned about market movement in the near future,

16

Tufanos delta-percentage should be higher than the percentage calculated in this research for the same company. That is, if the same measurement is employed, the gap of hedging level between American and Australian companies would be even more significant. Overall, American companies appear to be less risk averse than Australian companies. This result is similar to the conclusion in a survey conducted in New Zealand that across all firm sizes relatively more firms use derivatives. Nevertheless, most Australian companies still fall into the group of companies with low hedging level. Within the 35 companies, 22 (62.86%) have hedging levels less than 40 percent of their reserves. In terms of industry total reserves, companies having low hedging level (<40%) hold around 80% of total industry reserves. Those companies having hedging levels more than 40% of their reserve only hold 20% of the total industry reserve. Thus, even in a year that the whole industry has a bleak outlook, companies did not rush towards fully hedging.

17

4.0 Hypotheses and Methodology

4.1 Hypothesis I The motivation of this research originates from Tufano (1996). In his prize-winning paper, Tufano tests derivative usage in North American gold industry. Two classes of hypotheses are examined for management choices of hedging level: Shareholder Value Maximization Hypothesis and Managerial Utility Maximization Hypothesis. The first hypothesis mainly draws from the argument that derivatives may reduce the likelihood of costly financial distress as discussed in 2.1.1. Financial distress in gold mining industry is indicated by leverage, cash production cost. Cash position and company size are included in the test as part of firm characteristics. The second hypothesis is primarily based on agency theory. It is argued that managers whose personal capitals are poorly diversified would prefer managing risk for the best interests of themselves instead of shareholders. Managers with greater share ownership would prefer higher hedging level, whilst managers with greater share options would employ less derivative usage. This is because shares provide linear payoffs as a function of stock prices whereas options provide convex payoffs. To some extent, this hypothesis conforms to the academic argument discussed in 2.1.2 that derivatives reduce agency cost between management and shareholders. Tufano (1996) reject the Shareholder Maximization Hypothesis 7 and uphold the Management Utility Maximization Hypothesis. In order to compare the derivative usage by gold mining industry in Australian and North America, we establish the following hypotheses. H1a H1b The hedge level of a company is positively correlated with both cash production cost and leverage as indicators for financial distress. Risk averse management, having greater share holding, would be associated with higher hedging level, whilst having greater option holdings, would be associated with lower hedging level. Both univariate and multivariate methods are used to test these hypotheses. Since only one company of the 35 companies in this research did not hedge in 1997 financial year, samples
7

The evidence shows cash production cost is not a significant determinant for hedging level. The relationship between leverage and hedging is subtle in the result. When using the pooled sample (three years together), there is a statistically significant positive relationship between leverage and hedging level. Single year data are also analysed in the research. However, none of the p-values for the year by year analyses are significant to reject the null hypothesis. Tufano (1996) concludes that the hypothesis is not well supported.

18

are divided into two groups for the univariate test: (1) companies with small hedging level (<40%) (2) companies with large hedging level (>=40%). The cut-off point at 40% is consistent with Tufano (1996). For small sample size, Webster (1995) suggests a nonparametric analysis, Mann-Whitney U test, be used to test the mean difference between the two groups. For the multivariate test, ordinary least square regression is conducted in the multivariate analysis. The regression model is outlined as follows:

Hedratio = + 1 cash cos t + 2 leverage + 3 cashpos + 4 size + Hedratio = + 5 share + 6 option +

Our statistic technique here differs from the Tobit regression in Tufano (1996) mainly because many companies in his sample did not hedge. Tobit regression could avoid a bias towards large amount of zero value in his samples. Since gold companies with no reserve are excluded from our sample, we have only one company with no hedging in the sample. Furthermore, a single year data in the research effectively avoid the violation of assumptions for least square regression by time series data. Therefore, it is appropriate to use least square regression to test the hypothesis here. Table II shows the hypothetical relationship between hedging ratio and the variables.

