»z:>^;^>;*^x^^^^>^ ••3$v \-^-> v^v vwy v^v o > "of »". *a « -V K'*«'w V '» W W ^' •'••'^w'*°'''' • ^ , - \3 4*' - u ^^ • •• A V<; f /Vv o > ^o ,4 V ..... *rf>. ^ "• 1 . »i^o».'*. -^ I*' ^ IC 8877 Bureau of Mines Information Circular/1982 SIC-Based Demand Information System for Nonfuel Minerals By George K. Schenck, Balakrishnan K. Nair, and Kung-Lee Wang UNITED STATES DEPARTMENT OF THE INTERIOR Information Circular 8877 SIC-Based Demand Information System for Nonfuel Minerals By George K. Schenck, Balakrishnan K. Nair, and Kung-Lee Wang UNITED STATES DEPARTMENT OF THE INTERIOR James G. Watt, Secretary BUREAU OF MINES Robert C. Horton, Director V This report was prepared by The Pennsylvania State University, Department of Mineral Economics, University Park, Pa., under USBM grant number G0166213. The grant was initiated under the Research and Devel- opment Program. It was administered under the technical direction of the Bureau's Economic Analysis Branch with K. L. Wang acting as Techni- cal Project Officer. Joseph Pettus was the grant administrator for the Bureau of Mines. This report is a summary of the work, completed as a part of this grant during the period February 1 to June 30, 1979. This report was submitted by the authors on June 30, 1979. The research leading to the new data system was initiated by Dr. John Morgan, Chief Staff Officer and former Associate Director — Minerals and Metals Supply/Demand Analysis of the Bureau of Mines. Paul Meadow, former Director, Division of Analytic Studies, actively encouraged the work and made available the resources of his division. The valued cooperation of the commodity specialists for the minerals studied and the staff of the Division of Production/ Consumption Data Collection and Interpretation was most essential to the project. Carl H. Cotterill, Director, Division of Field Operation and former Assistant Director — Metals, Minerals and Materials, strongly supported the imple- mentation of this research and encouraged the Bureau's commodity spe- cialists on his staff to cooperate fully on this project. Dr. William A. Vogely, Head of Pennsylvania State University's Department of Mineral Economics, provided important advice and guidance, and Frederick Demler, Thomas Majcher, William Nagle, and Robert Whelan conducted case studies. Encouragement and active help were provided by many people in the indus- try. Although space does not permit acknowledging each participant by name, the authors wish to thank all representatives of the firms and trade associations listed in appendix A for their assistance. Appen- dix B provides a list of the commodities studied and a guide to the project report in which the end-use survey for each commodity is dis- cussed in detail. CONTENTS Page Preface i Abstract 1 Introduction 2 The new data system 3 The need for the new data system 3 Usefulness of the new data system 4 Elements of the end-use data system 5 The technical product class dimension 5 Criteria for selection of technical product classes 8 Additional criteria for classifying technical product classes 10 End uses — Description of the dimension 13 Distinction between commercial uses and intermediate product markets 13 Perspectives on the market versus product definition 14 Appropriateness of SIC to define end uses 16 Service sectors excluded 18 Two- and three-digit SIC classes preferred 18 Survey forms and methods 19 Design of the survey form 19 Separation of other uses to prevent overaggregation 24 A modified survey form 26 Lead: An example of a new SIC-based survey form 28 A statistical sampling alternative 31 Integrating data from other sources into the SIC-based information system. 31 Implementation of the data system 37 Bibliography 42 Appendix A. — Some of the firms and trade associations consulted in this study.. 45 Appendix B . — References to commodity studies 48 ILLUSTRATIONS 1 . Linkage in the supply and demand for minerals 3 2 . Schematic diagram of fluorine compounds 8 3 . Fluorine compounds consumed in glass and ceramics 9 TABLES 1. Comparison of Bureau of Mines intermediate product classes for cobalt with proposed technical product classes 12 2. TPC's common for ferroalloy group of minerals 13 3. End-use applications of fluorine compounds 15 4. Proposed canvass survey form for zinc compounds 21 5. End uses of zinc compounds 22 6. Questionnaire showing shipments of hot-dip galvanized steel by end-use industries 23 7. High-temperature alloys, end-use survey form 27 8 . Revised lead survey form 29 9. Old lead survey form 30 10. CDA market classification, an example using CDA end-use 310 Automotive, Nonelectric 33 11. Recommended translator for converting CDA data into SIC data base 36 12. TPC end-use matrix for information-critical minerals 39 SIC-BASED DEMAND INFORMATION SYSTEM FOR NONFUEL MINERALS By George K. Schenck, 2 Balakrishnan K. Nair, 3 and Kung-Lee Wang 4 ABSTRACT This Bureau of Mines publication describes and evaluates a Standard Industrial Classification (SIC) based end-use data system for minerals that is designed to link the demand for minerals to economic activity in commercial uses. The new data system is two-dimensional. The first dimension of technical product classes (TPC's) measures the consumption of minerals in major primary intermediate products. The second dimension measures the consumption of these TPC's in SIC end-use sectors in sufficient detail to identify the major determinants of demand for each TPC. Intercommodity comparisons are achieved at the level of SIC two-digit major groups. The methodology was developed so that all available sources of information from Government and industry can be utilized at minimum cost. This information will be supplemented with additional data to be obtained by modifying existing Bureau of Mines canvasses so as to elicit additional quantitative responses from mineral producers. The new data system is expected to improve significantly the abil- ity to forecast demand for minerals, and provide an information base that would, for the first time ever, permit resolution of such questions as trends in substitution among materials and the intensity-of-use of materials in end-use sectors. For the 42 TPC's identified in the report, data are available for the new system from current sources or, except for three TPC's that require new canvasses, can be obtained by modification of existing Bureau canvasses. 'The term "mineral" is used in this report in its broad, generic sense which encom- passes such diverse mineral-based materials as elements, metals alloys, compounds, and commodities. 2 Associate professor, Department of Mineral Economics, Pennsylvania State University, University Park, Pa. ^Research associate, Department of Mineral Economics, Pennsylvania State University, University Park, Pa. 4 Chief, Quantitative Economics, Division of Analytic Studies, Bureau of Mines, Washington, D.C. INTRODUCTION The SIC categorizes establishments by their major economic activities, qual- ified somewhat by product specialization. The choice of SIC to define end-use sec- tors in the Bureau of Mines mineral data system satisfies the primary objective of the data system; namely, to link, the demand for minerals to (1) economic activity and (2) other Government data sources. Public policy is now hampered by a lack of reliable information regarding the role of minerals in the economy, par- ticularly their effect in specific end- use markets. How far is the economy dependent on the availability of miner- als; what sectors of the economy are rel- atively more vulnerable to mineral short- ages than other sectors; which of the minerals are more critical? These are all questions relevant to current public policy as it is addressed to the adequacy of materials for sustained economic growth. The new data base discussed in this report is meant to correct the void that now exists in the information needed for establishing policy to resolve the above issues. Information is not static. Public policy not only demands information but also creates the demand for additional information. This changing need for information for minerals is reflected in the Mineral Availability System (MAS) earlier introduced by the Bureau of Mines as the Bureau sought to improve its information capability regarding domestic resources. This Information Circular introduces a similar effort by the Bureau to improve information regarding demand for minerals. The SIC-based information system being implemented by the Bureau will enhance understanding of the role of minerals in the economy and improve the forcasting of the demand for minerals. Evolving concerns about the preser- vation of environmental quality have raised questions not only about domestic mineral supplies, but also about the appropriate levels of use of minerals. Also the threat of nonprice restrictions such as mineral cartels that interfere with free trade in minerals may pose a threat to the economic welfare and national security of the United States. In order to understand and evaluate these issues, the data base regarding consump- tion of minerals in their numerous end uses should be significantly improved. There is a need to know about mate- rial substitution in mineral-consuming industries. There is also a need to enhance our understanding about the intensity-of-use of materials in various sectors of the economy. Finally, there is an urgent need to improve the quanti- tative base to improve forecasting of the demand for minerals, so that the Govern- ment and industry can anticipate and pro- vide for polices and programs to insure that such demand is appropriately satisfied. The feasibility of the two- dimensional information system presented here and jointly developed by Penn State and the Bureau was first rigorously tested for one commodity, namely sulfur. The test was a success. Then, under re- search contracts from the Bureau, and with the active cooperation of Bureau personnel, Penn State expanded the case studies to include all information- critical minerals. These are defined to be those minerals for which end-use data are relatively vital, either because of dependence on foreign sources of supply or because the existing data base may be inadequate for resource policy decisions. The Bureau is at present actively instituting the new end-use information system for nonfuel minerals. This new data system, when fully in place, is expected significantly to improve both understanding of the sectoral demand for minerals and forecasting of the demand for minerals. THE NEW DATA SYSTEM The new end-use data system is based on a two-dimensional approach using both technical product classes and end-use sectors. The new, two-dimensional system is consistent with the derived nature of the demand for minerals. The first dimension consists of a series of primary intermediate products of minerals that are designated as Technical Product Classes (TPC's). The second dimension consists of an extensive set of end uses defined by criteria established in the well-known Standard Industrial Classifi- cation (SIC) system of the Federal Government. The end-use dimension is itself more accurately characterized as "commercial uses," as distinct from final consumption in the economy. In other words, the new data system measures the consumption of TPC's in such commercial uses as agricul- ture, construction, and manufacturing, but ignores service sectors. The Standard Industrial Classifi- cation categorized establishments by their major economic activities, quali- fied somewhat by product specialization. The choice of SIC to define end-use sec- tors satisfies the primary objective of the data system; namely, to link, the demand for minerals to economic activity and to other Government data sources. 5 The Need for the New Data System Primary minerals rarely find direct application in final consumption sectors, except in the case of industrial miner- als. More often, minerals are used in the manufacture of intermediate products, the demand for which is derived from end- use markets. The demand for ferroalloy minerals, for example, is determined by the demands for steel, high-temperature alloys, and so on in end-use markets such as transportation. 5 Choice of SIC is justified in a later section entitled "Appropriateness of SIC To Define End Uses." Figure la shows schematically the linkage in the supply and demand for min- erals. Figures lb and lc illustrate how the demand for goods in the end-use mar- ket influences the price and output of intermediate products. A shift in the demand for automobiles from Q z to Q^ (fig. lc) calls forth an increase in the supply of steel from Q x to Q x (fig. lb). This, in turn, increases the quantities of iron ore and other ferroalloys demanded in the production of steel (from Q M to Qj,, not shown in fig. 1). The above example illustrates that the market for primary minerals is the intermediate product market and not final consumption sectors. Firms in the mineral-producing industries perceive the demand response in the intermediate prod- uct markets and not demand in final con- sumption sectors. The derived nature of the demand for minerals requires a two-dimensional information system. In terms of an PRIMARY MINERALS M INTERMEDIATE PRODUCTS X MINERAL END-USE MARKETS Z /Sx v s / / s 'x \\ //■■■ kJ \ V ; o £E p x \ X / ~ -¥ fc p v A/'\ /'A' v . 