• No results found

The Commodity Effect: Empirical Research into the Effect of Commodity Price Changes on Firm Value

N/A
N/A
Protected

Academic year: 2021

Share "The Commodity Effect: Empirical Research into the Effect of Commodity Price Changes on Firm Value"

Copied!
45
0
0

Bezig met laden.... (Bekijk nu de volledige tekst)

Hele tekst

(1)

The Commodity Effect: Empirical

Research into the Effect of Commodity

Price Changes on Firm Value

(2)

The Commodity Effect: Empirical

Research into the Effect of Commodity

Price Changes on Firm Value

T.W. Hoogenboom, 1059947 Amsterdam, July 2006

Faculty of Management and Organization Rijksuniversiteit Groningen, the Netherlands Graduate lecturers: dr. C.L.M. Hermes

dr. W. Westerman

The author is held responsible for the content of this thesis; the copyright of this thesis rests with the author and no quotation from it or information derived from it may be published without the prior written consent of the author.

(3)

Preface

With this thesis, I finish my study Business Management (Bedrijfskunde) at the Rijksuniversiteit in Groningen, the Netherlands. In the final year of following courses for my major in International Business & Management I became interested in the factors that influence stock value. The price of oil had only been recently linked to stock valuation and other commodities and I wondered whether commodities other than oil would show to have a similar effect on stock valuation. In writing this thesis, I was supervised by dr. C.L.M. Hermes and dr. W. Westerman of the Faculty of Management and Organization in Groningen.

Special thanks go to a diverse group of people who have helped me along the way to the completion of this research. I would like to thank my supervisors – and especially dr. C.L.M. Hermes – for supporting me in finding the right words and approach for writing this thesis. I would also like to thank my parents for their tremendous efforts and faith in me and the financial support which has enabled me to complete my study.

Amsterdam, July 25th 2006 Thijs Hoogenboom

(4)

Management Summary

Firm value determination is a central issue in empirical finance. Whereas many authors have addressed the role played by macroeconomic factors – such as the interest rate, inflation rate or the exchange rates – on the valuation of stocks, very little has been published on commodities in this context. This thesis seeks to evaluate the significance of commodity prices fluctuations for the determination of firm value. Whereas literature on commodity prices has focused mainly on the effect that price fluctuations have on the output price of goods, nothing has yet been published concerning the effect of commodity price fluctuations on the value of firms. Using stock prices as a proxy for firm value, the objective of this thesis is to expand on previous studies on the topic of stock price determination by evaluating the effect of commodity price changes on the value of firms between 1980 and 2005. In doing so, exposure to price fluctuations is measured on two levels. Using regression techniques this thesis evaluates the effect of commodity price changes on the valuation of firms on both national as well as on firm level. Firstly, fourteen national stock indices are examined for exposure to price fluctuations of the seventeen commodities included in the CRB Reuters Commodity Index. Consequently, this paper evaluates the extent of the effect that fluctuations of food commodity prices

have on the valuation of firms that specifically make use of certain commodities. In doing

so, the stock value of 257 food producing firms in Japan and the US are examined for exposure to price fluctuations of crude oil and eight other commodities specifically used in the production of food products. These commodities are wheat, corn, soybeans, lean hogs, live cattle, cocoa, coffee and sugar. The results show that on national level, a large percentage of the sample is significantly exposed to price fluctuations in crude oil and heating oil. This result confirms earlier findings by Driesprong et.al. (2005) and can be attributed to the dependency on energy as a resource by a broad variety of companies listed on the national indices of the countries examined. When examining firms that specifically make use of certain commodities, the results show that the food producing firms in the sample are insufficiently exposed to commodity price changes to allow for drawing conclusions for firms across the entire sector. The reasons for limited exposure can be attributed to the ability of firms to transfer production costs changes to the customer, diversification or successful hedging policies.

(5)

TABLE OF CONTENTS

1. INTRODUCTION ... 6

2. CONCEPTS AND THEORY ... 9

2.1FIRM VALUE... 9

2.2LITERATURE CONCERNING FIRM VALUATION... 12

2.4DETERMINANTS FOR COMMODITY PRICES... 15

2.5FIRM POLICY AND COMMODITY PRICE FLUCTUATIONS... 19

2.6STUDIES CONCERNING INPUT PRICE CHANGES... 23

3. HYPOTHESES... 25

4. METHODOLOGY AND DATA ... 27

4.1METHODOLOGY... 27

4.2DATA... 29

4.2.1 Stock Market Data ... 29

4.2.2 Commodity Data ... 31

5. RESULTS... 33

5.1RESULTS FOR HYPOTHESIS 1 ... 33

5.2RESULTS FOR HYPOTHESIS 2 ... 36

6. DISCUSSION AND RECOMMENDATIONS ... 40

6.1DISCUSSION... 40

6.2LIMITATIONS TO THIS STUDY... 42

6.3RECOMMENDATIONS... 42

(6)

1. INTRODUCTION

When a natural catastrophe ruins grain crop or a hurricane destroys a key energy distribution channel, the disruption of the supply and demand balance results in commodity price changes. Whereas financial assets are paper-based and can be stored safely in say a desk drawer, commodity distribution and storage is a complex and expensive endeavour so the production cycles of many natural resources are designed to reduce costs and limit spoilage. Given that storage and distribution costs are high, the production of many commodities is carefully attuned to the expected demand. In the case of a sudden shortage in supply or an increase in demand, governments, banks and companies have limited access to solutions leading to a short term stabilization of the commodity markets. Due to the lag time between planting and harvesting for instance, such a shortage cannot be instantly compensated. With limited tools to stabilize the market in the short-run and a slow production response, the market can only respond by adjusting the price of commodities.

Around 25 percent of world merchandise trade consists of primary commodities, making them a key determinant of developments in the world economy (Cashin and Cheetham; 2001). Global commodity demand is expected to increase considerably due to the growing economic wealth of large developing countries such as China, India and Russia. Due to this development commodities are becoming a more and more important assets for both investors and companies using them. The commodity markets are characterized as highly volatile (Pindyck; 2001) and are greatly dependent on changes in supply and demand. Fluctuations in the price of commodities may have substantial effects on the cost price of products containing these commodities. Traditional wisdom suggest that when prices and volumes are kept constant and the variable cost price of a product increases or decreases, this will affect a company’s net profits. In this situation, a change in production costs will directly influence the current and future expected cash flows.

With commodities representing a large share of the feedstock used in many industries in the world today, changes in their prices may have a significant effect on expected cash flows and firm performance, and therefore firm value. This thesis is an explorative study

(7)

aimed at finding out to what extent changes in commodity (input) prices can be shown to have an effect on the valuation of stocks.