Table III Hypothesis

Shareholder value maximization

Variable
Cash costs

Sign
+

Data Description
Cash costs include all direct and indirect costs of mining, crushing, processing and general and administrative expenses of the mine. Cash costs exclude noncash items, such as depreciation, depletion and amortization as well as interest expense, corporate SG & A, exploration and extraordinary costs. Total liability scaled by total assets Log of the value of common shares owned (excluding options) at average annual share price Number of options outstanding held by directors

Leverage Managerial utility maximiztion Directors shareholding Directors option holding Control factor Firm size Cash Position

+ +

Annual average of weekly market value of equity The percentage of cash in terms of total assets

19

4.2 Hypothesis II As per section 2.3 above, internal finance is the cheapest fund source. Managers would consider using internal funds as the first priority for investment. Sometimes, they may adapt investment level to the concurrent cash flow situation. When market moves to the detriment of the underlying business, internal cash flow will be constrained for existing business while required rate of return in both equity market and debt market would increase. To overcome the financial constraints on cash position, managers tend to limit their investments. However, this is less likely for companies with urgent investment commitment. In gold mining industry, the continuous operation depends on economically viable reserve. Continuous exploration and development for new reserves are important for the companys future survival. Restrained internal cash flow due to unfavourable market movement could cause managers difficulties in meeting with the investment commitment. Generally gold companies are sensitive to cash flow variation. This will be more significant for companies with lower existing reserves in terms of annual production capacity. As a result, these companies are more concerned about internal cash flows stability through hedging activity. H2 Companies with short production life remained tend to hedge more than those with long production life remained. The production life remained equals to the total prove and probable reserves scaled by the annual production in 1997. It is assumed in 1997, the company operated at full capacity. Companies started production in 1997 are excluded from the sample. An ordinary least square regression model is used to test this hypothesis.

hedratio = + ( reserve production ) +

The Mann-Whitney U test is also employed to test the mean difference of production life remained between the two groups of companies with high versus low hedging level.

20

4.3 Hypothesis III The sophistication of derivative markets makes derivative activities fall well in the area of information asymmetry as discussed in section 2.3. Financing hierarchy caused by information asymmetry is also applicable to derivatives. Companies with successful hedging policy may have difficulties in communicating the policy to normal investors in equity market. Lack of understanding in derivatives, shareholders could not differentiate companies with different hedging levels. By contrast, banks are in a better position to understand derivatives since many banks are actively involved in derivative markets. For risk-averse bankers, high hedging level could reduce risk in granting loans to a risky company. This is very true in a risky industry such as gold mining. A bleak outlook of the whole industry in 1997 can further aggregate the information asymmetry. The asymmetric information would widen the gap between required returns between bankers and shareholders for companies with successful derivative hedges. To avoid disadvantage of the information asymmetry, these companies should try to use debt market as much as possible. They may further increase hedging level for more borrowing instead of issuing shares. By contrast, a company has low hedging level may not differentiate between debt and equity market as much as those companies having high hedging level. Therefore, it is reasonable to assume that correlation exists between hedging levels and funding strategies. H3 If companies have to rely on external finance, companies using debt market tend to have higher hedging level than those purely using equity market. The sample is divided into three groups according to the cash flow statement of 1997: (1) companies only use internal funds for investment: these companies have a negative cash flow from finance. (2) Companies use both internal and external finance, however, only debt market is used. (3) Companies use both internal and external finance and new shares are issued in 1997. New shares from employee share scheme are not considered as external funding source. As we use dummy variables to represent each group, a univariate test is employed to examine this hypothesis. Considering the small sample in each group, non-parametric test is used. According to Webster (1995), the Kruskal-Wallis test applies for k-independent groups (k > 2), which is equivalent to ANOVA in parametric test.

21

4.4 Hypothesis IV As discussed in section 2.4, Stulz (1996) suggests that company should adjust derivative usage according to its financial position. Companies in financial distress or in healthy position would exhibit lower hedging level. Only those companies that may face financial distress would hedge more. This implies that financial distress does not necessarily have a linear relationship with hedging level and provides a plausible explanation why Tufano (1996) rejects the linear relationship between firm characteristics and hedging level. Tufano (1996) divided the sample into three groups according to its hedging level. According to Stulzs theory, companies in financial distress and those financially healthy could have similar hedging level. When these two types of companies are in the same group, the mean value indicating financial position would be evened up. In order to overcome this bias, we group our sample according to firm characteristics to test differences of the mean hedging level of each group. Stulz (1996) suggests firm leverage as the indicator for the possibilities of financial distress. Considering the specific industry characteristics, cash cost is also employed as an indicator for financial distress in this research. H4a H4b For companies with high or low leverage, they tend to hedge less. For companies with medium leverage, they tend to hedge more. For companies with high or low cash production cost; they tend to hedge less. For companies with medium cash production cost, they tend to hedge more. To test these hypotheses, the sample is divided into three groups according to firm characteristics rather than hedging level. All the companies are ranked in terms of cash production cost and leverage separately and then stratified in quartiles. The first quartile of each ranking is grouped as companies with strong financial position. The last quartile of each rank is grouped as companies with high possibility in distress. The remained companies are in the medium position. A Kruskal-Wallis test is employed for this hypothesis.