1 k Nd x i 1 A Q x Q' x QUANTITY (b) INTERMEDIATE PRODUCTS^ QUANTITY (c) END-USE MARKETS =Z FIGURE 1. - Linkage in the supply and demand for minerals. end-use data scheme for minerals, this means that the information should sepa- rate (1) the consumption of minerals in the production of intermediate products and (2) the consumption of these inter- mediate products in their end-use mar- kets. Practical considerations in gener- ating data, as well as cost effectiveness of the program, also justify this two- dimensional approach. Usefulness of the New Data System Effective management and rational policymaking require sufficient, reliable data. In the minerals sector, both the formulation of policy and the management of specific programs are to some degree hampered by insufficiently detailed data about the demand for minerals in end-use markets. By explicitly linking the demand for minerals with economic sectors defined by SIC, policymakers and managers in various industries will then be in a position to interrelate the sensitivity of material demand with economic activity. The wealth of census and other information about economic activity can then be used more effectively in forecasting mineral demand. In the few instances (for example, steel, aluminum, brass) where end-use data are now available, the usefulness of such data is limited because the "end- uses" classifications differ among them and often differ widely from those end- use sectors for which economic forecasts are available. The new data system alle- viates this difficulty by providing a common classification system, and does so without sacrificing the wealth of avail- able information. Translators^ are used now to convert existing data into the "These are tables in which non-SIC market classifications are correlated with the coincident SIC end uses. data system. This insures better infor- mation without abridging the traditional role of trade associations and other agencies and provides these data at mini- mum additional cost and without breaching confidentiality. By defining technical product class- es that are common for various minerals and thus, wherever possible, standardiz- ing information for different minerals, it will be possible to enhance intercom- modity comparisons. The new data system also recognizes that the information required for policy planning, for example, may differ in details from that required to forecast the demand for an individual mineral. An accommodation of these different levels of information is made possible by making effective use of the SIC two-digit, three-digit, and four-digit classifica- tions. This hierarchical feature further enhances the usefulness of the informa- tion to various users. Every effort has been made to pre- sent end-use data in a manner useful to industry in its market planning. In the new data system, this is achieved in sev- eral ways. First, where possible, exist- ing trade association data systems are used as sources of input data in their entirety. This enhances the traditional value of such data by integrating them with the new SIC-based system. Second, many industry professionals were con- sulted in determining the scope of each TPC and of the end-use classes (appen- dix A). This consultation provided both quality control and understanding of the needs of private users of the new data. The new SIC-based system provides a basis for individual segments of the mineral industry to enhance their understanding as to how other materials compete or interact with their products in numerous markets. ELEMENTS OF THE END-USE DATA SYSTEM The new end-use data system seeks to link the demand for minerals to economic activity. This linkage is achieved by means of a two-dimensional data system of TPC's and SIC's. The first dimension defines the demand for minerals in the production of technical product classes (TPC's) such as carbon steel. The second dimension mea- sures the consumption of these TPC's in SIC end-use sectors such as motor vehi- cles. SIC end-use sectors are so defined that intercommodity comparisons are pos- sible at SIC two-digit levels, while simultaneously insuring a detailed clas- sification that more accurately captures the demand for each TPC. The Technical Product Class Dimension Technical product classes (TPC's) are those groups of major intermediate products of a mineral for which separate end-use data are desirable, either because the particular TPC enhances understanding of the backward linkage to the demand for a primary mineral such as sulfur or because the TPC materially enhances the quality of end-use data by providing a category that is well under- stood by industry personnel. Symbolic Definition of Technical Product Classes 7 A mineral, M, may go either into the production of intermediate products, Xj, or directly into end uses, Z-. M = f(Xj, Z } ) i = 1, . . . ,n; j = 1, . . . ,m (1) 'This section is important for persons who will be working with the canvass forms or who wish to understand the logic of incorporating TPC's into the new data system. Others may skip this section. Intermediate products, X:, may be used to manufacture other intermediate products, Y | , or may be used in end uses , Z , . f(Z:, Y:) 1, (2) If Xj's are considered as primary inter- mediate products, Y|'s may be regarded as secondary or tertiary intermediate prod- ucts. An example of the primary inter- mediate product, X-, is sulfuric acid, which is made from elemental sulfur or is recovered from other sulfur sources such as smelter gases. Sulfuric acid may then be consumed in fertilizers, Zj, or may be used to manufacture a secondary inter- mediate product, such as aluminum sul- fate, Y|. Aluminum sulfate, in its turn, may be used in an end-use market, Z-, such as paper, or may be used to produce tertiary intermediate products such as sodium aluminum sulfate (alum) and so on. Yj's are, therefore, both secondary and tertiary intermediate products such as aluminum sulfate and alum. Y,'s themselves go into commercial use, Z|. Y ; = f(Z ; ) (3) Demand for a primary mineral, M, in end- use markets, Z., then is M = f D S2) D S1 = g(Z ,) i = 1, ... ,n D S2 = g'(Z,) i = 1, (6) (7) (8) In general, the demand for a mineral, M, in end-use sectors is given by a system of equations for technical product classes , Xj's. M = f(X,, X 2 ,...X m ) Xj = g(Z 1 > Z 2> •z n ) (9) (10) The formulation in equations 9 and 10 requires m canvasses to collect end- use data on these TPC's, and the solution sould involve a system of m+1 equations. Equation 10 does not show the direct con- sumption of M in final consumption sec- tor Z|. This is because, when relevant, the mineral itself is considered as one of the technical product classes for which end-use data are generated. For example, equation 5 is defined to include elemental sulfur, as well as sulfuric acid. One intrinsic advantage of this formulation is that the data base is fur- ther simplified because there are X-'s that might be common for several miner- als, which further minimizes additional canvasses. For example, consider steel, a TPC , which is common for all ferroalloy minerals. The advantage is in the sim- plicity of the concept, and the cost effectiveness of gathering information. One set of information — namely, the end uses of steel — gives end-use information for several minerals such as manganese, chromium, and vanadium. high-speed tool steel in larger propor- tion than other tool steels one would be underestimating the demand for cobalt in the aircraft sector if the demand for cobalt in aircraft was estimated as a, # Z k . Such an error traces to the use of the more general average coefficient instead of a more precise coefficient that is correctly weighted to reflect only the types of tool steel delivered to the aircraft industry. X, = g(Z.) (11) where X 1 is steel. Then, for each min- eral contained in steel, M. M, = f(a,Z,) (12) X where a t = — -*-. the contained mineral in 1 M j ' a unit of X,. Equation 12 assumes a direct proportional relationship between M contained in X, and M ; consumed in Z . . This will be true only if X 1 is a homo- geneous product. If X, is not a homo- geneous product but a group of related products, then the coefficient, a., is an average for the group of products as a whole, I|, and equation 12 may only approximate the consumption of M. in Z . . There may be some Zj's in which an alloy with a higher unit content of M| than the average a- may be selectively demanded or vice versa. Therefore, although the average coefficient, a f , may be a good estimate of M f in the produc- tion of X|, it may not be a good estimate for purposes of equation 12, which is the objective of the end-use data collection system. For example, say cobalt is the i +h mineral with X, being tool steel. The average coefficient, say, a,, which describes the ratio of cobalt contained in a unit of tool steel, is lower than the average contained cobalt, say, I 2 , in high-speed tool steel. However, the can- vass on tool steel provides data only on the consumption of tool steel in end-use sectors, Zj's. If one end-use sector, Z k (say, aircraft) selectively demands The problem illustrated in the above example can be eliminated only if end uses for each separate cobalt-containing steel alloy are known. Such detailed information is unavailable for a variety of reasons. Hence, the best that can be expected is to use the methodology where TPC's are carefully designed to get the average coefficient, a-, as close as pos- sible to the actual coefficient, a-, so that a | approximates the expected value of a | . In the illustrative case of cobalt, whereas a , if measured as an average coefficient for all steel could create a serious flaw, a 1 measured as an average for tool steel would be close to what can reasonably be expected from a commodity survey. Although using such average coefficients is a theoretical shortcoming in the two-dimensional end-use data scheme proposed in this study, no means were found during this study by which this problem could be entirely corrected. For example, it can be readily seen that high-speed tool steel is not a homo- geneous product either, and the argument to separate high-speed tool steels from tool steel can soon be applied to each of the many high-speed tool steels manufac- tured, but this leads to an explosion of data. The questions relevant to such a data explosion are can data be generated in such great detail, and are the mar- ginal benefits so obtained commensurate with the cost of generating and process- ing such data. The answers to both ques- tions are negative. Also, the greater the number of TPC's, the more narrowly each TPC is defined and the more likely that the canvass will ask for information considered by firms as proprietary and lead to diminished cooperation. Criteria for Selection of Technical Product Classes Technical product classes for a min- eral are those primary intermediate prod- ucts that adequately capture the demand information for minerals in relation to end-use markets. A TPC performs a dual function in the information system by providing a backward linkage to demand for primary minerals, while satisfying the forward linkage to end-use sectors. The selection of a TPC cannot be based upon a unique criterion. Rather, each TPC should be selected so that it represents the minimum number of primary intermediate products that together provide an information base adequate to quantify the flow of minerals in the economy. The selection of a TPC for a mineral is subjective, and the TPC's for a particular mineral are not fixed over time. Technical product classes may change as technology changes. The cri- teria for selection of technical product classes are best illustrated with an example. The case of fluorine is dis- cussed below. Figure 2 shows the three TPC's selected for fluorine — fluorspar, hydro- fluoric acid (HF), and fluorsilicic acid — drawn as diamond-shaped boxes. Rectangular boxes in figure 2 represent all major intermediate products derived from hydrofluoric acid or fluosilicic acid. Note that these intermediate products themselves may go into mineral SILICON TETRA- _ FLUORIDE (byproduct) (4) FLUORIDES (5) FLUORINE (7) SODIUM FLUORIDE A3L FLUO- SILICATES (6) FLUORO- CARBONS (9) CRYOLITE (10) ALUMINUM FLUORIDE FIGURE 2. - Schematic diagram of fluorine compounds. I i M I N E R A L E N D U S E M A R K E T S end-use markets or may go into the pro- duction of other intermediate compounds. For example, fluorides (manufactured from HF) may go into mineral end-use markets or may go into the production of other intermediate compounds. For example, fluorides (manufactured from HF) may go into final consumption (for example, water fluoridation, glass, and ceramics) or may be used as fluorinating agents that produce additional intermediate products such as organic compounds of fluorine (f luorocarbons) . The numbered boxes in figure 2 indi- cate that in one possible, hypothetical, detailed scheme, there are 10 possible technical product classes for fluorine that would require an equal number of questionnaires to elicit data from firms in the industry. Since more than 7 5 per- cent of hydrofluoric acid is consumed at captive plants, it also implies that a firm that manufactures HF would be the recipient of questionnaires 2, 4, 5, and 6. Similarly, a producer of byproduct fluosilicic acid could receive up to five questionnaires (3, 5, 8, 9, and 10). This complexity obviously would be unac- ceptable to industry. One of the end-use markets for flu- orine is in glass and ceramics, shown in figure 3. This sector uses various fluorine compounds for a variety of applications, such as providing trans- parency (fluorspar, cryolite) , etching and polishing (hydrofluoric acid) , MAGNESIUM FLUORIDE AMMONIUM BIFLUORIDE POTASSIUM BIFLUORIDE SODIUM FLUORIDE GLASS AND CERAMICS SECTOR ALUMINUM FLUORIDE SODIUM FLUO- SILICATE FIGURE 3. - Fluorine compounds consumed in glass and ceramics. 