The empirical knowledge about the factors that influence stock return is continuously being updated. Numerous studies have examined the effect that various macro-economic variables have on the behaviour and volatility of stock value. Some studies find that variables such as interest rate (Campbell; 1990), unemployment (Boyd, Jaganathan and Hu; 2001) and inflation rate (Murphy and Sahu; 2001) can be said to determine stock returns to some extent. The research done in this field typically investigates the effect that financial variables have on stock value and very limited research has focussed on the influence of input price changes on the value of stocks. Kia (2003) includes the commodity index1 in his model and concludes the commodity index can be seen as a determinant of stock price changes. Only recently Driesprong, Jacobsen and Maat (2005) were the first to establish a relationship between the price of oil and stock value. However, their research was limited to one commodity (oil) and only investigated this relationship on national stock indices (NSI). The question of whether individual commodities have an effect on stock value on firm level has remained almost untreated.

The objective of this research is to expand on recent research done on the topic of firm exposure to changes in macro-economic factors by looking at the extent to which commodity prices affect (1) the value of national stock indices and (2) individual firm value, for which the value of stock is used as a proxy. The analysis should contribute to the understanding of the relevance and the effect of macro-economic price changes for the valuation of firms by providing insight to the extent to which stock prices are subject to changes in the price of commodities.

Elaborating on the works of Driesprong et.al. (2005) mentioned above, this research aims to investigate whether price fluctuations of a total of seventeen commodities have an effect on the value of the national stock index in fourteen countries. Subsequently, this thesis is developed further by analysing the extent to which the firm value of companies that specifically make use of certain commodities is affected by changes in their prices. For this purpose, the effect of food commodities on stock value is analysed for firms

1Measured as a fixed-weight index of the spot or transaction prices of 23 commodities produced in Canada and sold in world markets.

(8)

operating in the food producing sector. The goal of this research is to find out whether there is a statistically significant relationship between fluctuations in the price of commodities and the value of stocks. In line with this objective, the following research question is formulated.

What is the extent of the effect that fluctuations of food commodity prices have on the valuation of firms that specifically make use of certain commodities?

By approaching the question of exposure to commodity prices on both NSI level as well as on firm level, I intend to show that the price of some commodities prove to be relevant for firm valuation across all industries whereas others are only relevant for the value of specific firms. Finding such a relationship would imply that investors and firms are paying insufficient attention to the price developments of raw inputs and that firm policy should be more focussed on limiting the risk run due to commodity price fluctuations.

This paper will be developed as follows. Section 2 defines the concepts of firm value and commodities and develops a theory that suggests a relationship between commodity prices and firm value. Section 3 constructs two hypotheses. Section 4 will outline the methods used to test the hypotheses and describes the data set used for this research. Section 5 discusses the results of the analysis. The final section offers a discussion, recommendations and conclusions.

(9)

2. CONCEPTS AND THEORY

2.1 Firm Value

The value of a firm can be said to equal the total price that an investor is willing to pay for a company. For publicly traded companies, this amount is found by multiplying the stock price2 by the amount of outstanding stocks. Firm value corresponds to the

market value – the total price that an investor would be willing to pay – of the firm’s assets (machines etc.) and commercial activities3. Or put differently, it equals the present

value of all the money that a firm is expected to make in the future; the net present value (NPV) of all future expected cash flows. These cash flows are dependent on how well a firm is performing (or is expected to perform in terms of sales and revenue), yet they may also be affected by movements in macro-economic factors upon which the firm may have little influence.

A company’s value – based on the earnings value of its assets – may change when the economy rises or falls, or when changes are brought about in the company. The key to understanding the value of a firm is that it is the price that an investor is willing to pay for the company (or part of it by purchasing stock) based on the information available to the investor. The investor’s perception of what price should be paid for a company relies on the way this information is interpreted by the investor. For example, the value of an oil company that strikes a new source of oil will rise because the earning potential and therefore (present value of) the expected future cash flow of the company is increased. Its stock will be in demand with investors (who expect higher future cash flows) which raises the price of the stock. On the other hand, if a new entrant threatening to take substantial market share is expected to join the market, a company’s value in that market will fall because of a potential loss in market share. Information about such changes in the economy can affect the perception of the investor and alter the price they are willing to pay.

2 The oldest and most conservative method of valuing stocks is the dividend discount model (DDM). When

assuming zero growth and a constant discount rate, a value is obtained by discounting and adding the value of all expected future dividends. If nothing else, the DDM demonstrates the underlying principle that a company is worth the sum of its discounted future cash flows. It is based on the formula for a perpetuity that calculates the Net Present Value (NPV) of all discounted future cash-flows

(10)

All publicly available information is priced by the market (i.e. all investors put together) and will cause the value of a firm to rise or fall according to the nature of the information. Investors convey their opinion about what the price of a stock should be by buying or selling stocks. Investors are therefore always looking for signs that may indicate how the price of a stock will develop. These signs are conveyed in changes in firm specific conditions (such as striking a new oil source) and changes in macro-economic variables. For example, a change in interest rate may increases a company’s cost of capital and a devaluation of the home currency may make foreign direct investments cheaper for a home firm. Higher interest rates increase the returns demanded by investors and a weak domestic currency will increase the costs of purchasing goods abroad. In both cases, the NPV of a company’s cash flows is changed due to changes in the economic environment.

Apart from a company’s projected earnings, the condition of the economy and the value of information perceived by investors, another less visible determinant of stock price is investors’ sentiment or emotion. Given the intangibility of emotion, behavioural finance will not be considered in this thesis.4

Information and certainty about future events will however, always be limited (to a certain extent) and therefore calculating a precise value is accompanied by a certain amount of risk. It is important to understand that prices change because of new information affecting the expected earnings – and therefore its stock – comes to the market. Both the buyer and the seller are faced with the risk that unexpected events change the expected value of the stock upon which they based their decision. Investors are compensated for their investments based on the amount of risk run commonly measured as the volatility of the return on a stock or investment. Investors in highly volatile stocks will demand a higher return on their investment due to an increased risk of losing capital.

4 Although in theory all market participants are assumed to be fully rational, investors may be influenced

by their emotions when trading stocks. Bollhorn (1999) identifies fear of losing (more) money as the emotion that causes traders to sell stock, and greed to buy them; hoping that they will generate a lot of money. (This effect was visible during the internet bubble at the end of the 20th century; investors were

buying lots of stocks driving prices up whereas the value of the assets underlying the stock price turned out to be worth much less.) Both situations result in a stock price that is not equal to the value of the assets available information. Although this factor contributes to the valuation of stocks, irrational behaviour by investors is not considered in this paper.

(11)

We can conclude that there are three generic elements affecting the present value of future cash flows and therefore the price determination of stocks. Conditions in the economy can give a firm a bright (gloomy) outlook causing expected earnings to rise (fall); new information about the firm can increase or decrease its perceived earnings potential and perceived value of information. Given the uncertainty around predicting future events, all these factors are all associated with a certain amount of risk. These elements are depicted in figure 2.1.