22

5.0 Testing Results and Discussion 5.1 Hypothesis I Table IV reports results for the multivariate regression. Except directors option holding, all other factors display the same sign on as hypothesised. However, only leverage passes the significant level less than 10 percent. As the explanatory power for the two regression model are fairly low, it only provides weak evidence suggesting leverage is positively correlated with managers choice on hedging level. The univariate test examines whether the mean ranks of firm characteristics are different between the high hedging level group and the low hedging level group. The results in Table V are consistent with those in the multivariate regression model. While the low hedging group has a lower mean rank in cash costs and leverage, it has a higher mean rank in directors shareholding and option holding. Except the directors option holding, differences displayed by the other three factors are in the same sign as hypothesised. However, none of them has shown significant differences. Even when possible correlations among independent variables are omitted, it is unable to conclude that the two groups are different. Thus, both hypotheses 1a and 1b are rejected. Our results differ from those of Tufano (1996) with respect to the Managerial Utility Maximisation Hypothesis. He found evidence that managers share and option holding are associated with the degree of risk management. Our results show no evidence to support the association. This difference could be due to cross ownership between Australian gold companies. The cross holding enable directors in one company to sit in boards of other companies although they do not have direct interest (share or option holding) in other companies. Indirect ownership is usually not disclosed in annual report. Thus, directors interests disclosed in annual reports may not reflect their real interests involved in the company, hence, distorts our results. Our results also differ from those in Love & Argawa (1997). They concluded that cash cost and leverage are the determinants of hedging by using data of 17 Australian gold mining companies between 1992 to 1996. As shown in Tufano (1996), pooled data and single year data could have different result because of time-series. Furthermore, the regression model in Love & Argawa (1997) has an adjusted R square at 0.10263. The relatively low explanatory power shows factors other than cash cost and leverage could significantly affect the hedging level.

23

Table IV: Multivariate test results for H1 Panel A: Ind. Variables 1 Intercept Cash cost Cashpos Leverage Size

Coefficients 0.437 0.010 0.231 0.331 -0.130

t-Stat. 0.666 0.054 1.199 1.733 -0.730

Sig. 0.957 0.240 0.094* 0.472

Dependent Variable: HEDRATIO Diagnostic R squared Adjusted R square F-statistics

0.121 -0.001 0.994

Panel B : 1 Ind. Variables Intercept Diroptions log(dirshval) Coefficients 0.461 -0.158 -0.200 t-Stat 1.786 -0.684 -0.862 Sig. 0.089* 0.501 0.398

Dependent Variable: HEDRATIO

Diagnostics R square Adjusted R square F-statistics *Significant at 10% level.

0.093 0.007 1.075

24

Table V: Univariate test result for H1 Low hedging level (<40%) No. Hedratio Cash costs Leverage Director shareholding Director option holding 21 21 22 20 17 Mean 0.2072 346.81 0.3521 14.9189 2283587 Std. Dev. 0.1163 115.62 0.1936 2.2096 3596345 Mean rank 11.5 16.3 17.14 16.5 14.03 High hedging level (>=40%) No. 13 13 13 10 10 Mean 0.6956 378 0.3962 13.682 1231231 Std. Dev. 0.1983 66 0.1953 3.084 1083554 Mean rank 29 19.42 19.36 13.5 13.95 U Sig