10 enamels and glazes (lithium fluoride) , frosting (potassium bifluoride) , and imparting opalescence (sodium fluosili- cate) . If the analyst knew only consump- tion of all fluorine in glass and ceram- ics, it might not be sufficient because this demand is satisfied using three dif- ferent sources (TPC's) of fluorine units. First, there is the consumption of fluorspar. Second, there is consumption as hydrofluoric acid and its compounds produced from acid-grade fluorspar, and third, there are products obtained from fluosilicic acid, which is a new source of supply of fluorine and is a byproduct of phosphate processing. There is also merit in knowing consumption of hydro- fluoric acid and products made from it versus consumption of fluosilicic acid and its compounds, because such a data base would indicate the substitution of fluorine compounds from two distinct and separate primary sources of fluorine. acid. Efficiency and cost dictate that the number of TPC's be limited in this case to three, the minimum required to satisfy the data objective. In summary, the approach by the Bureau is to collect end-use data for a limited number of TPC's for each mineral. These TPC's are major intermediate prod- ucts or groups of closely related inter- mediate products. The objective in selecting TPC's for a mineral is to limit them to the minimum number required to quantify the final consumption character- istics of a mineral subject to the avail- ability of data. Appendix B provides a list of the TPC's studied and a guide to the project report in which the sources for end-use data for each commodity are discussed in detail. Additional Criteria for Classifying Technical Product Classes The classification in figure 3 of fluorine into three TPC's (see diamond- shaped boxes for fluorspar, hydrofluoric acid, and fluosilicic acid) is amply jus- tified. But to classify as separate TPC's each of the eight compounds out- lined in figure 3 as rectangles (cryo- lite, sodium fluoride, etc.) would be unwarranted because such a data explosion would not be justified by a corresponding increase in usefulness of the data collected. What should be expected from an end- use data base is a report of the consump- tion of fluorine in end uses such as "Glass and Ceramics," aggregated in a way that would permit estimation of the supply requirements of fluorine given a forecasted level of activity in the glass and ceramics sector. This linkage to supply is amply satisfied if data can be collected for consumption of ceramic- grade fluorspar, acid-grade fluorspar (hydrofluoric acid) , and fluosilicic acid, the diamond-shaped boxes of fig- ure 3. The information base would not be strenghthened significantly if the con- sumption was determined separately for each of the compounds derived from either or both hydrofluoric acid and fluosilicic An important consideration in defin- ing a TPC for a mineral is that it should be limited either to one product or to a group of closely related products. In this manner consumption of mineral in that TPC can be determined with a degree of confidence. For a mineral such as iron ore, one TPC, namely steel, would be amply suffi- cient. However, steel is relevant not only to iron ore but also to all other ferroalloy minerals. Therefore, end-use data on steel will be provided by the Bureau in such a way as to give best results not only for iron ore but also for other ferroalloy minerals. In the Bureau of Mines' new end-use information system, steel is separated into six TPC's: Carbon steel, alloy steel, tool steel, stainless steel, heat-resistant steel, and high-strength, low-alloy (HSLA) steel. Demand for chromium, for example, is more closely related to the demand for stainless steel than to any other type of steel. Similarly, for molybdenum it is the consumption of tool steel that is most important, and for cobalt it is tool steel and perhaps specialty steels 11 (high-temperature alloys) . Therefore, even though at first steel is thought to be a single TPC for iron ore, once it is decided to generate end-use data on steel, it is wise to consider the several steel types as intermediate products com- mon for several alloying minerals. The basis for the definition and selection of a particular TPC is as fol- lows: (1) A link backward to supply, (2) a link with historical data, (3) a clas- sification common to several minerals, and (4) a minimum number of TPC's. Therefore, the task of selecting the appropriate number of TPC's for a primary mineral is one of finding the set of intermediate products that best satisfies all four criteria. For steel, the first two of these criteria are met by making maximum use of the existing data base of the Bureau. When the Bureau collects data on intermediate products it was determined (1) how far these satisfied other desir- able criteria and (2) whether these prod- ucts could be regrouped in a manner by which historical continuity (with the preexisting data scheme) was preserved while minimizing the number of TPC's. Table 1 compares the intermediate products for which the Bureau at present collects data with the seven proposed TPC's for the cobalt end uses. The seven letters (a to g) indicate the grouping of the Bureau's present list of intermediate products within the seven TPC's. For example, the Bureau's category Tool Steel (d) would be retained as an independent TPC, while all the other steel categories in the present Bureau classification would be grouped together under one TPC, Other Steels (e). This concept preserves the link between the new end-use data collection scheme and the existing data base of the Bureau. Column 1 of table 1 lists seven TPC's for cobalt (a to g) . This may seem to be a large number of TPC's for one mineral. However, for a mineral such as cobalt that finds diverse specialized applications, it appears unavoidable. One of the basic criteria for defining a TPC is that it will be a group of closely related products, so that reasonable assumptions can be made about contained mineral (in this case cobalt) in that TPC. The six metallic product classes for cobalt are common to a host of ferroalloy minerals, such as nickel, chromium, co- lumbium, tantalum, and tungsten (ta- ble 2). In other words, in the new sys- tem a TPC such as high-temperature alloys is not exclusive to cobalt. It is also a TPC for the five other ferroalloys shown in table 2. Therefore, across the spec- trum of ferroalloy minerals, the number of TPC's can still be kept to a manage- able level. Penn State has suggested a total of 42 TPC's to collect end-use data on 23 minerals. The test for an additional TPC is not whether one more TPC is desirable, but whether the objective of the data collection system can be met without it. 12 3 O O >-i O u 3 a S o cj PQ P. CO H o CO CL| CJ H ■H ^> s ,3 CO O cu 0) CO ^J CO cfl T3 rH ai CJ CO o a. o u PM >> >> o o rH rH r-l rH cfl CO u CO CJ T3 >. X 3 aj co H w —I CO CO CO rH M ts CU 01 O 0) CX rH 4-> 3^ co X* c CO O CO CO rH T) >. rH O O 3 r4 rH h -d M CO 3 3 CO -H CJ CJ •H bo CO 4-1 3 4H 0) vH X) CO H M rH CO eg 0) J= cfl a> Ol cfl S cfl co cfl 3 cfl bO J-i 3 CO •H -H 4-1 CO 4-1 0) 3 M >-. CO O 4-1 01 u 4H 01 C r« rH 3 Cfl T3 O M a. u 01 O xl to a a o rH .. X> iH Cfl rH J-I Cfl 4-1 0) Cfl 4-1 0) 01 CJ CO a 4J 00 C rH •H 0) cfl 0) 4-1 4-1 CO CO CO CO 0) r-t Xl O 3 O r^ U bO O 3 X "3 •H 111 fi 4-i — ccj CO •H >, CO CO O CO 01 "H 01 l-l rH rH < S cfl u CD CJ XI a CO CO cfl o CJ CU ■3 -a c CO CO 3 -3 CO I 4-1 CJ 3 3 -rH rJ 3 a, u cu 0) J3 CO -I CO cu 3 x: CJ CJ Oh CO 4-1 -3 CO C to s_ • • >, 3 CO 3 CO rH O 0) 3 Sh CO 4-1 CJ CJ O 3 3 CJ 3 J3 4-1 u 01 -3 O) iH 3 4-1-3 3 -H 3 rH O 3 rH 3 11 am CJ CO 4-1 CO 3 rH ■H 3 3 S w r~^ co o i — iD m ci — i oo vc co rH r^ CN rH rH 3 X> O xl 3 4H bO CO o cfl D o to a -3 U CO 3 3 >, J= u O 3 rH -3 in rH 3 3 3 3 B cj co 01 tH 0) 4J 4-1 TJ I 3 -H 3 C XI bO bO 1h 3 3 >, • O • <-t CO rH "3 3 3 D xi o 3 a, 3 a o 0) 3 a X. O U CJ SSUH CO 01 rH .3 01 4J .. 0) O O 4-1 I -H CO I rH 4-1 ^-{ IH rH 3 01 3 4-1 X XI 3 4-1 o a O CJ C O Z TABLE 2 . - TPC's common for ferroalloy group of minerals 1 13 Steel (e) (a) High- (b) Mag- (c) Mineral High Heat Carbides commodities Carbon strength, (d) Alloy Stain- resist- temperature netic and hard- low alloy Tool less ing alloys alloys facing alloys Chromium. . . X X X X X X X X Cobalt 2 . . .. X X X X X X X Columbium. . X X X X X X X X Nickel X X X X X X X X Tantalum. . . X X X X X X X Tungsten. . . X X X X Excludes manganese, molybdenum, tellurium, vanadium, etc. , which were not part of the present study. 2 Cobalt uses are coded the same way as in table 1. For example (d) means the com- modity currently is found in the tool steel TPC. End Uses — Description of the Dimension The end-use sectors of a mineral commodity provide the necessary second dimension to a fully specified mineral information system, the first dimension being the technical product classes (TPC's) discussed above. To define end uses the Bureau has chosen the SIC system. End use — for purposes of the new data system — is best defined as that use at which a TPC is delivered in the U.S. economy to a final commercial use, such as manufacturing, agriculture, or construction. SIC-based commercial uses are defined for each TPC such that the TPC is used in a consumer good (steel in automobiles) , or becomes untraceable fur- ther (sulfuric acid in the making of paper) . A mineral commodity may pass through several stages of manufacturing or fabri- cation before its final commercial appli- cation. In such cases, it may not be clear which point of consumption repre- sents final commercial use. This problem is compensated for by selecting the appropriate number of mutually exclusive TPC's for a mineral and collecting end- use data for each TPC. Data availability may further restrict the ability to obtain data in terms of commercial uses. Therefore, intermediate product uses such as "paints" may have to be retained as end uses , although conceptually more appro- priate commercial uses could be conceived (for example, automobiles, construction). Such exceptions are infrequent and are made only when respondents (during prep- aration of the questionnaire) indicated that they could not provide greater detail. Distinction Between Commercial Uses and Intermediate Product Markets Firms generally wish to have infor- mation about specific product markets in which their own products are used. These we have defined as tertiary intermedi - ate products . This desire of producers to identify product markets often creates a conflict with the economic analyst who wishes to select commercial uses rele- vant to measuring overall demand for minerals. Fluorine provides a good example to illustrate the disadvantage of defining intermediate products as "end uses" of minerals. Earlier, fluorine was used as an illustration to discriminate between primary and tertiary intermediate prod- ucts in making proper choice of TPC's. The same commodity is used as an example in this section to discriminate between tertiary intermediate products and 14 commercial uses for proper determination of end uses that measure demand on a national scale. Table 3 lists some of the major flu- orine compounds and the commercial uses in which these compounds are consumed. A product definition of the markets for fluorine is identical with column 3 of table 3, "Fluorine compounds." Use of such an end-use definition would provide information on the consumption of fluor- spar (and presumably hydrofluoric and fluosilicic acid) in the production of various compounds listed, such as alumi- num fluoride or various f luosilicates . But, it can be readily seen that (1) sev- eral fluorine compounds are demanded in the same end-use market, and (2) the same compound (for example, aluminum fluoride) is demanded in different end-use markets. Therefore, it can be recognized that a product-based definition is clearly inap- propriate to describe the end-use markets for fluorine. Column 2 of table 3, "End-use titles," on the other hand lists end-use markets for fluorine. Information based on these end-use markets will provide a basis for understanding the demand for all fluorine products in a well-defined manufacturing activity. Perspectives on the Market Versus Product Definition The principal reason for the per- sistence of the conflict between market definitions and product definitions is the underlying conflict between the need for product data by firms in the indus- try, and the market data an economic analyst needs to understand demand for minerals in the economy. Both are useful, but the two are not always coincident. In the methodology developed for this study and described in the earlier section, "Symbolic Definition of Techni- cal Product Classes," the "secondary" and "tertiary" intermediate products are ignored (Y ( in equation 2). It is assumed that information about the consumption of "primary" intermediate products in end-use markets (Z f ) is suf- ficient to satisfy the general objective of an information system that traces the flow of materials in the economy. Both practical considerations in collecting and processing the information and the cost-ef f ectiveness of the program justify the proposed methodology. However, the firms that produce the primary mineral products (X ; ) often sell them in the market for secondary and ter- tiary intermediate products (Y|). Because the primary producers (X,) sell to the intermediate market (Yj), they may not be concerned with the final end use; that is, they may not know their cus- tomers' markets (Zj). When there are several producers of primary products, the concern of each of them is to increase its share in the market for intermediate products. A producer of fluosilicic acid, for example, will be concerned primarily about a market for fluosilicic acid in the production of synthetic cryolite (Y,). Whether the customer uses its production of synthetic cryolite in producing aluminum (Zj), or sells it to ceramics manufacturers (A,), is usually of minor importance to the acid producer. Thus the level of information desired by firms in the industry may dif- fer considerably from the objective of the Government in initiating an end-use survey. Having recognized this differ- ence, the relevant questions then are 1. Whether the Bureau of Mines can generate end-use data in a step-by-step sequential manner, tracing consumption of a mineral through the several intermedi- ate uses to a final commercial use, and 2. If such a sequential canvass is not possible, whether the global data the Bureau needs will help the industry suf- ficiently in its business decisions to provide the necessary incentive for industry to participate in an end-use data scheme. 