Figure 2.1: Generic elements that influence investors’ interpretation and subsequently stock value

Investors as well as academic researchers attempt to evaluate determinants of stock price development by examining various variables within this framework aimed at constructing theories that can explain and quantify stock price behaviour. Shareholders are keen to know about these factors in order to safeguard their investments. Moreover, they will want to exercise their decision-making power if corporate policy does not adequately deal with exogenous factors influencing firm value. Scientists attempt to find explanations for movements in the economy and shifts in firm value, and try to build theories that will account for tendencies of markets to move in certain ways and that can predict future developments. The literature concerning variables that have been hypothesized to influence stock valuation is discussed in the following section.

Condition of the general economy Firm specific information available Investors’ sentiment Stock Value Investors Buy/Sell according to interpretation and take risk Market conditions of supply and demand

(12)

2.2 Literature Concerning Firm Valuation

Stock return predictability and the factors that determine the value of stocks have been a central issue in empirical finance. Literature written about the factors that may have an impact on stock return fluctuations include various categories of variables (see Cremers; 2002). Prior to 1981, much of the finance literature viewed the present value of dividends to be the main driving factor behind the level of stock prices (Balke; 2002). Since then many other macro-economic variables have been employed to attempt to forecast stock behaviour. Among the most popular predictors for the development of stock prices are nominal interest rate and the dividend yield. Whereas dividend yield appears to be the most favoured predictor in applied work (Ang and Bekaert; 2001), many authors claim that interest rates ar also a sound predictor of market movements (see Santoni; 1994). This relationship in the form of the expectations theory was developed by Fama and Schwert in 1977. Their model states that expected nominal stock returns equal the nominal Treasury bill rate plus a constant (Campbell; 1985). Although the expectations theory was rejected in various studies it did form the basis of extended research on the relationship between interest rate and stock predictability. It led Campbell, among others, to show that the variables which had been used as proxies for risk premia could also predict excess stock return. However, the opinions on the relationship between interest rate and stock return remain conflicting.

Research, however has not been limited to these factors. The inflation rate (e.g., Bodie; 1976, Fama; 1981); money stocks (Geske and Roll; 1983); aggregate output (Marathe and Shawky; 1994); the unemployment rate (Boyd, Jaganathan and Hu, 2001); interest rates (Campbell, 1990; Hodrick, 1992); and term and default spreads on bonds (Chen, Roll and Ross; 1986) have been considered in recent research to function as indicators for stock market movement. However, little consensus has been reached on the determinants of stock value. This can be concluded from the diverging results obtained by many researchers as noted by Cremers (2002).

Following the increase of businesses operating in an international environment, much attention has also gone to the effect of exchange rate fluctuations on the valuation of stock. The approach used in these analyses is based on the extent to which a firm’s NPV

(13)

is exposed to changes in exchange rates. Adler and Dumas (1984) were among the pioneers in this field of economic exposure research which was later developed more extensively by many others (see for example Jorion; 1990, He and Ng; 1998, Dominguez and Tesar; 2006) The basic idea behind economic exposure is that the valuation of a firm’s earnings, future cash flows and foreign investment is changed due to unexpected changes in exchange rates. Though a significant relationship between exchange rates and stock value is reported in the U.S. (Chamberlain, Howe and Popper; 1996), some provide relatively little evidence on exposure in major European economies (Rees and Unni; 2005). In both research within the U.S. and elsewhere, researchers are divided on the true effect of exchange rate fluctuations on firm value.

In conclusion, a lot of efforts have been made to determine how stock value is determined and which of the variables that affect the perception of investors can be said to predict how a stock price will develop. Although researchers have found some significant relationships, they are far from unified when it comes to which variable can be considered the soundest predictor of stock value. The variability in approaches by different authors, has led to the lack of consensus in the examination of the effects of various macro economic variables. Furthermore, diverging results and the fact that many different variables have been used in research on stock valuation suggest that the topic has not yet been fully covered. In this thesis I have identified a gap in the literature and notice that the effect of changes in the price of raw inputs on firm value has not been fully addressed. Given that the majority of products available to us today are manufactured using commodities, there is a good reason to expect that the price of commodities will have some effect on the performance and value of firms. This will be explained in the next section.

2.3 Commodities and Firm Value

Similar to a change in the cost of capital, one could argue that a change in the price of the inputs used in the manufacturing of products can be important for the success of a company and it’s NPV. When input prices change, they affect the marginal cost price of a produced good (ceteris paribus) which in turn may affect a company’s profit margin if all

(14)

other factors are kept constant. The theory behind this presumption will now be discussed in and argues that a change in the marginal cost price of goods can potentially lead to a change in the NPV of future expected cash flows of companies.

For this purpose, we assume a three factor Cobb-Douglas function in which the amount of products that can be produced (Q) is a function of the amount of Labour (L), Capital (K) and Materials and Supplies (M) in a product;

) , , ( ) )( )( ( * L K M f L K M A Q= α β χ = [ 2.1 ]

In which alpha, beta and gamma denote the percentage of factor used per product and A equals a parameter greater than zero that measures the productivity of available technology. Given that a firm can purchase each factor at factor cost price of wL, wK and wM respectively, the cost of producing Q units will be:

M wM K wK L wL Q c( )= * + * + * [ 2.2 ]

Hence a change in any of the variables will alter the cost price of the good.

There has been a focus in finance literature on the first two factors of this equation: The cost of Labour has been examined implicitly by looking at unemployment rate for instance (Boyd et.al.; 2001) and the cost of capital has been extensively researched in topics such as exchange rate exposure, inflation and more implicitly by looking at dividends in relation to stock price movements5. The third factor Materials and Supplies has not yet been researched widely even though it constitutes one of the factors that determine the amount of products a firm can produce and at what cost.

As opposed to labor and capital, materials and supplies encompass the physical (raw) inputs that are used to produce a good or service. In the context of this research, the inputs considered will be limited to commodities. Commodities form the raw material for every single (tangible) product we as people can acquire and the efficiency of these markets is vital for companies producing these products. The costs of commodities form

5 More dividend is returned to investors implying that the cost of capital is too high to ensure returns

(15)

an important determinant of the prices at which companies can produce and sell their products. The commodities discussed in this thesis are oil and those commodities primarily used in the production of food. They are commodities which are grown and harvested on the land or bred. Due to their exposure to nature, the supply in such commodities can change unexpectedly making their price trends unpredictable. We therefore look at which factors are determinant for the prices of commodities.

2.4 Determinants for Commodity Prices

This section looks at the development of commodities and the factors that influence their prices. In doing so, it will attempt to describe how and why commodity prices fluctuate. Answering this question will give insight into commodity price volatility and will provide an explanation as to why prices fluctuate.

Agricultural commodities have been shown to have three distinct characteristics that help to explain their price development and volatility6.

a) Commodity prices show a general decrease in real price level in the long run. b) Commodity supply is prone to unexpected changes and fixed in the short term. c) Price demand elasticity is very low due to low substitutability

These three characteristics are now described. a) Price decline over the long run.