111.5 124 80 84.5

0.381 0.533 0.397 0.98

25

5.2 Hypothesis II The Mann-Whitney U test in Table VI shows there is a significant difference between company with high hedging level and those with low hedging level in respect to production life remained. In consistent with H2, companies with shorter production life remained are more inclined to hedge. Companies with low hedging level have an average rank of production life at 19.95 comparing with 12.46 for companies with high hedging level. The regression model provides a negative of -0.25 (Table VII), which means one more production year would lead to a reduction of hedging level by 0.25. However, the relation is only significant at 15% level and the explanatory power of the regression model is fairly low at 3.4%. The lower explanation power suggests that production life remained may not have a linear relationship with hedging level. There could be a competing explanation for the different hedging level in the U test. Both Wharton survey and Dolde survey find that the focus of risk management is on near-term exposures. Further, market is more liquid for short-term derivative instruments. It is reasonable to assume that most hedging arrangements are for productions in the near future although some companies have long term hedging contracts8 . If companies have similar hedging level for productions in recent few years, the hedging ratio, which is the total forward selling by outstanding derivative contracts divided by total reserve, could be greater for companies with short production life remained. This is because companies with short production life have less reserve in terms of annual production. The higher hedging level could be simply caused by the measurement employed in the research.9 In order to test the competing explanation, another measurement of hedging level is also calculated. The new hedging percentage, scaled by annual production, represents how many times of annual production have been effectively sold forward. Results in Table VI show that companies with lower hedging level have less times of annual production that have been sold forward, using the new measurement -- hedging ratio2. While companies with low hedging level have a forward selling average at 2.22 times of annual production, companies with high hedging level have more forward selling average at 3.7 times of
8

In our sample, hedging contracts cover broad range of period. One company has a hedging contact with 12 years of maturity period. 9 For example, both company A and B has same hedging levels for productions in recent years: 50% for the first year, 30% for the second year and 0 for the third year. And they have the same annual production. However, the reserve for company A is only enough for production in next two years, whilst, company B has reserve for next ten years production. The hedging ratio would be smaller for B as the denominator would be greater when the two have the same nominator.

26

annual production. Therefore, companies with high hedging level have large volumes of forward selling in terms of both total reserve and annual production. This effectively dismiss the competing explanation that the high hedging level exposed is due to the measurement employed in this research. Thus hypothesis II is supported. Companies that are more sensitive to internal cash flow fluctuation hedge more. This result is consistent with survey findings that managing cash flows is the top choice for derivatives (Wharton 1996).

27

Table VI: Mann-Whitney U Test Low hedging level (<40%) No. Hedratio Hedging ratio2a Production year 21 21 20 Mean 0.2072 2.2241 14.1836 Std. Dev. 0.1163 1.5075 14.083 Mean rank 11.5 13.80 19.95 High hedging level (>=40%) No. 13 13 13 Mean 0.6956 3.6967 6.016 Std. Dev. 0.1983 1.7965 4.2153 Mean rank 29 21.92 12.46 U Sig

66 71

0.018** 0.03**

** Significant at 5% level a: hedging ratio2 refers to hedging level scaled by annual production Note: One company starting production in the middle of 1997 financial year was excluded from the test since the annual production in 1997 does not represent full operation capacity of the company.

Table VII: Regression test for H2 Ind. Variables Intercept Production year Dependent Variable: HEDRATIO Coefficients -0.252 t-Stat 7.835 -1.473 Sig. 0.000*** 0.151

Diagnostics R square Adjusted R square F-statistics *** Significant at 1% level.

0.063 0.034 2.169

28

5.3 Hypothesis III Results, using Kruskal-Wallis test, show that hedging levels for the three types of funding strategy are not different from each other (Table VIII Panel A). The mean hedging levels and the mean ranks for the three groups are fairly close. All three groups have broad range of hedging ratio since standard deviations are large for each group. Further, the mean hedging level for companies used equity market in 1997 is slightly higher than that for those only use debt market. This suggests management did not deliberately increase hedging level to gain further access to debt market. Therefore, hypothesis III is not supported.

Nevertheless, one cannot understand a companys funding strategy without taking consideration of the specific situation of the whole company. Other risk management activities could also affect a companys funding strategy. For example, a company with high leverage is more likely to issue new shares as creditors usually take gearing ratio as a major indicator for insolvency. In an attempt to distinguish other factors may significantly affect a companys funding strategy, indicators for other risk management activities such as cash cost and leverage are also tested for the three groups. However, leverage could be a determinant for the decision and also changed by the decision. While new borrowing would increase leverage level, new shares issued would reduce leverage. For this reason, leverage in the previous financial year may better served as a determinant for the funding strategy than current leverage. Moreover, market perception on a companys growth opportunity may indicate its capacity to raise external funds. If information asymmetry does exists in market, debt markets will be better informed than equity market. Within the same industry, a company with less growth opportunity tends to rely on equity market. The ratio, weekly average market value to net equity of a company, is used as a proxy for growth opportunity in the test. Table VIII Panel B reports test results for factors other than hedging level. Companies relying on debt market statistically have higher leverage in 1997 than the other groups. Except for 1997 leverage, none of the factors shows significance to determine funding strategies. However, as the 1996 leverages for companies using debt market (group 2) and equity market (group 3 in table VIII Panel B) are generally indifferent, the higher mean leverage in 1997 for companies resorted to debt market can be attributed to the borrowing activities during 1997 period.