15 TABLE 3 . - End-use applications of fluorine compounds SIC code End-use title Fluorine compounds 13 20 22 281 282 283 284 286 2801 291 321 324 325 326 329 331 332 3334 3301 3471 358 361 363 3601 376 38 Oil and gas extraction Food and kindred products Textiles Chemicals, and allied products: Industrial inorganic chemicals Plastics, materials, and synthetic resins rubber, and man-made fibres. Drugs Soaps, cleaners, and toilet goods Industrial organic chemicals Other chemicals Hydrofluoric acid, lithium fluoride Stone, clay, and glass products: Flat glass Cement Structural clay products Pottery and related products Abrasives and miscellaneous nonmetallic products. Primary metal industries: Steel mills Iron and steel foundries Primary aluminum Other primary metal industries Plating and polishing Refrigeration and service machinery Electrical and electronic equipment: Electric distributing equipment Household appliances Other electrical and electronic equipment. Guided missiles and space vehicles, and parts. Instruments and related products. Water treatment Chlorine trifluoride, ammonium bifluoride, hydro- fluoric acid, silicon tetraf luoride. Ammonium bifluoride, fluosilicic acid, copper flu- osilicates , sodium fluoride, perf luor os yclo butane. Boron trifluoride, fluoborates, perchlorof luoro acetones. Fluorspar, magnesium fluoride, potassium biflu- oride, hydrofluoric acid, fluosilicic acid. Hydrofluoric acid, aliphatic chlorofluoro hydro- carbons, poly (tetraf luorethylene) (PTFE). Perchloryl fluoride, silver f luosilicate, per- chlorof luoro acetones, fluorinated aromatic compounds . Fluosilicates of calcium, magnesium, and zinc; sodium bifluoride; sodium monofluoro phosphates; stannous fluoride. Fluoborates, hydrofluoric acid, potassium fluoride, etc. Aluminum fluoride, boron trifluoride, hydro- fluoric acid, etc. Aluminum fluoride, cryolite, ammonium bifluoride, fluorspar, hydrofluoric acid, lead fluoride, lithium fluoride, etc. Fluorspar, silicon tetraf luoride, zinc f luosilicate. Aluminum fluoride, lithium fluoride. Aluminum fluoride, cryolite, hydrofluoric acid, lithium fluoride, sodium fluoride, magnesium fluoride. Cryolite, fluoborates, fluorspar, f luorozirconates. Fluorspar, hydrofluoric acid. Cryolite, fluorspar, hydrofluoric acid, sodium fluoride. Aluminum fluoride, cryolite, fluoborates, hydro- fluoric acid, fluosilicic acid. Aluminum fluoride, fluoborates, hydrofluoric acid, f luorozirconates , PTFE, lithium fluoride, sodium fluoride, etc. Fluoboric acid, fluoborates, fluosilicic acid, sodium bifluoride, zinc tetraf luorosilicate. Sulfur hexaf luoride, carbon tetraf luoride, carbon tetrafluoride, CC1 2 F 2 , CC1 3 F, CBrF 3 , etc. Sulfur hexaf luoride, perchloryl fluoride, perf luoro eye lobutane. PTFE, carbon tetrafluoride, CHC1F 2 . Fluoborates, PTFE, Poly (chlorotrif luorethylene) (PCTFE). Nitrogen trifluoride, perchloryl fluoride, sul- fur hexaf luoride, fluoborates, PTFE, carbon tetrafluoride, CBrF 2 . Lithium fluoride, magnesium fluoride, fluorspar. Fluorsilicic acid, fluosilicates, cryolite. 16 The answer to the first question is clearly in the negative. The costs of collecting and collating data in such detail would be excessive. Firms that perhaps justifiably complain about their own escalating costs in the "unproduc- tive" exercise of providing a multi- plicity of data to various Government agencies also will not be indifferent to their share of the costs. Also, the more detail in which data are collected, the more it will encroach on proprietary information of various firms in the industry, a situation that neither the industry nor the Government considers advisable. Therefore, an end-use data collec- tion system sponsored by a Government agency must, of practical necessity, be such as to insure (1) a least-cost data base adequate for policy decisions and (2) the least expenditure of reporting time by industries as a whole. The ques- tion then is whether such a data base will give the industries a return commen- surate with their voluntary participation in collecting and reporting such data. Reporting efficiency is a significant consideration in designing mineral end- use surveys. It is important here to recall that the demand for minerals is derived demand. Therefore, for example, if the Government is in a position to estimate what impacts its policy regarding the construction market may have on the demand for mineral commodities, this will provide several benefits to mineral pro- ducers. The direct benefit is in having such information to guide business deci- sions. An indirect benefit will accrue if such information will help improve Government decisions concerning its pol- icies toward that sector or industry. For another, the data base itself will be valuable for market planning by firms. It is possible that a few large firms may already have such information and hence may apparently not be benefited by it. But even in such instances, private esti- mates are no substitute for industrywide data. The existence of trade associ- ations and their role in collecting statistics attest to the value of indus- try data. Similarly, trade associations themselves will benefit from SIC-based end-use information because it provides a comparable data base between competing industries and hence better interindustry data for association members. The new end-use data base described here will not substitute for a firm's own market research needs. A firm's "need to know" regarding the markets for its prod- ucts cannot be met fully by a service from the Government. What the Bureau is providing with the end-use data is a better information base for intraindustry and interindustry comparisons within which individual firms can pursue their own marketing interests. APPROPRIATENESS OF SIC TO DEFINE END USES Standard Industrial Classification (SIC) is the most widely understood and widely used "classification of establish- ments" to categorize the U.S. industrial structure. Therefore, the Bureau's objective of linking demand information for minerals with domestic economic activity is best achieved by defining end uses on the basis of SIC. However, the appropriateness of SIC to define end uses of minerals is not without its problems. There are questions about whether SIC "establishments" are appropriate to define end-use categories. There are also questions about the appropriate detail of SIC groupings, and there are doubts about the ability of the various industries to report end-use data on the basis of SIC. The first two questions are not entirely independent of each other. How- ever, in order to separate the method- ological problem from the choice of detail, these two questions are discussed separately. SIC classifies establishments based upon criteria such that establishments in 17 the industry (1) consist largely of those goods and services defining an industry, and (2) account for the bulk of the spec - ified goods and services provided by establishments in that industry. The underlined words emphasize the major criticism of using SIC to define end-use markets for minerals in terms of goods produced. It was a conclusion of studies at Penn State that such definitional prob- lems are not unique to SIC and would exist in any classification scheme. This can be readily illustrated. Those firms that participated in the present study agreed that "construction" is an appro- priate end-use category for mineral com- modities. But not all industries that serve construction markets would agree to have one end-use category, "construc- tion." It was generally believed that "building construction" should be treated as a separate subcategory. Those indus- tries that serve a specific market in building construction, such as "builders' hardware," wished to have that specific market identified as a second-order sub- category, and even then there was no agreement on how the category "builders' hardware" should be defined. There were those who insisted that furniture hard- ware is part of the building construction market and should therefore be classi- fied as "builders' hardware"; others disagreed. The above example illustrates that even in the case of a universally accepted category such as "construction," there was no agreement as to how many subcategories are adequate, much less as to the proper definition of the subcat- egories. In other words, definitional problems are inherent in any classifica- tion scheme, and using an SIC-based classification of "construction" is no worse than using any alternative scheme. The second challenge to use of SIC regarded choice of the appropriate level of the SIC system, which is based on a hierarchy of several levels of classifi- cation. SIC major groups are two-digit. These are further classified into three-digit groups that may in turn be classified into four-digit industries. 8 For purposes of collecting statistics, the Department of Commerce has further elaborated upon SIC to define five-digit product classes and seven-digit products in order to specify products manufactured in SIC establishments. It is necessary to distinguish the five- and seven-digit product classifica- tion used by the Deparment of Commerce from the Standard Industrial Classifica- tion of economic establishments that goes no further than SIC four-digit indus- tries. The five- and seven-digit product classification is entirely inappropriate for classifying end uses of minerals. The new data system of the Bureau is con- fined to the two- , three- , and four- digit industries defined in the Standard Industrial Classification. The new data system of the Bureau makes use of the hierarchical structure of the SIC. Intercommodity comparison is achieved at the SIC two-digit level. Demand information for each technical product class is obtained for all rele- vant SIC two-digit major groups and three-digit groups. In those instances where a particular four-digit industry represents a major end-use market, that industry is classified as a separate end use. The third and last question regard- ing the appropriateness of the use of SIC to classify end uses of minerals relates to the ability of firms in the industry to report data. This ability differs from one segment of the mineral industry to another and among different firms. Therefore, for each TPC , a separate 8 SIC 3 361 provides an example of the con- cept of the four-digit hierarchy. The first two digits, 33, provide the code of a major group, "Primary Metal In- dustries". The third digit, 6, shows it is the Nonferrous Foundry group within the major group, and the fourth digit in 3361 shows it precisely as the Aluminum Foundry portion of the Nonferrous Foundry group. 18 determination was made in consultation with industry representatives to assure that firms in the industry are indeed capable of reporting the required end-use information. The reporting burden on firms is also further reduced consider- ably by making use of translators to con- vert into the SIC end-use format data that are already available from existing reports from the Government or trade associations. Service Sectors Excluded For purposes of the new data system, the SIC end-use classification is limited to commercial uses in manufacturing. Service sectors are excluded. This is done because 1. In the overwhelming majority of cases, manufacturing is the appropriate end-use sector for minerals. 2. Consumption in manufactured goods, such as consumer durables, can be considered as the final market because over the life of a good, it is not avail- able for reuse. 3. Recycling of manufactured goods is a source of secondary supply for sev- eral minerals commodities. If the manu- factured goods generate secondary supply, that fact provides a good test that the buyers of those goods should be con- sidered as end users. 4. Even in those instances where a service sector may appear to be highly relevant as an end-use category (for example, the electrical utility market for copper) , it is possible to specify an appropriate SIC manufacturing sector without any loss of information. There- fore, for purposes of end-use data, it is unimportant whether a manufacturing sec- tor or a service sector is specified, except for consistency in practice. 5. More often, in place of one well-defined manufacturing sector, it would be necessary to specify several service sectors that are not mutually ex- clusive, thereby increasing definitional problems, as well as creating the poten- tial for inconsistent reporting of data. 6. The intensity-of-use of the materials and substitution between mate- rials are relevant only in terms of manu- facturing sectors. An information system that identifies these phenomena is cru- cial to determining the nature of the demand for minerals and in assisting in forecasting demand for minerals, a major objective of the new data system. Two- and Three-Digit SIC Classes Preferred One objective of the new end-use reports is to improve the data base for forecasting demand for primary minerals, given activity levels in end-use markets. This implies that the end uses specified for mineral commodities should be those (1) that are important for estimation of demand and (2) for which economic fore- casts are readily available. Both of these goals are best attained through use of a limited number of mutually exclusive end uses. It would also be easier for the firms responding to Bureau canvasses if the end uses listed on a survey form are held to the minimum number necessary to capture the significant consumption patterns for that commodity. The basic requirements of the end- use information system are best served if the proposed SIC end uses are limited to two-digit and three-digit groups, with exceptions provided when necessary. This can be done in an SIC end-use classifica- tion because SIC is structured so that four-digit industries are grouped within three-digit groups, which in turn are grouped within a two-digit major group. Hence, if the analyst specifies a three- digit group instead of a four-digit industry, or a two-digit major group instead of a three-digit group, there is no loss in distinction among groupings despite a loss of detail in a particular group. Each specified end use should be sufficiently descriptive of a particular major end use of the mineral commodity. In general, as proved in several case