In the long run, commodity price show a decrease in real price (Borensztein and Reinhardt; 1994). Overall the prices in real terms of agricultural commodities for instance have declined by nearly 50% over the past 40 years. This decline can be explained by growing technology (A in equation 2.1) due to improved machinery and pesticides accompanied by more advanced methods to grow and cultivate agricultural products by the development of more efficient fertilizers. Mainly due to the latter factor, the Food and Agriculture Organization of the United Nations (FAO)7 notes that increased productivity and efficiency and a decrease in the farm value in products (see figure 2.2), supply has typically exceeded world demand, thus repressing the price. Similar factors have caused

6 http://www.bized.ac.uk/virtual/ accessed on 20-07-06

(16)

declines in other commodities where technologies are replacing labor more and more, increasing the efficiency of the production processes.

US Farm Value in Food Products

1980 - 2002

0% 5% 10% 15% 20% 25% 30% 35% 40% 1980 1984 1988 1992 1996 2000 Year Fa rm V al ue Farm Value (%)

Figure 2.2: Farm value of food products in the US 1980 - 2002

b) Shocks in short term supply causes prices to change.

Secondly, commodity markets are subject to discontinuities or changes in trend, which range from natural catastrophes and political/military interventions to structural market changes (Labys; 2005). Such discontinuities can occur unexpectedly and lead to higher price levels in most cases. Weather conditions can also greatly influence the global supply of commodities as draughts or floods may ruin crops and hurricanes can instantly paralyze production on oil rigs or refineries. Assuming that global markets are efficient, new information regarding the supply of commodities will immediately be reflected in the price of commodities8. When these shortages appear unexpectedly, the consequences for companies depending on commodities for their inputs can be quite significant.

For annually produced commodities, such as corn and wheat, supply is fixed in the short run and is limited depending on the amount of land currently available to grow the

(17)

crops. When nature or political influences alter the supply of commodities, this will cause the price level to swing viciously (see figure 2.3)

Figure 2.3: Cobweb model (Kaldor): Prices fluctuate when supply is fixed and dependent on previous periods.

The illustrated model in figure 2.3 is known as a cobweb model which was identified by Nicholas Kaldor in 19349. It shows the effect that a short term change in supply can have on price levels. It is assumed supply is fixed and dependent on events that happened in a previous period (such as natural catastrophes or political intervention). When such an event decreases the total supply of a commodity from Q1 to Q2, The shortage of supply

will lead to an increase in price (from P1 to P2) according to demand conditions (curve D).

Given this higher price, farmers will increase the production (growing) of this commodity at the cost of other activities. Due to a lag between planting and harvesting, producers cannot anticipate how much of the commodity will be available in the next period. If all farmers increase their production, this will result in an excess of supply in the next period, causing the price to drop (to P3). The farmer will consequently reduce production of the

commodity in favor of other goods, leading to a shortage (and higher price P4) in the

following period. As this process is repeated, the magnitude of price volatility will excel. Based on the cobweb model we can state that for commodities, the supply in the short run is constant and dependent on the supply in previous periods; shortages in one period are compensated by over production in the next or vice versa, causing the price of the commodity to swing viciously, making it highly volatile.

9See: http://en.wikipedia.org/wiki/Cobweb_model Q4 Q2 Q1 Q3 P3 P1 P2 P4 S1 S2 S3 SLR D

(18)

Commodity prices are therefore very unpredictable and can vary widely from year to year. The implication for companies using commodities is that forecasting future commodity prices may prove to be extremely difficult. The question is whether firms possess adequate tools to anticipate changes in the supply of commodities.

c) Commodity Demand.

Demand factors also play a role in the development of commodity prices. Clem (1985) asserts that demand (and hence prices) for basic materials traded internationally may change rapidly because of exchange rate fluctuations, political turmoil or large purchases by governments. With higher prices when supply is low, companies will not refrain from purchasing commodities as they are vital inputs to many production processes. This occurs mostly in competitive markets for goods which have only limited substitutes. The demand curve for commodities is therefore said to be very inelastic. Basic economic theory predicts that demand for a good in this case will change less in relation to changes the product’s price. Figure 2.4 illustrates this effect.

Figure 2.4: The inelastic demand curve causes prices to fluctuate more as supply changes.

The curves Delastic and Dinelastic represent the demand curve of goods that is respectively

relatively elastic and inelastic and the curve S1 represents the supply curve. The Nash

equilibrium in the initial situation is achieved at point A where the quantity and the price of the products are Q0 and P0 respectively. A decrease in the supply of a good shifts the

supply curve from S1 to S. The equilibriums for both products are now represented by the

S S1 P Q Delastic Dinelastic P0 Pe Pi Qe Qi Q0 A

(19)

intersection of the two demand curves with the new supply curve S. As can be seen in figure 2.4, the effect that this shift has on the price and quantity of both products is different for demand curve. The product with the relatively elastic demand curve experiences a decrease in price (P0 Pe) as well as quantity (Q0 Qe) that is greater than

that experienced by the product with the relatively inelastic demand curve (Q0 i,P0 i).

Conversely, this implies that given an inelastic price demand curve, the price of a product will increase more when supply is decreased.

We have described three factors that influence the price development of commodities. A decrease in farm value and technological improvements cause a decrease in the long term price development, supply shocks cause large price volatility and an inelastic demand curve results in relatively high price changes. In the short term, the characteristics of supply and demand conditions cause changes in the price of commodities. Even though many governments provide subsidies to stabilize the commodity market in the long run, firms using commodities will still be faced with changes in the cost of production when unexpected events cause price fluctuations in the short run.

2.5 Firm Policy and Commodity Price Fluctuations

The question following is: what do extra costs (benefits) of production arising from increased (decreased) materials prices imply for the value of a firm? We have asserted that in the short run, no tools are directly available to counter the effect of commodity supply and demand destabilization apart from altering the price of commodities10. Therefore, when the supply of a commodity is insufficient to meet current demand due to a shortage of supply or an excess in demand, the price of that commodity will rise. Firms that use the commodity in the manufacturing of their product are faced with a higher average cost price of their product.

A producing firm faced with a change in the cost of an input does have several options in a short run perspective. The Economic Research Service of the US Department of Agriculture asserts that if the input cost increases, the firm can

(20)

i. absorb the higher costs by keeping its prices steady and accepting a lower profit level,

ii. pass on some of the higher costs by raising the price of products.

iii. adjust its production process and employ fewer units of the higher cost input by substituting one or more other inputs.

iv. hedge against price fluctuations by purchasing futures.

Logically, the firm is faced with the opposite options – higher profits, lower output prices or expanded input use – if the input cost decreases. Whereas the first option leaves retail prices constant, the others can directly affect food prices either by the firm raising the price of its food products or by food production adjustments that influence the amount of food available and thus its price.

i. Absorption.

Absorbing the extra costs of production due to price increases in inputs means that in case of constant output, these costs are taken out of the firm’s profit margin. A firm may choose to absorb extra costs in order to maintain the price level of their products and keep consumers satisfied.

ii. Raising Prices.