29

Table VIII: Results for H3 Panel A: HEDRATIO Cash flow 1a 2b 3c Firms 13 7 12 32 Mean . 3808 . 3733 . 3786 Std. dev. . 2676 . 3093 . 2692 Mean Rank 16.77 15.71 16.67 Chi-square df. Sig.

Total Panel B: Cash cost

0.064

0.969

1 2 3

Total Leverage 1 2 3

13 7 11 31 13 7 12 32 13 7 12 32

332.31 336 392.72

101.57 106.1 99.69

14.19 14.36 19.18 2.090 2 0.352

. 3632 . 5215 . 3262

. 1662 . 1522 . 1990

15.46 23.71 13.42 5.596 2 0.061*

Total Leverage96 1 2 3

. 3609 . 2902 . 3001

.2531 .1239 .1770

18.46 15.71 14.83 0.996 2 0.608

Total Growth opportunity

13 5.0058 5.6741 17.92 7 3.6985 2.4497 18.71 12 2.4787 1.5764 13.67 Total 32 1.784 2 a: refers to companies only use internal fund. b: refers to companies also use external funds from borrowing. c: refers to companies use external funds but not from borrowing. Note: The two companies with positive cash flow from investment (no investment during the year) were excluded in this test.

1 2 3

0.41

30

Although not statistically significant, both cash costs and growth opportunity display a better mean value and rank for companies that resorted to debt market comparing with those used equity market. Companies using internal finance and borrowing from debt markets have lower mean cash production costs at about A$336 comparing to A$392 for companies using equity market. Similar, companies using internal finance and debt market exhibits higher ranking in growth opportunity than companies using equity market (Table VIII Panel B). It seems that information asymmetry between debt market and equity market still exists to some extent. However, for an apparent area causing information asymmetry like derivatives, we found no evidence at all. One plausible reason is the small sample sizes in the research. Another possible explanation could be that managements do not link derivatives with external financing.

5.4 Hypothesis IV Hypothesis IV tests hedging activities in a non-linear pattern. Results are reported in Table IX. Contradicting to the suggestion in Stulz (1996), there is no evidence supporting the hypothesis with respect to leverage. Actually, the results, to some extent, support that leverage is positively correlated with hedging ratio. Companies with lower hedging ratios consistently embrace lower leverage among the three groups. This positive relationship is consistent with Tufano (1996) when using the pooled data and Love & Argava (1997). As predicted, companies with high and low cash costs hedge less than those with medium cash costs. Companies with medium cash production costs in the industry have considerably high mean value and high rank hedging ratios when comparing with companies with low or high cash production costs. The difference is significant at 10% level. The scatter plots clearly depict the relationship.
Figure 2
cashcost 600

Figure 3
Leverage 1

500

0.8
400

0.6
300

0.4
200

0.2
100

0
0 0 0.5 1 1.5

0.5

1.5

31

Figure 2 shows the relationship between hedging level and cash production costs. Although companies with medium cash production cost have broad range of hedging level, companies with high hedging level (50%) are exclusive companies with medium cash costs. Figure 3 exhibit the relationship between hedging level and leverage. The plot virtually shows no association between the two variables. According to Stulz (1996), companies could bear a certain level of risks to take advantage of possible profit windfall by speculating in market, providing this will not increase the likelihood of financial distress. Our result suggests that managers are using cash production cost rather than leverage to monitor the likelihood of financial distress when deciding hedging strategy. One possible reason relates to the purpose of derivative usage. As derivatives are used to manage cash flows, variable costs act as a direct standard for hedging contracts. Companies with high cash costs but still below spot price are likely to fall into financial distress if market price moves further unfavourably. These companies are more inclined to fix selling price above costs to a certain extent as suggested by Stulz. Furthermore, the difference between cash production costs and gold market price serves as an indicator for future production. Tufano (1997) points out gold mining companies hold a call option of gold, with the exercise price being their marginal production costs (cash production costs). For example, when the gold price falls below marginal costs, the firm can choose to temporarily or permanently suspend production. In this sense, cash production costs determine the maximum volume (future production economically viable) and minimum exercise price (marginal production cost) of hedging contracts. The 24 companies excluded from the sample may further suggest this association. Indeed, none of the 24 companies report derivative usage when they have no reserve at all. Another explanation relates to accounting standard in Australia. The standard requires mining companies write off deferred exploration and development costs from assets for reserves not economically viable any more. A large reduction in book values of a company may signal to the market that the company is in financial distress. Fixed price through derivatives could maintain reserves economically viable even when market price moves to an unfavourable direction. Thus, hedging is more compelling for companies having cash production costs close to market spot price. By contrast, leverage is not as direct as cash costs acting as a signal to market for possible financial distress, especially for equity market. In share market, some high return companies actually have high leverage.