studies, this distinction is achieved at 19 the SIC three-digit level. For cases where total consumption at any two-digit level is a minor but reportable part of the total consumption, distinction may be attained at the two-digit level. As a general rule, limiting end-use classification to SIC two- or three-digit groups also helps to minimize some of the common problems associated with the end- use classification itself or in reporting data. It is certainly easier for a min- eral producer to report his shipments to the two-digit group, SIC 25, Furniture and Fixtures, than, for example, if he were asked to specify end uses in three- digit groups, SIC 251, Household Furni- ture, and SIC 255, Office Furniture. By the same token, it would be many times more difficult if the respondents were asked to specify four-digit industries such as Metal Household Furniture (SIC 2514) and Mattress and Bedsprings (SIC 2515). Even if a mineral producer has the ability to report shipments to such four- digit industries as SIC 2514 or SIC 2515, that does not imply that data reported at such levels would be more accurate that data reported only at the corresponding three-digit level. In fact, the converse may be true. At the more aggregated lev- els (smaller SIC numbers), customer iden- tification becomes more clear and hence may result in more consistent reporting among respondents. One advantage of SIC is its hierarchical classification system which permits different firms with dif- fering abilities to report data at the level most convenient to each, without losing comparability for data among dif- ferent reporting firms. By collecting data at higher levels of aggregation (SIC two- and three-digit groups) , there is a reduction in data incompatibility among reporting firms. A producer may combine into SIC 3519 its shipments to Marine Engines (SIC 3519) and also to Ship Building and Marine Equipment (SIC 373). This error could then invalidate the data for SIC 373 received from this respondent. However, if SIC 37, Transportation Equipment, is considered, the error at the two-digit level caused by the producer's reporting procedure would be negligible, because shipments to marine engines then becomes such a small percentage of total ship- ments to all transportation equipment. In general, data tend to become more reliable at higher levels of aggregation (smaller SIC numbers). As a final justification of higher aggregation, we note that some firms have indicated that in certain cases if data are reported at the four-digit level, it may amount to divulging their customers' identity, and hence, they wish to report data only at three- or two-digit levels. Therefore, by specifying two- or three- digit groups, problems of confidentiality also tend to be minimized. SURVEY FORMS AND METHODS Identification of the TPC dimension and the end-use dimension is only the necessary first step in developing sta- tistics about consumption of minerals in commercial uses. It is next necessary to integrate these principles into a con- venient format to gather information. In the new data system, this is achieved by preparing translators to con- vert existing data into SIC format and by designing new survey forms for those TPC's for which a new canvassing format is required. Design of the Survey Form Before a canvass survey form is designed for a given mineral, it is essential to determine (1) the TPC's for the mineral, (2) whether there is a need for a separate canvass survey form for each of the TPC's or if several TPC's can be accomodated as separate columns com- bined into one canvass survey form, and (3) a simple form design to capture the essential information and to facilitate uniform reporting by all firms that are canvassed. 20 The canvass survey form may be designed for a mineral or for each TPC for that mineral. The general principle adopted in the end-use surveys is to have a separate questionnaire for each TPC. This follows essentially from the defini- tion of TPC's and the care exercised in defining TPC's. Sometimes, however, the cost of the survey is minimized by designing one com- bined questionnaire for several TPC's. This occurs when (1) the end-use sectors of two or more potential TPC's are more or less common to both, (2) the TPC's are produced by firms in the same industry, or (3) both of these factors are com- bined. An example of a combined form with three TPC's is shown in table 4, a questionnaire for zinc compounds. Here an end-use survey for three zinc compounds — zinc oxide (the most important zinc compound) , other zinc compounds, and zinc dust^ — are all combined into one canvass survey form. Table 4 also illustrates the basic design of an SIC-based questionnaire to canvass firms in a particular industry. The first two columns describe the SIC classifications relevant for this TPC. Data on shipments are entered in the last three columns. "zinc dust either substitutes for zinc compounds in some end-use sectors, such as paints, or is the starting point for the continued production of other zinc compounds. The first stage in the design of this questionnaire was the process of identifying the several secondary and tertiary intermediate products of the TPC, which led to a table such as table 5. This second table serves two purposes. First, it insures that all the relevant end-use sectors are considered for inclusion in the canvass survey form. Second, it gives a preliminary idea as to whether it is sufficient to ask for data for all zinc compounds together, or whether a distinction is required among the more important products. It can be seen from table 5 that in several end-use sectors both zinc oxide and other zinc compounds derived from zinc oxide are used. Thus, this type of table allows the analyst to draw a first conclu- sion that data on zinc oxide should be collected separately from data on other zinc compounds. The second step in the design of a questionnaire is to correlate the mine- ral's end uses such as are shown for zinc in table 5 with SIC groups or industries. Whether a three-digit group or a four- digit industry is selected will depend upon (1) the percent of total consumption accounted for by that SIC category and (2) the extent of overlap in consumption among various four-digit industries in the three-digit group. The general rule that has been developed, with the help of several case studies, is that a four- digit industry will be specified only in exceptional cases when required to give clarity to the data or when the consump- tion in the specified industry is highly significant. 21 TABLE 4. Proposed canvass survey form for zinc compounds SIC code End-use industries Shipments , pounds of contained zinc Zinc Zinc Other dust oxide zinc compounds 01' AGRICULTURE 22 TEXTILES 24 LUMBER AND WOOD PRODUCTS 26 PAPER AND ALLIED PRODUCTS 28 CHEMICAL AND ALLIED PRODUCTS 2816 Inorganic pigments 2822 Synthetic Rubber (vulcanizable elastomers) 2823 Synthetic manmade fibers, rayon 284 Soap, cleaners, and toilet goods 285 Paints, varnishes, and allied products.... 2899 Soldering and welding flux 2801 2 Other chemicals 28 CHEMICALS AND ALLIED PRODUCTS Subtotal 2992 LUBRICATING OILS AND GREASES 30 RUBBER, NATURAL 32 CERAMICS AND GLASS PRODUCTS 325 Floor tile and other structural clay products 326 Pottery and related products 3201 2 Other ceramics and glass products 32 CERAMICS AND GLASS PRODUCTS Subtotal 347 ELECTROPLATING, PLATING AND COATING 3679 FERRITES OTHER IDENTIFIED END USES (Specify) UNIDENTIFIED END USES EXPORTS TOTAL 'Corresponds to Division A, "Agriculture, Forestry, and Fishing"; 01 is used to avoid alphanumeric coding. 2 This is not an SIC code. Subscript 01 is added to a 2-digit code to report all other shipments to that 2-digit category, except those 3- and/or 4-digit categories specified. 22 TABLE 5 . - End uses of zinc compounds End-use description Zinc oxide Zinc chloride Zinc sulfate Other zinc compounds Synthetic rubber. , Natural rubber. . . Rayon Textiles Wood preservative, Cosmetics Fungicides Paper Lubricants Dry batteries. Electroplating. . . Galvanizing Ceramics Phosphors Laundry soap Agriculture Dyes Cement plaster. . . Pharmaceuticals. . Paints , X X X X X Table 6 shows an end-use question- naire for hot-dip galvanized steel, one of the more lengthy forms. A basic fea- ture of the standardized design of this questionnaire is the use of the tiered structure of SIC that is intended to increase the accuracy of data reported at higher levels of aggregation, such as the SIC two-digit group 34. This tiered structure satisfies the objective to col- lect data as accurately as possible for all significant consumption (that is, greater than or equal to 1 percent of total) within SIC two-digit major groups. In the case of agriculture, SIC end use is specified at the two-digit level. In other instances, a three-digit group (341) or a four-digit industry (3446) is specified. In such cases where three-digit groups or four-digit indus- tries are specified, the questionnaire has the respondent calculate a group sub- total for each relevant two-digit major group. As a general rule, all survey forms are designed with separate categories for each SIC end use that constitutes more than 1 percent of total shipments. The arbitrary cutoff point of 1 percent has been established below which an SIC end- use questionnaire entry line will not be assigned. This cutoff serves as a sig- nificance test that, while it limits entries on a questionnaire to a manage- able number, also provides data for all significant uses. 23 TABLE 6 . - Questionnaire showing shipments of hot-dip galvanized steel by end-use industries SIC code End-use industries Shipments , net tons 01 ' AGRICULTURE CONSTRUCTION 1 52 Residential building construction 154 Nonresidential building construction 16 Highway and other heavy construction CONSTRUCTION Subtotal 25 FURNITURE AND FIXTURES 2514 Metal househole furniture 2501 2 Other metal furniture and fixtures 2 5 FURNITURE AND FIXTURES Subtotal 34 FABRICATED METAL PRODUCTS 341 Metal cans and shipping containers 3446 Architectural metal work 3462 Iron and steel f orgings 3496 Miscellaneous fabricated wire products 3401 2 Other fabricated metal products 34 FABRICATED METAL PRODUCTS Subtotal 35 MACHINERY, EXCEPT ELECTRICAL 352 Farm and garden machinery 353 Construction, mining, and related machinery 3551 Food products machinery 3501 2 Other machinery , except electrical 35 MACHINERY, EXCEPT ELECTRICAL Subtotal 36 ELECTRIC AND ELECTRONIC EQUIPMENT 361 Electric distributing equipment 363 Household appliances 3644 Non-current-carrying wiring devices 3601 2 Other electric and electronic equipment 36 ELECTRIC AND ELECTRONIC EQUIPMENT Subtotal 37 TRANSPORTATION 3714 Motor vehicle parts and accessories 3701 2 Other transportation 37 TRANSPORTATION Subtotal 39 MISCELLANEOUS MANUFACTURING 394 Toys and sporting goods 3901 2 Other miscellaneous manufacturing industries 39 MISCELLANEOUS MANUFACTURING Subtotal OTHER IDENTIFIED END USES (Specify) UNIDENTIFIED END USES TOTAL Corresponds to Division A, "Agriculture, Forestry, and Fishing"; 01 is used to avoid alphanumeric coding. 2 This is not an SIC Code. Subscript 01 is added to a 2-digit code to report all other shipments to that 2-digit category, except those 3- and/or 4-digit cate- gories specified. 24 Separation of Other Uses To Prevent Overaggregation In quantitative questionnaires, a category such as "all other" is useful to capture all the miscellaneous categories of consumption not included among those separately listed. In the canvass forms designed for the end-use data collection system, this miscellaneous "all other" category is kept as small as possible. For example, a firm that ships hot- dip galvanized products to the category Transportation may sell the bulk, of these shipments to automobile manufacturers (SIC 3714). The same firm may also sell products to other transportation equip- ment manufacturers, such as railroad equipment and shipbuilding, neither of which may be individually signifi- cant, although they may be significant collectively. If automobiles, railroad equipment, and shipbuilding were each separately listed, there would be three end-use cat- egories for Transportation, which adds length to the form while in this case only adding trivial data. Moreover, the significance test should not be violated, or the end-use classes will multiply rap- idly. On the other hand, if the Trans- portation classification is limited to automobiles, the respondent's shipments to other types of transportation manufac- turers would be reported in a general miscellaneous catchall category, "All other. " To avoid such overaggregation into a general miscellaneous category, a method was developed wherein a coded "other" category is provided for each SIC two- digit group (for example, 3701, Other Transportation) . This means that the residuals appropriate for each two-digit major group can be separately reported. This insures, at the two-digit level, that information is as accurate as could be expected. However, such liberty within SIC numbering causes a conflict with SIC itself. In the SIC system, there are miscellaneous categories. All the cate- gories not specified elsewhere for a two- digit category ordinarily are classified within a three-digit category in which the last digit is a 9. Thus, SIC 359 is Miscellaneous Machinery, Except Electri- cal. This methodology of using 9 is also repeated at four-digit levels, such as with SIC 3589, Service Industry Machines, Not Elsewhere Classified. The existence of categories such as 359 complicates the problem. Because SIC 359 is defined in SIC to include only the miscellaneous machinery specifically mentioned in the SIC manual, the use of an XX9 category for a general miscellaneous category would be misleading for our purpose. What is needed instead is a true "all other" category to capture shipments not only to SIC 359, but also to all other three-digit categories not speci- fied within 35X in the survey form, such as 354, 356, or 357. The special codings such as 3501 were developed by Penn State as a solution