There have been several studies supporting the notion that food producers are able to pass on increased production costs to the consumer. In the US, the ERS has published various papers than examine the effect that input prices have on Consumer Price Index (CPI). Given the inelastic nature of the demand curve of food products, food producers are able to pass the extra costs of input on to the consumer by raising prices without losing on sales volume or market share. The pass-through method is based on the observation that in a perfectly competitive market with constant returns to scale, average cost equals marginal cost, which in turn equals the output price. Thus, any increase in input costs will add to the industry’s average and marginal costs, and therefore to the output price by the amount of the affected input’s share of operating cost. The higher cost is thereby “passed through” to ultimately be paid by the industry’s consumers (Lee;

(21)

2002). Presumably, raising the price of food to cover the extra incurred production costs leaves the profit margin and therefore the stock value unchanged. However, it is perceivable that producers dealing with increasing production costs may pass on more than this cost over to the consumer which could lead to a higher profit margin and hence increased stock value.

Several key factors influence how an input cost increase might affect the prices of products under conditions of competition among numerous firms. Given an increase in the price of a commodity used to produce a certain good, the increase in the product’s price will be larger when11:

• The input has fewer good substitutes in the production process

• Consumers have few good substitutes for the product, in which case demand is not decreased when prices are higher.

• A short period is considered. For example, weather or transportation problems can temporarily cause sharp price increases, with prices returning to previous levels once the problem ends.

• The share of the input in the total cost of producing food products is larger. In a short-run competitive market, food producers are price-takers in input and output prices since they are but one of many firms competing in the sector buying and selling goods. Moreover, the short-run perspective assumes that all production facilities are established in plants of a fixed size and cost.

iii. Adjusting the Production process.

Given that production capacity and capability is constant in the short run, adjustments to the production process can only take place in the long run. When considering the price of food products in the long-run, two types of responses affect its development. On the one hand, consumers more readily find and use substitute food products as more time passes, which would tend to make the price increase of a particular food larger in the short run than in the long run (mimicking the earlier case of weather or transportation

11 Source: USDA Briefing Room

http://www.ers.usda.gov/Briefing/CPIFoodAndExpenditures/howchangesininputcostsaffectfoodprices.htm accessed on 20-07-06

(22)

problems). On the other hand, permanent input cost increases will encourage producing companies to produce their products more cost-efficiently given that they operate in a competitive market by:

- changing (the composition of) their product in terms of the intensity of the use of a commodity or by substituting a commodity in the production process or;

- developing more efficient ways to acquire raw materials by investing in research and development (R&D) activities that increase the quality or quantity of available commodities.

Once again the first option – of changing the characteristics of a product – is presumably unlikely to occur as many companies will have to consider the marketing costs of selling a revised product to the consumer. Moreover, substituting an input may result in a loss of customer base. Increased efforts to invest in the development of more effective fertilizers or harvesting machines will lead to cheaper acquisition of commodities in the long run. This idea is supported by the gradual decrease in the (real) cost of commodities as observed in a longer time frame.

iv. Hedging.

Hedging for price fluctuations is a very common practice in the foreign exchange market in which forward contracts are bought and sold that allow the exchange of future units of one currency for future units of another currency at a fixed price. By using these contracts, a firm can hedge against its exposure to exchange rate fluctuations. Similarly, a firm can purchase commodity futures that allow buying or selling a commodity at a specific price and on a specific delivery date. Knowing before hand what the cost of an input will be eliminates the risk of facing unexpected price changes.

In conclusion, the possible effect of change in the price of a product’s inputs may have an effect on the valuation of a firm based on the possible change it may cause to a company’s expected NPV. In the short run, absorption of extra costs incurred due to increases in commodity prices will lead to a narrower profit margin. Over time, this effect may be reversed if the volume of sales goes up due to the competitive position if the firm offers its products at lower prices relative to their competition. Pushing increased costs

(23)

over to the customer will leave the profit margin constant if the price increase is equal to the extra costs. Anticipating increases in the price of commodities, companies may choose to hedge against commodity price changes in order to cover the risk of having to produce at a higher cost price. Of these four options, only the first would hypothetically lead to a change in the firm’s stock valuation. Absorbing the higher costs reduces the profit margin and may therefore decrease future expected cash flows, which will lead to a decrease in the value of the stock.

2.6 Studies concerning input price changes

There have been several studies that address the effects that price changes on inputs used in manufacturing processes. A common perspective from which these studies are concentrated around the effect of commodity price changes on the retail price. Reed, Hanson, Elitzak and Schluter (1997) assert that if food producing firms use the same proportions of inputs in the short run, and if consumers are assumed not to respond to retail prices changes, the full increase in average cost of the good (due to higher energy prices) will be passed on to consumers in the form of higher food prices. They estimate an average of 1.82 % increase in food prices in the short run and a 0.27 % increase in the long run in the hypothetical case that energy prices have doubled. Essentially therefore, food producers are able to push the some of the extra costs incurred due to higher commodity prices over to the paying customer (Furlong and Ingenito; 1996). This effect is supported in research done by Lee (2002). Moreover, the theme of a group of studies published in the American Journal of Agricultural Economics in 1997 was retail food price forecasting. Each of these studies emphasizes the importance of farm products and their price volatility in affecting food prices.

With respect to the influence of commodity price changes on the value of stock, very little has been published to date. Kia (2003) finds a commodity price index to be one of (many) significant factors to influence stock value in the (highly open) economy in Canada. In his paper, he notes that “…no study so far incorporated the commodity price

index in the determination of stock price or return” (Kia; 2003, p.39). The recently

(24)

price changes on the value of stocks. They do this by regressing the change in the price of oil on the value of the national stock index in various countries and find a statistically significant predictability for twelve out of eighteen countries as well as for the world market. They conclude that investors react immediately to obvious and observable effects of oil price changes, yet underestimate the impact of more general economic effects of oil price changes.

In summary, no widely published research has accounted for the effect that commodities have on stock valuation. Furthermore, the work by Driesprong et.al. (2005) only estimated the effect of oil price changes on national level. This thesis elaborates on the topic of firm exposure to exogenous variables by looking at the effect of commodity price changes on national level as well as the effect of these changes on firm level. The following section will develop the hypotheses that will be used to evaluate this effect.

(25)

3. HYPOTHESES

In the previous section, we have addressed some important issues concerning the conditions that could imply that commodity prices can be said to influence stock prices. Firstly, the cost of production of goods containing commodities rises and falls with the price of these commodities, secondly it is difficult to forecast the highly volatile commodity market and lastly the stock value of firms can be affected by commodity price changes if the extra costs (benefits) are absorbed in the firm’s profit margin. In order to test for any significant relationship between commodity prices and stock value, we formulate two hypotheses that are tested using regression techniques. The commodities used to perform the tests are those listed in the CRB Reuters commodity index described in section 4.2. The relevance of commodity prices for stock valuation is tested on national stock index level with the hypothesis (H1) that:

H1: Fluctuations of commodity prices have a significant effect on the value of national

stock market indices.

