THE RESERVES OF EXTRACTED RESOURCES: THE HISTORICAL DATA Julian L. Simon, Guenter Weinrauch, and Stephen Moore ABSTRACT The available data on known reserves of extracted resources have not previously been tabulated and graphed. This article compiles the reserves data into one place for the first time. This serves to a) show the trends in the reserves series, b) make the series available to others, and c) set the reserves data in the context of price data for which longer series have long been available, and with which they can now be seen to be consistent in showing increasing availability rather increasing scarcity. INTRODUCTION It is common sense that as a resource is extracted from the earth, there should be less left to be extracted in the future, and hence that scarcity (by whatever measure) must increase. Dire forecasts of future exhaustion have frequently been made in terms of currently-known reserves. Yet observers with long memories remember that current estimates of known reserves - even in terms of years of consumption, let alone physical quantities - have not declined. Frank Notestein, founder of the Office of Population Studies at Princeton, commented when he was in his 80's during the 1970s that "We've been running out of oil ever since I've been a boy". The main purpose of this article is to present the series of known reserves as far back as available. Seen more broadly, one of the great issues throughout human history has been the relationship of economic growth to availability of extracted resources. Increased scarcity always has seemed to threaten humanity because of the concept of a fixed non-renewable stock being depleted by consumption. Yet economic progress has not ground to a halt because of a lack of minerals and energy. Indeed, these resources have become relatively less important as they have constituted an ever- decreasing proportion of the value of production, and of employ- ment (Barnett and Morse, 1963; Simon, 1981/1994). The measure of economic importance is price. We desire minerals and energy for their productive powers. That is, we want the services they provide; this is their economic value. The economic measure (and definition) of scarcity is the price of one has to pay to obtain the a good and the services it provides - the rate at which it trades against other goods, and especially against human labor. As shown by the price data reproduced here along with the reserves data, the prices of raw materials have been declining historically, both relative to consumer goods and to wages. (For longer price series, see Simon 1981/1994.) Physical scientists, however, tend to be reluctant to rely on trends in prices to forecast future scarcity trends. Rather, they draw our attention to physical quantities and especially to known reserves. Hence we must here address that issue as well. To present those data on reserves is the main task of this paper. Explanation of the observed trends in theoretical and historical contexts is outside the scope of this paper; its contribution is entirely the data on reserves. THE CONCEPTUAL FRAMEWORK Let us conceptualize the matter from the physical point of view, as well as from the economic point of view, in the follow- ing fashion. If resources are indeed becoming more scarce due to declining unrenewable stocks of them, we would expect to see the following: 1. Price would be going up in the long run if scarcity is increasing. This is the point that economists emphasize. Physical scientists sometimes argue that prices are not good indicators of scarcity because of supposed measurement defects - such as the power of a present price to indicate present expecta- tions about future scarcity; economists generally do not agree that these aspects of prices are flaws. 2. Stocks of the resources would be going down, especially the proven reserves which are the basis of the forecasts about running out. Physical scientists emphasize this point, and economists do not disagree. 3. A more meaningful measure than the absolute size of stocks is the ratio of stocks to production. This measure would be declining if scarcity is increasing. 4. Production would be declining, though falling production need not be a bad sign unless price is going up; production could be falling because of substitutes, as in the historical cases of flint and coal. The aim of this article is to present in full the data on these matters. Various chunks of of the evidence have long ago been provided by various writers, thereby laying the intellectual foundation for the subject; particularly noteworthy are Barnett and Morse (1963); Potter and Christy (1962); and the sources of the data used here - Historical Statistics of the United States (1976) together with the annual Statistical Abstract of the United States, and the annual The Minerals Yearbook (plus the range of other sources referred to in the source notes for the tables in the Appendix, and the occasional Commodity Trade and Price Trends produced by the World Bank); also relevant is Simon (1981/1994). But the main body of evidence has not previously been assembled in one place, especially with an eye to coverage of a) data on reserves, b) a comprehensive set of resources, and c) as long historical series as are possible. To present such a compilation is our aim here. THE DATA The tables of data that underlie the graphs require much more space than is possible here. These tables are available in machine-readable form (IBM) and in a printed Appendix from the authors, who hope that others will take advantage of the large amount of scut work that went into compiling this material - especially the data on known reserves, which are the core of this paper. The displays generally speak for themselves. For each resource, the following data are provided in a series of graphs: 1) World known reserves. 