to this problem. The "01" is suffixed to an SIC two-digit category, such as SIC 35, to provide a place to report shipments to that particular SIC two-digit category rather than to the three-digit groups or four-digit indus- tries separately specified in each ques- tionnaire. This solves. the problem. To return to the previous example, the questionnaire asks the hot-dip gal- vanizer to report shipments to SIC 3714, Motor Vehicle Parts and Accessories, and to 3701, Other Transportation Equipment. This format insures that (1) data are accurate at the two-digit level and (2) a single catchall miscellaneous category will not overwhelm the data base. How- ever, in a technical sense, 3701 is not an SIC category. The "01" coding concept is used because no such code exists in SIC; hence, its use will not conflict with any of the existing SIC categores. Consider two firms conducting hot- dip galvanizing, and that although one firm is identical to the above case, the other is a firm that for some reason (because it sells through a sales agency or is otherwise unable to make a 25 determination of a customer's activity) does not know what portion of its ship- ments is consumed in "automobiles" versus "other transportation equipment." How- ever, the firm can report how much is consumed in all "transportation equip- ment." The above-suggested format allows the second firm to report its data aggre- gated at the suggested level, without forcing it to report at a level of detail at which it is incapable of reporting, the example below shows how the second firm will report its "other" shipments on the line, SIC 3701. SIC code 37 3714 3701 37 Description of end-use industry TRANSPORTATION EQUIPMENT Motor vehicle parts and accessories Other transportation equipment TRANSPORTATION EQUIPMENT Subtotal Shipments (example) 120 120 If for a different commodity, for example, aluminum, shipments to both motor vehicles and aircraft and parts are significant, the questionnaire designed as shown below: can be SIC code Description of end-use industry Shipments (example) 37 TRANSPORTATION EQUIPMENT 371 Motor vehicles and equipment 80 372 Aircraft and parts 10 3701 Other transportation 10 37 TRANSPORTATION EQUIPMENT Subtotal 100 Note that in both cases the two- digit (SIC 37) information provides a good total. This flexible questionnaire design format allows collection of data for shipments to specific three-digit and four-digit SIC categories, the signifi- cance of which varies among minerals. The format captures all residual ship- ments to the SIC two-digit code catego- ries specified for each mineral commodity and also provides reliable data at the SIC two-digit level for all mineral commodities. The design of a survey questionnaire also includes a few categories that are not SIC-coded end uses. Of these, "Other Identified End Uses (Specify)" is a cat- egory specifically inserted in an effort to induce a respondent firm to report shipments to an end use not otherwise included on the questionnaire but which is significant for that particular respondent firm. The respondent is expected to briefly identify that special end-use sector in such instances. Space is provided on the questionnaires for this purpose. This category, "Other Identified End Uses," also provides for respondent firms to report shipments to new end uses as they arise. The additional questionnaire cate- gories "Other Unidentified Shipments," "Exports," etc., are self-explanatory. Whenever interfirm sales are common in an industry, provision is also made to report such sales. Questionnaires are designed so that firms report their shipments by weight, because quantity data are needed to link the intermediate product to the contained weight of the mineral and, therefore, to U.S. mineral requirements. However, there may be special situations such as metal carbides for which the weight of material sold is so small that some firms maintain records only on the basis of dollar sales. In such instances, data can be collected by the Bureau on the basis of dollar sales, provided the 26 survey form also requires the firm to supply a conversion factor to convert the dollar sales figures into a measure of contained mineral. This conversion fac- tor of necessity will have to be an aver- age for each respondent of all his/her products and for all the sales contained in the special TPC. Because prices may fluctuate from time to time, the firm should be asked to report the conversion factor for each year in such cases. But collection of end-use data on a dollar sales basis is a poor approximation of the desired end-use data by weight. Hence, monetary values should be resorted to only when absolutely necessary. A Modified Survey Form The design of the typical survey form discussed in the previous section is flexible enough to be adapted to the par- ticular information needs of an unusually complex mineral market. Table 7 is a canvass survey form designed for survey- ing producers of high-temperature alloys that contain many minerals. The survey form for high- temperature alloys consists of a Work Sheet that is retained by the respondent and a Reporting Form submitted to the Bureau. Each of these sheets is further subdivided into Section I, Average Nom- inal Composition of High-Temperature Alloys Produced, and Section II, Ship- ments by End Uses. Section II of the reporting form is identical in all respects to tae general case discussed earlier. The reasons for the differences between the Work Sheet and the Reporting Form are explained below. Section I provides space for respondent firms to report a weighted average of the composition of each firm's product mix. This feature is designed to improve the information base that is used to link the demand for the primary metals to end-use sectors. Entries in section I are coefficients for various metals, as discussed earlier. Section I provides space for the respondent to estimate the amount of cobalt, chromium, and other ferroalloy metals incorporated in the products shipped to various end-use sectors. Information reported in section I considerably minimizes the problems asso- ciated with the use of an average coef- ficient for purposes of mineral end-use estimation. This is because, although high-temperature alloys are a very diver- gent group of alloys, firms in the indus- try tend to specialize in a limited range of products. This fact was used in the design of the survey form so that, by directing each respondent firm to report the average composition relevant to itself, the average metal composition for all firms can be brought as close as pos- sible to true average for all metal con- tained in the total product shipped to an end-use sector. In other words, by suit- ably designing the survey form, the Bureau c^n obtain an average measure of contained cobalt that is unique to the particular product mix of each firm instead of using a national average coefficient to allocate cobalt for all high-temperature alloys. This measure is also linked to the relevant end- use markets for the particular product mix. Section II of the Work Sheet is fur- ther segmented to parts A, B and C. This is done to insure more accuracy and more uniformity in reporting. Because respondents return the Reporting Form but retain the Work Sheet, the respondent can place confidential data on the Work Sheet in order to calculate data requested on the Reporting Form. (Note. — A distinc- tion was made between the design of the Work Sheet and the Reporting Form. Some firms in the industry felt that the detailed exercise required to complete the Work Sheet would provide information that is confidential. The asymmetric design devised by Penn State means that confidential data need not be placed on the Reporting Form. ) 27 o , CD > u p (/) CD w I (3 en n 3 ■u cfl I CD J-) I .£ M ■H S3 28 High-temperature alloys are not always sold directly to the customers by the producing firms. Often, shipments are made to intermediate fabricators, such as forge shops or casting works.. During the case study on high-temperature alloys, several firms insisted that forge shops and casting works be designated as end-use markets for high-temperature alloys. This would have caused serious dilution of the end-use information. Furthermore, it would also have allowed those firms in the industry that did not want to report market details to report ambiguously much of their shipments as going to forge shops and casting works. They could do this despite the fact that sales practice in the industry is such that end-use destinations are rather well-known to producers through customer and product identification. Also, there was no guarantee that all the firms in the industry would report data in the same manner, because the more con- cientious firms would report estimates shipments to end-use sectors, while other firms might not do so. All these complications were avoided by providing a Work Sheet that allows each respondent firm to record the firm's direct shipments to end uses in part A. Part A of the Work Sheet also lists non- captive forge shops and noncaptive cast- ing works as two quasi end uses. Part B of the Work Sheet directs the firms to show shipments made directly to forge shops in terms of the end-use markets that the firm estimates for their forge shop customers. Part C does the same for shipments made to casting works. The horizontal summation for each row across parts A, B, and C then gives the total shipments by each respondent to each SIC end-use market. For example, the quan- tity shipped to SIC 372, Aircraft and Parts would be the sum of the quantities shown for the SIC 372 row on part A, part B, and part C of the Work Sheet. This total is then entered for SIC 372 on the Reporting Form, which is returned to the Bureau. This survey form, which is unique to high-temperature alloys, also points to the cost effectiveness of the two- dimensional information system now being introduced by the Bureau. Such a two- dimensional approach would enable the Bureau of Mines to collect, by means of one annual canvass, all the relevant information for eight or more metals without any loss of accuracy. If instead, the information system was designed for each primary metal, there would be a minimum of eight separate can- vasses, one each for each metal specified in section I of table 7. Lead: An Example of a New SIC-Based Survey Form Table 8 shows section 4 of the Bureau of Mines' 1978 revision of the lead survey form. 10 Section 4 provides an example of an SIC-based replacement of an older Bureau form. The old format of section 4 of the previous survey form is shown in table 9. The new format repre- sents a compromise between the standard format recommended by Penn State for end- use questionnaires and the desire of the Bureau's analysts to have historical con- tinuity with statistics derived from the earlier form; thus, the format shown in table 8 should be considered an inter- mediate step to the general SIC format recommended for the Bureau's end-use questionnaires. The new format differs from the old primarily in having two new columns inserted after the first column, "Inter- mediate products." The two new columns are labeled "SIC Code" and "End-use industries," column 2. The next column, "ADP Code," serves merely as numerical identification for data processing clerks. The last four columns — "Soft lead" (201) to "Lead in copper-base scrap" (204)--are identical to the last four columns in the older lead survey. 1 °Lead Secondary Smelter and Consumer Report" (Form 6-1108-MA, Nonf er. ) . 29 TABLE 8. - Revised lead survey form 4. Consumption of lead by class of product and ond-uso industry (Short tons of load content) Intermediate product! (1) SIC Code End-in* industrial Code Soft lead (201) lead in ontimoncol lead (202) Lead in alloys (203) copper-bate wrap (204) METAL PRODUCTS: 3482 - 3482 Shot 301 Ammunition Bullets 302 35 36 371 370 U Machinery except 304 Bearing metals Electrical and electronic 305 Motor vehicles and 306 'Other transportation eauipment 307 Brass and bronze- 3351 Brass and ingots, billets bronze 309 r ui J 3o1 Cable covering 1 310 Communication coble 311 Calking lead 15 Buildina construction 313 36 371 Electrical machinery and 314 Casting metals (die, permanent mold, and slush-weights Motor vehicles and 315 370 1-^ Other transportation equipment 3443 Nuclear radiation shielding 316 317 Pipes, traps and other extruded products 15 3443 319 Storage tanks, process vessels, etc 320 15 Building construction 3443 Storage tanks, process vessels, etc 3693 Medical radiation shielding 322 323 324 326 341 367 360 U Metal cans and shipping 327 Solder Electronic components 328 Other electrical machinery 329 371 Motor vehicles and equipment 330 Storage battery "36911 Storage batteries-SLI automotive 36912 Storage batteries- industrial and traction 332 grids, posts, etc. metal 333 Storage battery oxides 3691 1 Storage batteries-SLI automotive 36912 Storage batteries- industrial and traction 335 336 _ 371 Motor vehicles and Terne metal equipment 338 _ 27 Printing and allied Type metal industries 339 Other metal 34^/Other metal products 340 285 32 281 Paints 342 PIGMENTS: Red lead and litharge - Glass and ceramic 343 (for other than Color 344 storage batteries) 28 Other pigments 345 CHEMICALS: Lead alkyls 2911 Petroleum refining 347 MISCELLANEOUS USES: (please specify)**' 350 TOTAL 399 ^Thit is nol an SIC product clatt- .!/ Include 1 collaptible tubes, toil, a Vinclude s powder, rubber and pi i added to a 2-digit code to report other conso aling, galvaniiing, plating and fishing weights additives, anodes, steel products, etc. 30 TABLE 9. - Old lead survey foi 4. Consumption of lead by doss of product (Short tons of lead content) Products HI Code Sail lead (2011 leod in anlrmontol lead 12021 leod in oltoyi 12031 lead in copper baie scrap {204| METAl PRODUCTS Ammunition — billets shot, wire 301 Renrinrj mptnk (nvrj % Sh) 302 Brass and brome ingot 303 Coble covering and sleoving Invg J, Sbj 304 Calking lead Casting metals (die, permanent mold, 305 306 Collapsible tube blanks, disks 307 Foil 308 Pipes, traps, and other extruded 309 310 311 Sloroge bottery grids, posts, etc., of SbPb lavg * Sb| 312 313 Terne metal 314 315 PIGMENTS: While lead 317 Red lead and litharge (for other than storage batteries! 