32

Table IX: Test results for H4 Panel A Leverage Hedratio 1 2 3 Total N 9 18 8 35 Mean .3583 .3613 .4842 Std. dev. .3758 .2214 .3040 Mean Rank 15.11 17.94 21.38 1.584 2 0.453 Chi-square df. Sig.

Panel B Cash cost Hedratio 1 2 3 Total N 9 17 8 34 Mean .2435 .5047 .3044 Std. dev. .1731 .3119 .2488 Mean Rank 12.56 21.24 15.13 5.066 2 0.079 Chi-square df. Sig.

Note: Cash cost of one company in the sample was unable to be obtained. 1. The lower 25% companies according to the rank of leverage (panel A) or cash production cost (panel B) 2. The middle 50% companies according to the rank of leverage (panel A) or cash production cost (panel B) 3. The highest 25% companies according to the rank of leverage (panel A) or cash production cost (pan

33

6.0 Conclusion
This paper investigates corporate hedging activities by using data from gold mining industry in Australia. Similar to North American companies, gold mining companies in Australia have employed a broad range of hedging levels. The paper finds that Australian companies in 1997 are more actively involved in gold derivative markets than their counterparts in North America in 1993. This may suggest a general risk-averse attitude among Australian managers when comparing with North American managers.

This paper also finds that corporate hedging activities are associated with some of firm characteristics, however, in different patterns. In contrast to the linear relationship suggested by previous researches, this paper finds evidence supporting non-linear relationship between the possibility of financial distress and hedging levels. High and low cash production costs are found to be associated with low hedging percentage, whilst medium level cash production cost is associated with high hedging percentage. This finding is consistent with the hypothesis in Stulz (1996) that companies in healthy position and in financial distress tend to speculate in market. However, not all indicators of financial distress act in this non-linear pattern. We find no evidence to support the nonlinear relationship between leverage and hedging level as Stulz (1996) originally suggested. In consistent with some earlier researches, this research only finds weak evidence supporting the positive correlation between leverage and hedging level. This paper also tests the relationship between hedging level and cash flow since most surveys report that the primary objective of hedging is to manage cash flows. Evidence is found that hedging percentage is associated with internal cash flow constraints. However, this paper finds no evidence at all to suggest the relationship between hedging activities and cash flow from external funding sources. Surprisingly, hedging is not associated with funding sources from debt market as information asymmetry theory strongly suggested. Overall, this paper concludes that hedging level in Australian gold mining industry tends to be associated with internal operation factors of companies. Cash production costs and internal cash flow from business operation, which are difficult to manipulate through finance activities, are the major determinants of managements choice on hedging level. The close relationship between companys operation factors and derivatives evidenced in the research drives us to have an inference on derivative usage different from the

34

traditional view: Corporate may use derivatives as part of operation decisions. As suggested by Tufano (1997), gold companies themselves have real options in production (flexible production according to market movement). Derivatives could be used to complement the call options a gold mining company already has on its operation. By using derivatives, company can maximise its production and minimise possibility of financial distress. From this point, derivatives differ themselves from a pure finance activity such as changing leverage that virtually have no impact on production. This may suggest that Modigliani and Miller framework is not applicable to derivatives. Therefore, future theory on derivatives should also look at their impact on companys operation rather than purely focus on their impact on corporate finance.

7.0 Limitation
Tufano (1997) pointed out that it is dangerous to draw conclusion on derivative usage by experience in a single industry. While gold mining industry provides an ideal environment for derivative research, the specific environment also limits broad implication of findings in gold mining industry for other industries having very different backgrounds. Another major limitation of this research is the small sample size. It is usually difficult to reject null hypothesis if the effect is not particularly strong with small samples. Also, statistical test results could change dramatically with a single movement in the data. There is a problem of representative of the sample for the whole population. This problem could be severe when using single year data. Further research is clearly needed to test the conclusions drawn in this research.