Given that national stock indices are composed of companies from many different sectors, we furthermore believe that crude oil will show a more significant effect than any other commodity. The reason behind this extension is that many firms (such as those in the service sector) in NSI do not produce goods containing commodities, yet are still dependent on energy prices for running their business. By taking this stance, this thesis addresses the question as to whether commodity prices other than crude oil (as shown by Driesprong et.al) can be shown to influence the value of NSI.

I have asserted that no recent literature has examined the relationship between commodity price changes and stock value on firm level. I therefore formulate the following second hypothesis (H2):

H2: Commodity price fluctuations have a significant effect on the stock value of firms that

(26)

To test H2, food commodities and firms in the food producing industry will be used. This

is described in more detail in the following section. Testing this hypothesis will determine whether firms using certain inputs are finding themselves to be exposed to the prices of these inputs. The purpose of examining the effect on both national stock index level and individual firm level is to distinguish between the effect that commodity price changes have from a macro- and microeconomic perspective. In case of a significant effect on macroeconomic level, the results imply that the effect is relevant across all sectors and therefore independent of the industry in which firms are active. When examining individual firms, any found effect is only relevant to those specific firms. A distinction is made between generally applicable and used commodities across sectors – energy commodities crude oil, heating oil and natural gas – and those commodities almost exclusively used in a particular industry. I expect the commodities in the first category to show a more significant effect on macroeconomic level and those in the second category to have a greater impact on firm value of individual firms specifically using these commodities. The following section describes the methods used to test these hypotheses.

(27)

4. METHODOLOGY AND DATA

4.1 Methodology

The analysis used to determine whether commodities have an influence on stock value is based on regression techniques. The hypotheses imply that there is a linear relationship between commodity price and stock value; a change in the price of a commodity should be reflected in a value change in the value of stocks. A linear relationship between two variables can be expressed mathematically as a regression line and tells us how Y is related to X.12 Given this equation, we can say that the slope of the equation (β , the coefficient) is found by differentiating the regression model13 and denotes how much the value of Y will change when X changes by a small (marginal) amount. The coefficient can be interpreted as the marginal effect that X has on Y. In the exchange rate literature, it is interpreted as the measure of economic exposure. Here the coefficient βis assumed as a measure of the risk of observing a change in stock value when commodity prices change. In this regression β constitutes the amount of exposure – and therefore risk – of stock value to commodity price fluctuations. Mathematically we use:

ε

β

α

+ ∆ + = t−x mt C R 1 {1}

To test our first hypothesis. In equation {1}, Rmt and Ct-x equal the independent and

explanatory variable respectively at time t, is the constant,

β

the regression’s coefficients and

ε

is the standard error term. What we look for is a statistically

significant causal relationship between the dependent and the explanatory variable. The predictive ability of X is typically assessed by examining the t-statistic and the OLS estimate of in the equation (see Koop; 2000). I test whether the coefficient of is significantly different from zero. In order to test the hypothesis that

β

≠0, the

12 i.e. Y =

α

+

β

X 13 i.e. =β

dX dY

(28)

statistic (t-stat) is calculated14 and translated into a P-value. The analysis will attempt to find any statistical significance at 0,10 and 0,05 denoted by any P-value15 less than 0,05 and 0,10. Where

β

1 is significant, the hypothesis of a commodity effect is accepted.

When testing the second hypothesis, I take into account that there are many factors that influence the price development of stocks. Following the literature written on exchange rate exposure (see for example Jorion 1990), in order to filter out the effect of commodity prices, a second explanatory variable is added to the regression in order to evaluate influences to stock value independent of the return in the general market. Similar to the inclusion of a risk free rate in the CAPM model, such a variable controls for market movements by including the return on the market portfolio estimated by the change in value of the NSI, Rmt. This variable is used as a proxy for fluctuations that influence the

value of Rit besides the variable under investigation, C. Therefore,

β

1measures the firm’s exposure independent of market’s exposure (

β

2) to these changes. Equation {1} is expanded for the analyses on firm level to:

ε β β α + ∆ + + = jtx mt it C R R 1 2 { 2 }

In which Rit denotes the monthly change16 in stock value of firm i [1…n] at time t, Ct-x is

the monthly change in the price of commodity j at time t-x. The following section will describe the data used for the variables in the regressions above.

14 using

b

s

t= βˆ in whichβˆ is the estimated value of β and sb the standard deviation (or standard error) of

βˆ where 1 ) ( 1 2 − − = = N Y Y s N i i .

15 The P-value provides a direct measure of whether t is large or small. 16 Computed as (P

(29)

4.2 Data

Having formulated and discussed our hypotheses and methodology, the following section will describe the data used to test the hypotheses and answer the research question.

4.2.1 Stock Market Data

All data concerning national stock indices and firm specific stock values have been collected using DataStream. The time period over which this analysis is done is a maximum of 26 years using changes in the average monthly values of the indices. The series start between January 1980 and December 1995 and all end at the end of 2005. For reasons of the reliability of the analysis, data is included if a series is available over a period of at least 10 years with a minimum of 120 monthly data points.

Although intending to use the same eighteen countries as Driesprong et.al. (2005) did in their research, due to insufficient data availability, we are limited to the following fourteen: Australia, Austria, Belgium, France, Germany, Hong Kong, Italy, Japan, the Netherlands, Singapore, Spain, Switzerland, the United Kingdom, and the United States. No NSI data was available for Canada, Norway and Sweden, and the available series for Denmark started after December of 1995. The listed monthly closing value are used in the analyses for the effects of commodities on national level (i.e. H1) and are used as a

proxy to estimate the average market return contained in the model for H2. Data from the

national stock indices of these countries is used in all analyses.

For the analysis on firm level, the food producing industry was chosen for two reasons. Firstly, the supply of agricultural commodities is greatly dependent on weather conditions and developments in a previous period. For this reason, unexpected fluctuations in supply may occur that can lead to sharp commodity price changes. Secondly, such commodities have very few substitutes which making the price demand curve very inelastic. Furthermore, as opposed to most other products, food will always need to be produced to feed people around the world.

The data for the analysis on firm level is also provided by the DataStream database. In here, companies listed on national stock exchanges for the fourteen countries used in this paper are categorized according to their activities. For the purpose of this analysis,

(30)

those companies active in the food producing industry were selected. Table 4.1 shows the amount of firms that meet these criteria in each respective country:

Food Producers

Market Total >120 observations

Australia 88 15 Austria 7 5 Belgium 13 6 France 50 22 Germany 29 14 Hong Kong 73 15 Italy 11 2 Japan 205 144 Netherlands 28 13 Singapore 42 9 Spain 17 4 Switzerland 22 9 United Kingdom 79 24 United States 439 113 Grand Total 1112 401

Table 4.1: Number of listed food producing companies in 14 countries and the amount of these traded prior to 1996.

The choice of using the food producing companies in the United States and Japan is twofold. Firstly, both countries have substantially more food producing firms within their borders and can therefore be said to collectively represent a majority of the global food production industry (57.9% of the sample). Secondly, when testing for the criteria of at least 120 observations in the past 26 years, the United States and Japan are the only two countries that contain a reasonable sample size needed to make a reliable analysis (113 and 144 firms respectively). Using a total sample of 257 companies, we test hypothesis 2.