318 Other (please specify) 319 CHEMICALS: 321 Lead arsenate 322 Other (please specify/ 323 MISCELLANEOUS USES: 325 Got vani zing Leod plating Weights and ballast OTHER USES (please specify) 326 327 328 329 TOTAL 399 Signolurt I, lie Dal, 31 A major difference between section 4 of the 1978 lead survey form and the SIC survey forms proposed for other com- modities is the rigidity of the end-use classifications of the new lead survey. In the revised lead survey, each respond- ent firm is asked to report its shipments to the selected few end uses specified for each intermediate product. There is no provision in this intermediate form to improve the accuracy by accumulating data at SIC two-digit levels. 11 The reasons for the intermediate form are fully detailed in the appendix to Chapter XII, Lead, in Penn State's Sep- tember 1978 report, "SIC Based End-Use Scheme for Information Critical Minerals" (28). 12 A Statistical Sampling Alternative An alternative to a comprehensive industry canvass is to use a suitable sampling method to estimate statistically valid data for industry shipments. If feasible, this method can reduce the cost of the survey and reduce the reporting burden on the industry. A sampling approach is particularly suited to those atomistic industries that serve largely regional markets. sampling about 10 percent of establish- ments in this particular industry. The results were mixed. On one had, the study proved that a stratified random sampling both is feasible and can provide valid results for SIC-based end-use data, even when the respondent firms are not well acquainted with SIC. On the other hand, the study also showed that even an intensive effort that involved two sepa- rate mailings and extensive telephone canvassing was not sufficient to persuade firmsin this industry to participate in the voluntary canvass. In the case of steel service cen- ters, it was possible to design a sam- pling method, but it was impossible to implement the scheme. The respondent firms often serve narrow, local markets and focus on present customers. They did not perceive sufficient benefits for themselves from nationally oriented, SIC- based data to complete the canvass. The voluntary sample survey of steel service centers produced only a 20-percent response, an inadequate basis for drawing statistically valid conclusions. Integrating Data From Other Sources Into the SIC-Based Information System A pilot study of steel service cen- ter shipments was conducted by Penn State to determine the feasibility of a survey 1 1 Penn State carried out a study on an SIC-based end-use data scheme for U.S. lead in the summer of 1976. Two alternative survey forms were pro- posed in that study. The Lead Indus- tries Association (LIA.X, which coop- erated with the Penn State study, however, wished to have a more detailed product to end-use classifi- cation scheme. The alternative form, shown in Penn State's 1978 report, reflected LIA's concern for details and was more consistent with the standard format of end-use questionnaires . 1 Underlined numbers in parentheses refer to items in the bibliography preced- ing the appendixes at the end of this report. Several trade associations, such as the American Iron and Steel Institute, Copper Development Association, and Alu- minum Association, collect end-use data periodically from their member firms. For the SIC-based information system pro- posed in this study to be cost effective, it was necessary to examine whether end- use data that are now available from such associations could be integrated into the SIC-based information system. Reports of end-use data published by trade associations tend to be issued very promptly and usually in such detail that the data completely satisfy the informa- tion system proposed in this study. Until now, the use of such data has been inhibited by two major considerations. First, each trade association follows a market classification of its own, appar- ently suited to the particular needs of the trade group involved. Therefore, 32 data from different trade associations for different materials are not gathered on a comparable basis. Second, until now, information from trade associations has been dissemenated only to the member firms and generally has not been available to the Government or the public. Even when trade associ- ation data are provided to the public, the data are often released after major aggregations or other recasting, so that the published information is not com- parable to SIC. For example, the Copper Development Association (CDA) published data on annual consumption of copper in five major end-use markets. But the CDA itself collects data from brass mills on shipments to 54 "end uses" and for 4 dif- ferent products. The CDA combines its primary data for brass from the 54 end uses with estimates of copper wire mill shipments, and only these aggregated data are provided to the public. However, if the data that are made available by trade associations can be integrated into an SIC-based information system, the public interest would be served. Hence, the attempt is made in the Bureau's new end-use reporting system to use trade association data whenever such data are available. To achieve the above objective, it is essential to "correlate" the trade association classification with SIC. Such attempts at correlation were made unsuccessfully in the past by the Ameri- can Iron and Steel Institute, Inter- national Nickel Company (for steel), and Copper Development Association (for brass mill shipments) . Table 10 illustrates the reason for the failure of these industry efforts to correlate existing market classifications with SIC. CDA 310, Automotive Non- electric, consists of the 95 items listed in table 10, but CDA 310 itself is but one of the 54 CDA end uses. CDA has defined the 310 "end use" to include products such as automotive hardware, piston rings, and passenger car bodies, as described in column 2. As shown in column 1 , CDA then proceeded to iden- tify each one of these products with the SIC four-digit industry or five-digit SIC-based product class that described these products. Column 1 of table 10 shows that CDA end-use category 310 cannot be identified with one unique SIC code. Automotive Hardware is produced by firms in SIC 3429, Piston Rings in SIC 3529, Pas- senger Car Bodies in SIC 3711, and so on. Because each of these examples belongs to three entirely different SIC two-digit major groups, the difference cannot be reconciled even at the two-digit major group level. Because of such extensive overlaps into different SIC groups and industries, CDA concluded that it is not possible to compare CDA end uses with SIC. Two points should be noted here. First, CDA was not willing to redefine its end use 310 (and other end uses) in a way that would minimize the above prob- lem. Instead, CDA attempted (and failed) to achieve a perfect one-to-one correla- tion between two entirely different clas- sification systems. Second, this failure is not a sufficient indication that SIC is a poorly defined classification sys- tem. All that it indicated is that SIC is defined in a different manner than that utilized in CDA's classification system. A reasonable basis for comparing such individualized industry classifica- tions as that of CDA with SIC does exist, however, if such comparisons are made with SIC commercial uses. 33 TABLE 10. - CDA market classification, an example using CDA end-use 310, Automotive, Nonelectric SIC No. Description 3429 4 Automotive hardware. 3429 4 Bus hardware. 3429 4 Hardware, motor vehicle. 3429 4 Motor vehicle hardware. 3429 4 Truck hardware. 3465 Automotive trim. 3465 Grills, automotive radiator. 3465 Hubcaps, automotive. 3465 Trim, automotive body. 3519 1 Bus engines, diesel. 3519 1 Diesel engines, truck, bus, and auto. 3519 1 Truck engines, diesel. 3585 3 Air conditioners, auto, truck, and bus. 3585 3 Automotive air conditioners. 3585 3 Bus air conditioners. 3585 3 Truck air conditioners. 3592 Carburetors, automotive. 3592 Floats, carburetor. 3592 Piston rings. 3711 1 Ambulances, assembly. 3711 1 Assembly — cars, trucks, and buses. 3711 1 Bodies and body parts, automotive. 3711 1 Bodies, motor vehicle, other than truck or bus. 3711 1 Buses, assembly. 3711 1 Chassis, automotive, for sale separately. 3711 1 Fire engine assembly. 3711 1 Motor vehicle bodies, other than truck or bus. 3711 1 Motor vehicles, assembly. 3711 1 Passenger car assembly. 3711 1 Passenger car bodies. 3711 1 Taxicabs, assembly. 3711 1 Tractors, motor truck (highway), assembly. 3711 1 Truck tractors and trailers, assembly. 3711 1 Trucks, motor, assembly. 3713 Ambulance bodies. 3713 Bodies, truck and bus. 3713 Bus bodies. 3713 Cabs, for industrial trucks. 3713 Coach and bus bodies. 3713 Truck beds. 3713 Truck bodies. 3713 Truck cabs. 3713 Truck tops. 3714 1 Accessories, automotive, nonelectrical. 3714 1 Automobile parts and accessories, nonelectrical. 3714 1 Bearings and bushings, automotive. 3714 1 Brakes and brake parts, automotive. 3714 1 Bumpers, automotive. 3714 1 Bus engines, except diesel. 3714 1 Bus parts and accessories. 34 TABLE 10. - CDA market classification, an example using CDA end-use 310, Automotive, Nonelectric —Continued SIC No. Description 3417 1 Bushings, automotive. 3714 1 Clutches and parts, automotive. 3714 1 Cores, automotive radiator. 3714 1 Cores, tire valve. 3714 1 Disk brakes, automotive. 3714 1 Drum brakes, automotive. 3714 1 Engine bearings, automotive. 3714 1 Engines and engine parts, automotive. 3714 1 Filters, automotive. 3714 1 Fuel systems and parts, except carburetor. 3714 1 Gears, automotive. 3714 1 Governors, automotive. 3714 1 Headers, radiator, automotive. 3714 1 Heaters, automotive. 3714 1 Lubrication systems and parts, automotive. 3714 1 Motor vehicle parts and accessories, nonelectrical. 3714 1 Passenger car parts and accessories, nonelectrical. 3714 1 Pollution control equipment, automotive. 3714 1 Power steering and brakes, passenger car. 3714 1 Pumps, fuel and water, automotive. 3714 1 Radiator and parts, automotive. 3714 1 Rebuilding, motor vehicle engines, factory. 3714 1 Rebuilt engines and parts, automotive, except carburetors. 3714 1 Shock absorbers, automotive. 3714 1 Steering mechanisms, automotive. 3714 1 Strainers, gasoline, automotive. 3714 1 Tanks, radiator, automotive. 3714 1 Thermostats, automotive. 3714 1 Tire inflators, automotive. 3714 1 Tire valves. .3714 1 Transmission equipment, automotive. 3714 1 Truck engines, nondiesel. 3714 1 Valves, automotive, tire. 3714 1 Wheels and parts, automotive. 3715 Chassis, truck trailers. 3715 Horse vans. 3715 Semitrailers for truck tractors. 3715 Tank trucks. 3715 Trailers, motor truck. 3715 Truck trailers. 3715 Vans, freight. 7539 Automotive radiator repair shops. 7539 Automotive repair shops, n.e.c. 7539 Radiator repair shops. 7539 Repair shops, automotive, n.e.c. Source: Copper Development Association, End-Use Classification Scheme for Copper and Copper Alloy Mill Products, 1977. 35 Consider again the previous example of CDA category 310, Automotive, Non- Electric. This category represents con- sumption of copper and copper alloys in nonelectric uses in motor vehicles. The appropriate end-use sector then is motor vehicles. Irrespective of whether a par- ticular product is produced by a metal fabricator or an automotive plant, 13 the material is consumed in the manufacture of motor vehicles. Therefore, SIC 371, Motor Vehicles, correlates extremely well with CDA 310. In other words, the quan- tity of copper used in CDA 310 is con- sumed in the production of motor vehic- les, whether we choose to designate the consumption in terms of SIC or in terms of DCA classification. In this explana- tion, then, we have a basis to compare CDA's classification system with the Bureau's SIC-based system. The above example was meant to show the basic reason for the failures of past industry attempts, as well as to explain that within the objectives of the present study, it is highly probable that indus- try data can be integrated into an SIC- based information system without a sig- nificant loss in accuracy. This logic assumes, for example, that a brass pro- ducer will report the same tonnage when he is asked to report shipments for con- sumption in motor vehicles, whether or not motor vehicles are described as CDA 310 or SIC 371. It should also be noted that the quantity shown for 310 in CDA's own data base tells the industry only the consump- tion of brass mill products for automo- tive use. There is no reason why these product listings have to be viewed dif- ferently when the data are analyzed in terms of SIC. The above example shows that, at least as a first approximation, it is possible to correlate a trade 3 The producing industry is the basis used by CDA for assigning the SIC numbers shown in table 10, column 1. association's end-use categories with SIC-defined end uses, even when there does not appear to be a one-to- one correlation between the two classifications . Table 11 provides a means to trans- late the CDA 54 end-use categories into a minimum number of SIC categories. The CDA 54 end-use categories are regrouped in terms of 19 SIC categories using the following criteria: (1) Significant vol- ume of shipments as indicated by avail- able data on brass mill shipments, (2) the clarity of data at the SIC two-digit major group level, and (3) the desirabil- ity of a list of end uses common for both brass mill and wire mill shipments. The format of table 11 collapses to 19 the number of SIC end uses that cor- respond to the 54 CDA end uses. In table 11, CDA data for 1972 brass mill shipments are used as a measure of the volume of shipments. Column 3 lists the percent of brass mill shipments demanded in each SIC category specified. The reader also should be aware that brass mill shipments are but one part of copper shipments, which are what the Bureau wishes to measure. This brass-copper relationship is the reason that a major exception to the volume cutoff point apparently is made in the choice of SIC end uses in electrical and electronic equipment. Among SIC 361, 362, or 366, none is critical if con- sidered on the basis of volume of brass shipments. These three SIC categories are critical for end-use data relating to copper wire mill products and are, there- fore, retained. The percentages shown in table 11 are misleading because they per- tain only to brass mill shipments. Another major exception that is made in table 11 is the inclusion of SIC 38, Instruments and Related Products. This item is included to insure clarity of data at the SIC two-digit level. 