3 rd AGRI-FOOD COMMODITIES: AN INVESTMENT ALTERNATIVE

Transcription

1 3 rd AGRI-FOOD COMMODITIES: AN INVESTMENT ALTERNATIVE OCTOBER 2010 for Third Quarter of 2010 Ned W. Schmidt,CFA,CEBS Agri-Food Value View Version 3/2010/October Year Schmidt Management Company, 2010

2 ABSTRACT Investing in commodities has become a more popular theme over the past decade. To assess the wisdom of such a move, the Agri-Food Price Index was constructed. It was used in a set of tests of returns on that index versus U.S. equities and bonds. A reasonable conclusion of that evaluation is that portfolios void of exposure to Agri-Food commodities are not efficiently modeled. Tests on the returns of the Agri-Food Price Index versus the S&P 500 were conducted. Tests on both the arithmetic mean returns and the geometric mean returns generally found the returns on Agri- Food Price Index to be higher in a statistically significant manner. Correlation coefficients suggest diversification benefits from Agri-Food commodity related investments. Diversification of all equity portfolios into the Agri-Food Price Index improved efficiency in a mean-variance framework up to 60% in Agri-Food Price Index. A price index for the 10-Year U.S. bond was also used in similar tests. Those tests on both the arithmetic mean returns and the geometric mean returns generally found the returns on Agri- Food Price Index to be higher in a statistically significant manner. Correlation coefficient between Agri-Food Price Index and bonds can be assumed to be either zero or negative. Diversification of a bond portfolio into the Agri-Food Price Index improved efficiency in a meanvariance framework up to 45% in Agri-Food Price Index. Agri-Food Commodities: An Investment Alternative Premier independent research on the returns produced by Agri-Food commodity prices. Published quarterly: Jan, Apr, Jul, Oct. Delivered in PDF format only by . Annual subscription link: Version 3/2010/October Page 1 of 55

3 TABLE OF CONTENTS Topic Page Introduction 3 Research Theses 6 Literature Search 9 Data & Testing 17 Agri-Food Vs. Equity Tests Results 21 Agri-Food Vs. Bond Tests Results 37 Conclusions 50 References 52 Appendices 53 The Agri-Food Value View Reviews the current state of Agri-Food investment world. Premier newsletter on the topic of Agri-Food investing. Monthly issues are delivered by . A subscription can be obtained by using this link: Version 3/2010/October Page 2 of 55

4 INTRODUCTION This is the third of our quarterly reports on the topic of Agri-Food commodity price returns. This is not a general interest publication intended to be witty and entertaining. Our intention is to update the findings and conclusions of the previous quarterly report. Some of that which you will read will seem familiar. It should, as the questions have not changed since last quarter. What may change are some of the answers, and that is the important part. Finally, this report is not intended as entertainment. It is intended to be dry and statistical. Emotion and wittiness will be minimized, and for that we apologize. Range of asset classes in which investors have developed an interest seems to have exploded over the past decade. When compared to the acceptable asset classes of past generations that which is considered acceptable for investment today is far broader. The investment world has moved from almost exclusively a narrow list of blue chip equities and bonds acceptable in the 1960s to an environment of almost anything is permissible if used in what might be deemed an appropriate fashion. However, to use an asset class appropriately requires a full understanding of the nature of the asset class. What are the economic drivers? What are the characteristics of the returns produced by the asset class? It is to the goal of answering the second question that this research is directed. In doing so, we must keep in mind that research is much like an onion. Another layer always seems to exist. What follows in not the answer, but just one thin layer of skin. Commodities became part of that list of investments of interest as the 21 st century opened. While considered as a single asset class initially, analysts began to further define the categories within this broad asset class. That commodities is not a homogenous class was soon realized. Commodities is actually comprised of three asset classes with distinctly different characteristics and return drivers. Version 3/2010/October Page 3 of 55

5 Commodities can be subdivided into the following asset classes.! Energy: Oil, natural gas, coal, etc.! Minerals: Iron, copper, gold, etc.! Agri-Food: Soybeans, wheat, corn, etc. Commodities is further separated from other investments by the difficulty of investing directly in the commodity. All have storage issues that may preclude or might make impractical long-term investment directly in a commodity. Deterioration over time of the stored commodity s quality can, in some cases, be a major issue. With regards to this consideration, financial assets or derivatives reflecting the economic value of the commodities may have preferred characteristics for the investor. Investors are primarily interested in the economic benefits of investing in the commodity. For that reason, financial assets that convert the return on a commodity or the return produced by a commodity into usable form are often substituted for direct investment in the commodity. Derivatives were a natural evolution in this investment approach. Traditional derivatives such as already existing commodity futures, as well as newly created ones, have been used as substitutes for direct commodity investing. The financial securities or derivatives used as the primary investment vehicle to gain the economic benefits of commodities produce no economic return in and of themselves. All value produced by these investment proxies is derived from either the underlying commodity, the production of that underlying commodity, or the use of the underlying commodity. For that reason, the beginnings for all analysis of these investments and related instruments must be with the commodities themselves. If the returns produced by the underlying commodity are not attractive, the returns from a derivative based on them will be unattractive. If the underlying returns are not attractive, derivatives will not alter that situation. Leverage, used with some directives, might widen the range of realized returns, but leverage in and of itself provides no return. Version 3/2010/October Page 4 of 55