(31)

4.2.2 Commodity Data

The trading of commodities occurs at various exchanges world wide There are four global dominating commodity indices on which world prices are quoted They are the Reuters/Commodities Research Bureau (CRB) index, the Goldman Sachs Commodity Index (GSCI), the Dow Jones AIG Commodity Index (DJ-AIG) and the Rogers International Commodities Index (RICI). When correlating the values of the indices (see table 4.2), it becomes clear all four indices are very closely correlated so choosing one is a practical and reliable way to characterize global market price movements. As Akey (2005) notes, any of these indices may provide investors with a similar commodity exposure. In terms of market selection, it is the Reuters Commodity Research Bureau (CRB) Index that attempts to create a broad measure of overall commodity price trends. The Index is calculated as the geometric average of each of its 17 market’s average price. For reasons of data availability and accessibility, this paper uses the data of the CRB Index. For nearly 50 years, this world-renowned index has served as the most widely recognized measure of global commodities markets and will therefore serve as a good indicator for price developments of global commodities.

DJ-AIG GSCI CRB Reuters RICI

DJ-AIG 1

GSCI 0.953 1

CRB Reuters 0.880 0.710 1

RICI 0.984 0.976 0.887 1

Table 4.2: Correlation matrix based on monthly changes in values The Commodities Research Bureau (CRB) Index

The only complete and accessible data source for commodity data over the period between 1980 until 2005 was the CRB Reuters Index. The CRB index is composed of 17 commodities divided into six main groups; energy, precious metals, grains, industrials, live stock and softs. Table 4.3 shows the various commodities and sub-groups the CRB index consists of and the weight of each individual commodity within the index.

(32)

CRB Index Composition Cotton 5,88% Copper 5,88% In du s-tr ia ls Sector Total 11,76% Gold 5,87% Platinum 5,87% Silver 5,87% Pr ec io us M et al s Sector Total 17,60% Crude Oil 5,87% Heating Oil 5,87% Natural Gas 5,87% E ne rg y Sector Total 17,60% Corn 5,87% Wheat 5,87% Soybeans 5,87% G ra in s Sector Total 17,60% Lean Hogs 5,88% Live Cattle 5,90% M ea ts Sector Total 11,80% Orange Juice 5,88% Cocoa 5,88% Coffee 5,88% Sugar 5,88% So ft s Sector Total 23,50% Grand total 100%

Table 4.3: Components and weight distribution of the CRB Index

The commodities used to test hypothesis 1 are all of the above. Model 2 uses the price of crude oil and those commodities associated with the production of food. These are: corn, wheat, soybeans, lean hogs, live cattle, cocoa, coffee and sugar. These are the most basic raw materials specifically used by food producers to produce their products within the CRB Index. In the regression, the monthly changes of the price of each commodity (C) as quoted in the CRB Index are used.

(33)

5. RESULTS

5.1 Results for Hypothesis 1

Test of H1: Fluctuations of commodity prices have a significant effect on the value of

national stock market indices.

In order to test hypothesis H1, we regressed the change in oil price on the change in

closing value of the national stock index of our fourteen countries using the following formula: ε β α + ∆ + = t−x mt C R 1 { 1 }

It is assumed that changes in stock value and commodity price do not occur simultaneously and therefore a lagged response time is used to examine the relationship. I test which value of x gives the most significant results for the C values for the commodity oil (see table 5.1). The one-month lag used by Driesprong et.al. (2005) serves as the best lag time to use and this lag time will be used throughout all calculations.

ε β α + ∆ + = t−x mt oil R 1 X=0 X=1 X=2

Sign. Level N Sign. % of Sample N Sign. % of Sample N Sign. % of Sample

0,05 (0,10) 3 (5) 21% (36%) 9 (10) 64% (71%) 1 (1) 7% (7%)

Table 5.1: Using a lag of 1 month when regressing oil on NSI gives the most significant results

The results of regression {1} for all commodities are summarized in table 5.2, displaying the averages of the constant, coefficient (

β

1) and t-values that provide a measure of significance of the proposed relationship. The last two columns show the amount and percentage of the sample in which the change in commodity price (C) provides statistically significant explanatory power for Rmt at the 5% (10%) level of

(34)

x=1 Average N Sign. % of Sample Commodity (C) Constant Coefficient T-Value 0,05 (0,10) 0,05 (0,10)

Copper 0,0081 -0,0486 -0,9756 1 (3) 7% (21%) Industrials Cotton 0,00821 -0,3492 -0,4086 0 (0) 0% (0) Gold 0,0078 -0,0390 -0,4602 1 (1) 7% (7%) Platinum 0,0078 -0,0141 -0,3256 0 (0) 0% (0%) Precious Metals Silver 0,0077 0,0512 1,0913 2 (4) 14% (29%) Crude Oil 0,0091 -0,1042 -2,7713 9 (10) 64% (71%) Heating Oil 0,0088 -0,0646 -2,0809 8 (9) 57% (64%) Energy Natural Gas 0,0079 0,0130 0,4196 0 (1) 0% (7%) Corn 0,0080 0,0110 0,1544 0 (0) 0% (0%) Wheat 0,0087 0,0336 0,6540 1 (2) 7% (14%) Grains Soybeans 0,0080 0,0226 0,3972 0 (0) 0% (0%) Lean Hogs 0,0079 0,0698 1,8178 6 (6) 43% (43%) Meats Live Cattle 0,0083 -0,0702 -0,9595 2 (3) 14% (21%) Orange Juice 0,0084 -0,0422 -1,0131 0 (1) 0% (7%) Cocoa 0,0082 -0,0026 -0,1291 0 (0) 0% (0%) Coffee 0,0081 0,0258 0,7903 0 (1) 0% (7%) Softs Sugar 0,0085 0,0239 0,7043 0 (1) 0% (7%)

Table 5.2: Summary results of Model {1}: Rmt= + Ct-x+

The results show that the effect of commodity prices only has limited significant explanatory power for the values of NSI in the sample. However, crude oil, heating oil and lean hogs price changes show to have a significant effect the national stock indices (NSI) of 14 countries using available data from the CRB Index between 1980 and 2005. In this 25 year sample, 9 and 10 out 14 countries show a significant influence of respectively heating- and crude oil price changes on the value of the NSI at the 10% significance level in those countries. The similar relationship and significance of both types of oil can be accredited to the fact that they are closely correlated (see table 5.3). This correlation serves as an indicator that the price of heating oil is dependent on the price of crude oil causing its price to fluctuate similar to that of crude oil. Similar to the findings of Driesprong et.al. (2005), the results show a negative relationship between oil price change and changes in NSI although the effect is stronger with crude oil. With respect to crude oil, we can say that H1 is rejected for 28.5% of the sample and accept this

(35)

Given these results, a 10% increase in the price of crude oil will cause an average of 1.02% decrease in the value of the NSI for this remaining percentage. Similarly, 64.3% of the sample shows to have a statistically significant exposure to price changes in heating oil. These results imply that a rise in oil – both crude and heating oil – prices substantially lowers future stock market returns and vice versa and therefore the price of oil can be used to predict market fluctuations to some extent.