35

Bibliography Batten J., R. Mellor & Wan V., (1992), Currency Risk Management In Australia, Working Paper, Monash University. Berkman H. K., M. E. Bradbury & Magan S., (1997), International Comparison of Derivative Use, Financial Management, Winter, Tampa. Breeden D. & S. Viswanathan, (1990), Why Do Firms Hedge: An Asymmetric Information Model, Working Paper, Duke University. Bodnar G. M., G. S. Hayt & Marston R. C., (1996), 1995 Wharton Survey of Derivative Usage by US Non-Financial Firms, Financial Management, Winter, Tampa. Calomiris C.W., & R. G. Hubbard, (1988), Firm Heterogeneity, Internal Finance and Credit Rationing, Working Paper 2497,National Bureau of Economic Research, January. Colquitt L. L. & Hoyt R. E., (1997), Determinants of Corporate Hedging Behaviour: Evidence from the Life Insurance, Journal of Risk and Insurance, December, Mt.Vernon. Cumming J. D., R.D. Phillips & Smith S. D, (1997), Derivatives and Corporate Risk Management: Participation and Volume Decisions in the Insurance Industry, Paper presented at the 1997 conference of the international Association of financial Engineers, Boston, September 22-24. ---, --- & ---, (1998), The Rise of Risk Management, Economic Review - Federal Reserve Bank of Atlanta, First Quarter, Atlanta. Dolde W., (1996), The Trajectory of Corporate Financial Risk management, Journal of Applied Corporate Finance, Vol 6, 33-41. DeMarzo, P. & Duffie D., (1992), Corporate Incentives for Hedging and Hedge Accounting, Working Paper, Northwestern University. Edwards F. R., & M. S. Canter, (1995), The Collapse of Metallgesellschaft: Unhedgable Risks, Poor Hedging Strategy, or Just Bad Luck?, Bank of American: Journal of Applied Corporate Finance, Vol. 1, Spring, Pp86-105. Fazzari S. M., Hubard R. G. & Petersen B. C., (1988), Financing Constraints and Corporate Investment, Brooklings Papers on Economic Activity, Pp141-206. Federal Reserve Bank of Atlanta Research Division, (1993), Financial Derivatives: New Instruments and Their Uses, Georgia. Froot K. A., D. S. Scharfstein & Stein, J. C., (1993), Risk Management: Coordinating Corporate Investment and financing Policies, The Journal of Finance, Vol. 68, No. 5, Pp1629-58. Gunther J. W. & T. S. Siems, (1995), The Likelihood and Extent of Bank Participation in Derivative Activities, Federal Reserve Bank of Dallas working Paper, May. Hull J., (1989), Option, Futures, And Other Derivative Securities, Prentice Hall, New Jersey.

36

Love R. & R. Argawa, (1997), The Economic Benefits of Gold Price Risk Management, Working Paper, Monash University Main S. L. (1996), Forward market, Stock markets, and the Theory of the Firm, Journal of Finance, Vol. 42, 1167-85. Miller, M. & F. Modigliani (1961), Dividend Policy, Growth, and the Valuation of Shares, Journal of Business, Vol. 34, Pp411-33. Modigliani F., & M. Miller, (1958), The cost of Capital, Corporation Finance, and the Theory of Investment, American Economic Review, Vol. 48, Pp261-97. Nance D. R., C. W. Smith Jr & C. W, Smithson (1993), On the Determinants of Corporate Hedging, Journal of Finance, Vol. 68, Pp267-84. Scholes M. S., (1998), Derivatives in a Dynamic Environment, The American Economic Review, June, Nashville. Sinkey J. F. &D. Carter (1994), The Determinants of Hedging and Derivatives Activities by US Commercial Banks, Working Paper, University of Georgia. Smith D., (1997), Aggressive Corporate finance: A Close Look at the Proctor & Gamble / Bankers Trust Leverage Swap, Journal of Derivatives, Vol. 4, Pp67-79. Smith L. W. & R. M. Stulz, (1985), The Determinants of Firms Hedging Policies, Journal of Financial and Quantitative Analysis, Vol. 20, December, Pp391-405. Strange S., (1986), Casino Capital, Basil Blackwell, Oxford. Stulz R. M., (1984), Optimal Hedging Policies, Journal of Financial and Quantitative Analysis, Vol. 19, No. 2, June, P127-40. Stulz R. M., (1996), Rethinking Risk Management, Bank of America: Journal of Applied Corporate Finance, Fall, Vol. 9, Pp8-24. Tufano P., (1996), Who Manages Risk? An Empirical Examination of risk Management Practices in the Gold Mining Industry, The Journal of Finance, September. Tufano P., (1997), The Determinants of Stock Price Exposure: Financial Engineering and the Gold Mining Industry, Working Paper, Harvard Business School.
Webster A. L., (1995), Applied Statistics for Business and Finance, 2nd Edition, Irwin, Chicago.