6 As analysis must be focused, this effort will be limited to Agri-Food commodities. The intention is to, (1), measure the returns produced by Agri-Food commodity prices, (2), compare those returns to those earned in other markets, and, (3), describe the characteristics of Agri-Food commodity price returns. In this issue we will compare the return on Agri-Food commodities to both U.S. equities and U.S. 10-year Treasury bonds. Neither of these choices seems optimal, but comparisons are necessary. The dissimilarities of those two choices should provide a sufficiently wide range of characteristics as to satisfy the information needs of most investors. U.S. equities may not be the best alternative with which to compare Agri-Food returns, but it does have some qualifications for that role. The U.S. equity market remains one of the largest liquid markets in the world. Second, it has the least number of restrictive covenants of the global equity markets. Third, it represents a substantial portion of the portfolio of many investors. Fourth, doing so seems the accepted practice. For comparison to U.S. bonds we will use the return on an index based on the yield to maturity of the 10-year U.S. Treasury bond. Admittedly, that choice is arbitrary and subject to reasonable criticism. As the 10-year Treasury is considered a benchmark for many fixed income investments, it seems a reasonable starting point. Investors with interest in other sectors of the bond market should be able to work with the results built on this benchmark. Version 3/2010/October Page 5 of 55

7 AGR-FOOD VS. EQUITY TESTS RESULTS In the form of wealth indices, historical returns are portrayed in the chart below for both the Agri- Food Price Index and U.S. equities. Three observations are readily made from this picture. First, over the entire period covered by the graph, the performance of the Agri-Food Price Index appears greater than that for U.S. equities. In the tests that follow those returns will be tested for statistical significance 1 to assess the wisdom of this naive conclusion. Our interest, however, extends beyond simply the returns, but also to the character of those returns. What kinds of portfolio returns might be created through their combination? $200 $180 $160 $140 $120 $100 $80 $ $40 Agri-Food Price Index S&P 500 Second, in the first part of the graph the indices seem to move in different, or near opposite, directions. In the more recent period, they seem to move together. What is the correlation between the two indices? Was that correlation stable and reliable? Will the diversification of a portfolio be improved by combining them? 1 Many times the statistics on samples will appear to be different. They indeed may look different. However, we must consider how variable those numbers might be. If the numbers tend to vary considerably over time, the current observations may exist purely due to chance. Statistical testing helps us assess the probability that the observed numbers are different. Version 3/2010/October Page 21 of 55

8 In summary, several conditions have been established. Returns on Agri-Food Price Index and U.S. equities, both arithmetic and geometric, were not statistically different from zero, and these results are probably due to the large variances. Returns on Agri-Food Price Index, both arithmetic and geometric, are more likely to be different from zero than those on U.S. equities. Returns, both arithmetic and geometric, on Agri-Food Price Index were generally greater than those on U.S. stocks. Correlation between the two asset classes is sufficiently low to suggest benefits from diversification, and statistically significant. Agri-Food Price Index returns should be assumed to have the smaller variance, a measure of total investment risk. Next step is to compare the returns and variances of portfolios with different levels of investment in Agri-Food Price Index and stocks. Twenty-one portfolios consisting of the two asset classes were created using the historical return data. At opposite ends of the range of portfolios are one comprising 100% Agri-Food Price Index and one composed of 100% stocks. In creating these portfolios, the level of investment in Agri-Food Price Index was increased in 5% increments and that in stocks was decreased by 5%. For each portfolio, the mean return and variance were calculated. The ratio of the mean return to standard deviation of returns was also then calculated for each of the portfolio. Those results are visually portrayed in the following chart. That portfolio where the slope of this line is the highest is the most efficient portfolio. Version 3/2010/October Page 33 of 55

9 AGRI-FOOD VS. BOND TESTS RESULTS Bonds are today, because of the vast quantities issued by governments to finance their expenditures, a significant holding in the portfolios of many investors. Bonds are also the oldest form of investment due to the reliance of governments down through the ages for financing through debt. Stocks are much younger, as the corporate form of business governance did not develop until the 18 th century. The form a bond takes is that of, one, a promise to repay the original loan, and, two, possibly some form of interest on the loan balance. Neither payment is particularly predicated on the payment of the other. Value of a bond is the sum of the present values of the two different cash flow streams, the principal stream and the income stream. We have previously taken the position that an income stream is tied to Agri-Food commodities, though it is unobservable. That income stream is the additional value that develops when the Agri-Food commodity is processed into something consumable or useful. For that reason and because this is our first contact with bonds, we will be looking at only the value of the principal component of the bond. At a later date we may indeed consider the inclusion of the value of the income stream Agri-Food Price Index Bond Index Version 3/2010/October Page 37 of 55