Crude Oil Heating Oil Natural Gas Lean Hogs

Crude Oil 1

Heating Oil 0,7865 1

Natural Gas 0,2978 0,3691 1

Lean Hogs 0,0806 0,0369 0,0575 1

Table 5.3: Correlation Matrix of price changes of crude oil, heating oil, natural gas and lean hogs

The results of regressing the value of lean hogs are surprisingly irregular and can not be accredited to correlation with the price of crude oil or heating oil (see table 5.3). The results show that 43% of the sample is significantly related to price fluctuations of lean hogs. Contrary to crude oil and heating oil, the relationship between the price of lean hogs and the value of the NSI is positive. This implies that any increase in the price of this commodity results in a positive change in the value of the NSI for 43% of the sample at a 0,10 significance level. One could propose that lean hogs being a meats commodity play a significant role in stock price determination due to a dependency on meats for nutrition.

The possible reasons for the limited effect of the remaining commodities may lie in the fact that many companies of which the stock is quoted in a NSI do not make use commodities but are active in the tertiary or service sector. Furthermore, the commodities that are used by companies in NSI could constitute a minimal factor in the price determination of production. We conclude that commodity prices other than crude oil, heating oil and lean hogs do not have any explanatory power to the value of NSI in our sample. In order to find out whether this power is present for firms specifically using certain commodities, we test hypothesis 2.

(36)

5.2 Results for Hypothesis 2

Test of H2: Commodity price fluctuations have a significant effect on the stock value of

firms that make use of commodities in their production process.

For this hypothesis, a sample of 257 companies in the food producing industry of the US and Japan was analyzed to test for exposure to oil price changes in their stock value using the following regression equation:

ε β α + ∆ + + = i tx mt it C R R { 2 }

The monthly changes in the price of eight individual commodities are used for the value of C. These are: soybeans, wheat, corn, live cattle, live cattle, cocoa, sugar and coffee. From the analysis, seven US firms were defined as outliers and are therefore excluded from the results. For consistency, the value of x is pinned to 1. The results are summarized in the following tables:

Grains

x=1 Soybeans Wheat Corn Statistics Total Japan US Total Japan US Total Japan US

Mean 0,0248 0,0087 0,0456 -0,0184 0,0055 -0,0482 0,0334 0,0231 0,0467 Positive 143 79 64 123 78 45 153 91 62 Negative 107 65 42 124 66 61 95 53 44 Sample Size 250 144 106 250 144 106 250 144 106 Pos. Sign 5% (10%) 4 (10) 2 (5) 2 (5) 6 (11) 5 (9) 1 (2) 4 (15) 2 (7) 2 (8) Neg. Sign 5% (10%) 4 (6) 1 (2) 3 (4) 2 (3) 2 (3) 0 (0) 0 (3) 0 (1) 0 (2) Total Sign. 5% (10%) 8 (16) 3 (7) 5 (9) 8 (14) 7 (12) 1 (2) 4 (18) 2 (8) 2 (10)

(37)

Softs

x=1 Cocoa Sugar Coffee Statistics Total Japan US Total Japan US Total Japan US

Mean -0,0885 -0,0057 -0,2009 -0,0023 0,0315 -0,0482 0,0476 0,0538 0,0245 Positive 110 65 45 135 89 46 106 54 52 Negative 140 79 61 115 55 60 144 90 54 Sample Size 250 144 106 250 144 106 250 144 106 Pos. Sign 5% (10%) 2 (5) 2 (3) 0 (2) 8 (11) 8 (10) 0 (1) 2 (5) 2 (3) 0 (2) Neg. Sign 5% (10%) 8 (13) 5 (10) 3 (3) 2 (7) 0 (2) 2 (5) 8 (13) 5 (10) 3 (3) Total Sign. 5% (10%) 10 (18) 7 (13) 3 (5) 10 (18) 8 (12) 2 (6) 10 (18) 7 (13) 3 (5)

Table 5.5: Summary Results H2 for Softs Commodities – Cocoa, Sugar and Coffee

Live Stock

x=1 Live cattle Leans Hogs Statistics Total Japan US Total Japan US

Average Beta -0,0698 -0,0076 -0,1543 0,0048 -0,0268 0,0477 Positive 101 56 45 101 45 56 Negative 149 88 61 149 99 50 Sample Size 250 144 106 250 144 106 Pos. Sign 5% (10%) 1 (7) 1 (3) 0 (4) 7 (10) 0 (1) 7 (9) Neg. Sign 5% (10%) 8 (14) 4 (10) 4 (5) 6 (19) 5 (13) 1 (6) Total Sign. 5% (10%) 9 (21) 5 (13) 4 (9) 13 (29) 5 (14) 8 (15)

Table 5.6: Summary Results H2 Live Stock Commodities – Live Cattle and Lean Hogs

Crude Oil x=1 Total Japan US Average Beta -0,2191 0,0375 -0,5461 Positive 183 110 73 Negative 67 34 33 Sample Size 250 144 106 Pos. Sign 5% (10%) 22 (40) 13 (24) 9 (16) Neg. Sign 5% (10%) 5 (6) 4 (4) 1 (2) Total Sign. 5% (10%) 27 (46) 17 (28) 10 (18)

Table 5.7: Summary Results H2 Crude Oil

The previous figures show that the effect of price fluctuations of commodities on the valuation of stocks of food producing firms is very limited. Given the limited effect, only the outcomes for results at the 0,10 significance level will be discussed here. The price of oil and lean hogs show to have the highest percentage of exposed firms at 17.9% and 11.6% respectively. The prices of the remaining commodities are significant in less than

Referenties

GERELATEERDE DOCUMENTEN

Auch eine Anwendung zu erzwingen, die Graphenstruktur des Werkzeugs zu verwenden, hilft nicht viel, da die Graphenstruktur des Werkzeugs nicht ausdrucksfähig genug für die

What merits further research then is a social model, one which moves the focus away from the problematizing health related illness critique and sexual violence discourse -

An exogenous fiscal stimulus on the other hand is successful in increasing the share of extrapolating agents while also increasing the level of their expectations, ensuring

The figure shows that this simple constitutive assumption is not valid away form the center of the channel (as discussed in Section IV), since the shear stress and strain rate

ology I use is a two-step regression; firstly, weekly stock return data for the period January 1998 to December 2010 for 78 Dutch non-financial listed firms is regressed against

Cumulative abnormal returns show a very small significant reversal (significant at the 10 per cent level) for the AMS Total Share sample of 0.6 per cent for the post event

In order to make a mechanical model describing the forces and moments acting on the whisker some assumptions are made. 1) The mechanical response can be modelled as a combination of

During the temperature variation parametric study, it was found that as the hot leg pipe wall or containment temperature was increased, the heat absorption rate