2) Known reserves as a ratio with world production. 3) World production, together with prices relative to consumer goods and to wages, each with a different vertical scale. The order of presentation is: All metals in rough order of economic importance - pig iron and iron, copper, bauxite, lead, mercury, zinc, molybdenum. Then coal, oil, and uranium. Then phosphate and sulfur. Figure 1 DISCUSSION Using short time series, and not attending to relevant measures of scarcity, can seriously mislead. Figure 2 shows a forecast for mercury made in 1976 by Cook. He combined a then- recent upturn in prices with the notion that there is a finite amount of mercury on the earth's surface, plus the mathematical charm of plotting a second-degree polynomial with the computer. Figure 1 (exhibits for mercury) shows that the forecast was almost immediately falsified, and price continued its long-run decline. (Figure 1 also shows that reserves of mercury have increased rather than decreased over the years, just the opposite of naive "finitist" ideas). Figure 2 Economists, too, can be influenced by physical considera- tions into focusing on too-short-run price series, consequently making wrong forecasts. For example, in 1982 Slade published an influential analysis of trends in commodity prices based on a theoretical model including grades of ores. Her series ran from 1870 (or later) through 1978. She fitted quadratic concave- upwards curves to the data and concluded that "if scarcity is measured by relative prices, the evidence indicates that nonrenewable natural-resource commodities are becoming scarce" (1982, p. 136). If she were to conduct the same analysis with data running until the 1990s, and using data before 1870 where available, she would surely arrive at quite the opposite conclusion. Meadows (in the famous Limits to Growth by Meadows et. al., 1972) shows how one goes wrong with respect to minerals by using the known-reserves concept for forecasting scarcity. For exam- ple, in 1972 he estimated the world supply of aluminum to be exhausted in a maximum of 49 years. But aluminum is the most abundant metal in the earth's crust, and the chance of its supply becoming an economic problem is nil. (Meadows also made the error of counting only high-grade bauxite, while lower grades are found in much greater abundance). The history of aluminum in Figure 1 shows how aluminum has become vastly more available rather than more scarce since its early development in the 19th century. And in the two decades since Meadows wrote, the price has continued to fall and reserves have continued to rise, both are sure signs that the trend is toward lesser rather than great- er scarcity. CONCLUSION The data on known reserves - even measured in years of consumption - show upward trends over the decades since 1950, the opposite of what "common sense" and Malthusian theory suggest. PERSONAL AFTERWORD It is the judgment of the authors that these data, in combination with the price trends that also show declining trends and together with theory expanded beyond that of Malthus, are consistent with a forecast that all natural resources will continue to become less scarce indefinitely. For the philosophical basis of this forecast, see Simon (1985). page 1/article3 resource/May 15, 1994 REFERENCES Barnett, Harold and Chandler Morse,Scarcity and Growth (Baltimore: Johns Hopkins Press, 1963). Cook, Earl, "Limits to Exploitation of Non-Renewable Resources", Science, 191, 20 Feb, 1976, pp. 677-682 Meadows, Donella H., Dennis L. Meadows, Jorgen Randers, and William W. Behrens III, The limits to growth (New York: Potomac Associates, 1972). Potter, Neal and Francis T. Christy, Jr., Trends in Natural Resource Commodities (Baltimore: Johns Hopkins, 1962). Simon, Julian L., The Ultimate Resource (Princeton: PUP, 1981), 2nd edition forthcoming 1994. Simon, Julian L., "Forecasting the Long-Term Trend of Raw Material Availability." International Journal of Forecasting 1, 1985, pp. 85-93. Slade, Margaret E., "Natural Resources, Population Growth, and Economic Well-Being," in D. Gale Johnson and Ronald D. Lee, Population Growth and Economic Development: Issues and Evidence (Madison, Wisc: U of W Press, 1987), pp. 331-372. page 2/article3 resource/May 15, 1994 APPENDIX TO SIMON, WEINRAUCH, MOORE This appendix contains 1) the tables and the 2) sources from which the graphs in the paper were drawn, plus 3) the graphs drawn to a larger scale and 4) in some cases the separate series plotted separately. It also contains 5) some additional graphs containing longer-run price data; the sources for these graphs are those given for the price series plus some additional sources given in Simon (1981/1984). Also included are 6) data for some resources not mentioned in the article, such as gold. All of this material will be made available upon request in electronic media to readers of the published article. page 3/article3 resource/May 15, 1994