37

Appendix I
35 Companies Included in the Sample AAA AFN AUG CNG CTR DOR GCM GGS GMA HAO KGM LYG MMC NFM OTR PEM RSM SGW Acacia Resources Ltd Australia Gold Field Aurora Gold Ltd Central Norseman Gold Centaur Mining and Exploration Dome Resources NL Great Central Mines Ltd Ghara Gold Mines Ltd Gold Mines of Australia Ltd Haoma Mining Kidstor Gold Mines Ltd Lynas Gold NL Macraes Mining Co Nnorth Flinder Mines Ltd Otter Gold Mines Perolya Mines Ross Mining NL Sons of Gwalia Ltd ABN ARS CAT CRS DGD EMP GGR GLD GMS HAR LHG MLG NCM NML PDG RGS SBM Morgans Gold Mines Ltd Australia Resources Ltd Camelot Resources NL Crosus Mining NL Delta Gold Emperor Mines Ltd General Gold Resources Goldfield Ltd Gold Mine of Sardinia Ltd Hargraves Resources NL Lihir Gold Ltd Mt. Lyshon Gold Mines New Crest Mining Miugini Mining Placer Dome Gold Ranger Minerals NL St. Babara Mines

Companies Using Derivatives but not Fully Disclosed AGR GWC UGM Australian Gold Resources Ltd Gwalia Consolidated Union Gold Mining Co CTO WER Charters Tower Gold Werrie Gold Ltd

Companies Have No Reserves Economically Viable at the End of 1997 Financial Year BCD CMX EQR KCN LEO MZG SCN GDM ALX BDG COG Beaconsfield Gold NL Climax Mining Ltd Equinox Resources NL Kingsgate Consolidated LEO shield Exploration Menzies Gold Mines Star Mining Corporation Goldsteam Mining All State Exploration NL Bendigo mining Coolgardie Gold CPC COM JWM LVG LSE PSR TRY SRI ARC CCR WGR Carpenter Pacific Resources NL Dominion Mining Ltd Johnsons Well Mining Laverton Gold NL Lone star Exploration Preston Resources NL Troy Resources NL SIPA Resources International Amalg Resources NL Capricorn Resources Australian NL Westgold Resources

38

Appendix II
In their breakthrough paper, Black and Scholes (1973) use differentiation technique to obtain the exact formula for the prices of European call and put options. The price of European options can be determined by several factors: 1.The current spot price (S) 2.The strike price (X) 3.The time to expiration (T-t) 4.The volatility of the underlying instrument ( ) 5.Dividend expected during the life of the option The price of a call option (c) and the price of a put option (p) are:

c = SN ( d 1 ) Xe r ( T t ) N ( d 2 ) p = Xe r (T t ) N ( d 2 ) SN ( d 1 ) ln( S / X ) + (r + 2 / 2)(T t ) d1 = T t ln( S / X ) + ( r 2 / 2)(T t ) d2 = T t

In a risk neutral world, as assumed by Black-Scholes model, St has a lognormal distribution. That is,

ln St ~ (ln S + ( r 2 / 2)( T t ), T t )
A European put option will be exercised only when the spot price at maturity date (St ) is less than the strike price (X). The cumulative probability that St is less than X equals to N(z), where z is:

ln X ln S + (r 2 / 2)(T t ) T t ln S ln X + (r 2 / 2)(T t ) = T t ln( S / X ) + ( r 2 / 2)(T t ) = T t = d2 z=

39

Figure 1

N(-z)

P =1-N(z) =N(-z)

-z

Therefore, the cumulative probability that X is greater than St is: Delta for put options =N(z) = N(-d2 ) Similarly, a call option will be exercised only when the strike price (X) is less than the spot price at maturity (St ). As the standard normal distribution curve is symmetric with the mean equal to 0 and variance equal to 1. The cumulative probability that strike price (X) is less than St can be obtained, as shown in figure 1: Delta for call options = 1 N (z) = N(-z) = N(d2 )

40