36 TABLE 11. - Recommended translator for converting CDA data into SIC data base SIC SIC description Brass shipments, percent of total Corresponding CDA end uses code CDA CDA Description 1 code 15 3 190 522 Onsite heat exchangers. 33 6 910 Rerollers and redrawers. 34 Fabricated metal products: 343 6 130 Plumbing and heating. 345 10 610 810 Fasteners. Screw machine products. 346 Metal forging and stamping. . 6 820 830 Forging. Stamping. 348 6 760 Ordnance. 3401 Other fabricated metal 9 110 Builders' hardware (342). products. 120 521 530 230 290 Architectural products (344). Industrial heat exchangers (344). Industrial valves and fittings (349). Utensils and plated ware (3401). Other household products (3401). Subtotal 890 Other component parts (3401). 37 Nonelectrical machinery: 35 358 6 4 140 529 590 Cooling. Other heavy industrial equipment. 3501 Electrical and electronic 10 Other industrial machinery and equipment. 36 equipment: 361 Electrical transmission and 1.5 421 Transformers. distributing equipment. 422 Switchgear and industrial circuit breakers. 362 Electrical industrial 2 410 Motors and generators. apparatus. 423 429 Industrial controls. Other electrical distribution and control. 366 1.5 431 Telecommunications. 3601 Other electrical and elec- tronic equipment. 7 210 220 440 432 490 Major appliances except air conditioners (363). Portable appliances (363). Wiring devices, lamp shell, and sockets (364). Electronics and other communication (3601). Other electrical and electronic products (3601). 12 Transportation equipment: 371 13 310 320 Automotive, except electrical. Automotive, electrical. 3701 1 330 Marine. Subtotal 390 710 740 Other transportation. 14 1 38 Instruments and related products . Timing devices (387). Instruments, except electrical measuring (3801). Miscellaneous manufacturing industries: 396 Jewelry, notions, buttons, 2 620 Closures. closures. 7 30 Jewelry. 3901 Other miscellaneous manu- facturing industries. 2 790 Other uses. 4 Other identified end uses: 12 1 920 930 Export. 100 lumbers in parentheses indicate appropriate SIC codes for CDA end uses grouped together in this translator for convenience. 37 Table 11 retains (1) all the major SIC end-use sectors relevant for copper and copper alloys and (2) the benefits of an end-use classification that is common for both wire mill (copper) and brass mill products. Also, since the end-use categories in table 11 are aggregated, there is less need for concern about the confidentiality of the source informa- tion. This was a problem that precluded direct use of CDA's 54-end-use data base. IMPLEMENTATION OF THE DATA SYSTEM The practical aspects of the imple- mentation of the SIC end-use data system were first tested using sulfur. After an industrywide trial run of the survey form to elicit industry comments and sugges- tions, the scheme for sulfur was imple- mented in 1976. The initial work was followed by an extended study to cover all information- critical minerals. Studies were com- pleted for the following mineral commodities: Aluminum Antimony Asbestos Bismuth Cadmium Chromium Cobalt Columbium Copper Fluorine Iron and steel Lead Mercury Mica Nickel Rare earths Sulfur Tantalum Tin Titanium Tungsten Vanadium Zinc Together, these 23 mineral commodi- ties represent a total of 42 techni- cal product classes (ferrous minerals, 18; nonferrous minerals, 17; and nonmetals, 7). Table 12, which is similar in con- cept to the traditional transaction matrix of an input-output table of the economy, lists the 42 TPC's and the SIC end uses relevant for each. In such a table, the column entries would represent the input of each TPC in the production of goods in SIC-based com- mercial uses , and the column totals would represent the total consumption of TPC's in the production of all domestic goods. Similarly, the row (line) entries would indicate the demand for the various TPC's in each SIC commercial use. Table 12 here consists of three pages; in reality it should be one sheet of 42 TPC columns (beginning with "car- bon steel" and ending with "asbestos") and 71 end-use rows (beginning with SIC 01, Agriculture and ending with SIC 39, Miscellaneous Manufacturing Industries) . The ratio of each actual entry in the table 12 matrix to the total output of the corresponding SIC commercial use yields the intensity-of-use of a TPC in that SIC commercial use. In the section "Symbolic Definition of Technical Product Classes" equation 4 was given. This equation can be used to calculate the intensity of use of a mineral as a func- tion of the intensity of use of its TPC's. It was also shown how properly specified TPC categories (the X's and Z's of figure 1) could be combined to yield a reasonable approximation of the intensity-of-use of a mineral (M) in SIC commercial uses. In other words, when the data system is in place for all min- erals, it will be possible to construct a second matrix similar to the one in table 12 with minerals as columns instead of TPC's, thus providing an intensity- of-use matrix for different minerals in SIC commercial uses. Many of the data needed to construct the matrix in table 12 are not now col- lected by the Bureau. The purpose of the new data system is to gather, for the first time, information leading to the construction of such a matrix. The end-use data for the 42 TPC's in table 12 can be obtained by the Bureau with a minimum of additional canvasses. 14 Other than the new survey forms already in use for sulfur, only three 14 See appendix B for references to speci- fic commodity studies. 38 new canvasses are needed: (1) High- temperature alloys, (2) magnetic alloys, and (3) carbides and hard-facing alloys. These three TPC's are common for a number of ferrous minerals, yet no end-use data are at present collected for them. Data for all other TPC's can obtained either by translating currently available data into the SIC end-use for- mat or by modifying existing Bureau canvasses. The statistics will be collected from the producers of the various TPC's on an annual basis. Annual collection reduces the burden on the reporting firms and also avoids the necessity for data on inventories. The annual time horizon implies that "shipments" are more likely to coincide with "consumption" than is the case with the monthly and quarterly data of many of the Bureau's current statistical reports. When fully implemented, the two- dimensional TPC end-use matrix shown in table 12 will for the first time provide and complete input-output matrix of mate- rial consumption in SIC end-use sectors. The table entries will help in improving the forecasting of the demand for mate- rials (TPC's), given forecasts of eco- nomic actibity in the specified end-use sectors. The matrix in table 12 also will yield data over time useful in iden- tifying the intensity-of-use of TPC's in various end uses, as well as the substitution between different technical product classes in the same end-use sectors. The new data system of the Bureau will constitute a major step in gaining a better understanding of the nature of demand for minerals in the economy. IPC end-use matrix for information-critical minerals (X's indicate data that will be available when implementation of the data system is complete) ■a C bO a) E 1 >» „ SIC End-use industries n a s " ai ra u ST? 2 n c BO t3 c 1 1 E fl code B H » •O Uj J3 H ■a s tt a *J 3 >»-H I >> ai >. N 2 < 4> -1 01 o ot ^H V £ o c o a » h 2 fi- Jrf 4-» J£ o fl 8- ■-H Q> td n 00 H now IM 1-. J3 bJ 8% as 1-3 S 2 -C O hi J3 O 3 X to H CO en uj a: « O J3 u o Eg u U O u e z a. Z (D Z U H H > < 01 10 15 16 20 22 25 26 27 281 282 283 284 285 286 287 289 28 291 29 30 31 Agriculture X Mining X Building construction X Heavy construction X Food and kindred products Textile mill products Furniture and fixtures X Paper and allied products Painting Industrial inorganic chemicals Plastics, materials, and synthetics Drugs Soap, cleaners, and toilet goods Paints and allied products Industrial organic chemicals Fertilizers and agricultural chemicals Explosives and miscellaneous chemical products Chemicals and allied products Petroleum refining Petroleum and coal products Rubber and miscellaneous plastics products.... Leather and leather products 01 10 15 16 20 22 25 26 27 281 282 283 284 285 286 287 28 291 29 30 31 X X X XX X X X X X X X X X X X X X X XXX X X X X Agriculture Mining Building construction X Heavy construction Food and kindred products Textile mill products Furniture and fixtures X Paper and allied products Painting Industrial inorganic chemicals Plastics, materials, and synthetics Drugs Soap, cleaners, and toilet goods Paints and allied products Industrial organic chemicals Fertilizers and agricultural chemicals Chemicals and allied products Petroleum refining Petroleum and coal products Rubber and miscellaneous plastics products.... Leather and leather products > a at m &o o> 1 *-3 o •O 3 CO •H w a co C >. •^ C n a. o TJ c td W 3 m 0) o 3 0) r-l 01 O X 3 1 4) ■a > E <0 so e e O TJ i- -a IM T3 to C 00 E ~H c e m U •a 3 "H TD >H u ■H « — o 2 g 2 8 en n os 0) >> 3 U m N CJ 3S B r-q u N O J C/5 ~ < u Eh '-J- cd x to to CO S X XX X X XX X XXX X X XXX X X XX X XX XX X X X X X XXX XXXXXXX XX X XX X XX XX X X X X X X X X XXX X X X X X X XXX XXX XX XX X X X X X XXX XXX X X TPC end-use matrix for information-critical minerals (X's indicate data that will be available when implementation of the data system is complete) SIC code End-use industries 321 324 325 326 329 32 331 332 333 334 335 33 341 342 343 344 345 346 347 348 349 34 351 352 353 321 324 325 326 329 32 331 332 333 334 335 33 341 342 343 344 345 346 347 348 349 34 351 352 353 Glass Cement Structural clay products Ceramics Abrasives and miscellaneous products Glass, ceramics, and clay products Steel mills X Iron and steel foundries Primary nonferrous metals Secondary nonferrous metals Nonferrous rolling and drawing Primary metal industries X Metal cans and containers X Cutlery , hand tools , and hardware X Plumbing and heating X Fabricated structural products X Screw machine products X Metal forgings and stampings X Plating, coating, and polishing Ordnance X Miscellaneous fabricated metal products Fabricated metal products X Engines and turbines X Farm and garden machinery X Construction machinery Glass Cement Structural clay products Ceramics Abrasives and miscellaneous products Glass, ceramics, and clay products Steel mills Iron and steel foundries Primary nonferrous metals Secondary nonferrous metals Nonferrous rolling and drawing Primary metal industries Metal cans and containers Cutlery, hand tools, and hardware X Plumbing and heating Fabricated structural products Screw machine products X Metal forgings and stampings X Plating, coating, and polishing Ordnance X Miscellaneous fabricated metal products Fabricated metal products X Engines and turbines Farm and garden machinery Const rue t ion machinery X X X X X X X X I > §■3 ■o 3 CO ■H W a to c •H C si n a. t: c -a 3 ui 3 . C « a) i-i C CO C r-» rO 00 *J U B T3 3 -H TJ "H ■H O -H O 5 1 § £ S 5 V >. U •H OQ M a X C N >-i N u -J ''■ 03 U fa [14 (4 X to UJ ffl z X X X X X X XX XX XX X X XX X X XXX X X X X XX X X X X XX X X X X TABLE 12. - TPC end-use matrix for information-critical minerals — Continued (X's indicate data that will be available when implementation of the data system is complete) SIC End-use industries code c O —1 ■C 01 U 0) u w < CI .J 0) o oj 3S o -u 01 C H m a) D. s H 08 1 >. J= O HO -H c o b0-4 i « C DO iH C o « ll •^ i O Q, i- s £ o u a 2 1 U O 01 JZ al U *-» 3 i c 3 M 9 at a* o. t-> a. - o 3 X X X X X X X X X X 355 356 357 358 35 361 362 363 364 365 366 367 369 36 371 372 373 374 376 379 37 38 39 354 355 356 357 358 35 361 362 363 364 365 366 367 369 36 371 372 373 374 376 379 37 38 39 General industrial machinery X Office and computing machines Refrigeration and service machinery X Nonelectrical machinery X Electric distributing equipment Electrical industrial apparatus Household appliances X Lighting and wiring equipment* Radio and TV receiving equipment Communication equipment X Electronic components Batteries and miscellaneous equipment Electric and electronic equipment X Motor vehicles and equipment X Aircraft and parts X Ship building and repairing X Railroad equipment X Guided missiles, space vehicles, parts Miscellaneous transportation equipment Transportation equipment X Instruments and related products Miscellaneous manufacturing industries X Machine tools, dies, etc Special industry machinery General industrial machinery Office and computing machines Refrigeration and service machinery X Nonelectrical machinery X Electric distributing equipment Electrical industrial apparatus Household appliances X Lighting and wiring equipment Radio and TV receiving equipment Communication equipment Electronic components Batteries and miscellaneous equipment Electric and electronic equipment X Motor vehicles and equipment Aircraft and parts Ship building and repairing Railroad equipment Guided missiles, space vehicles, parts Miscellaneous transportation equipment Transportation equipment X Instruments and related products X Miscellaneous manufacturing industries X 1 > .-< — < n) 0) U <1> m b0 w •a 1 33 in •H o 3 y w w Q. , •H C a. T1 TJ 3 tfl 3 T3 TJ n at O at a CU O .fl n O 3 1 01 u a. *d c ^ 0? '•\*° > V ■^ A *' "by V » o ' • '. c» & ..1^L%. *> -♦♦ ,-isS^-. X/ .*afe\ V** ' :Mfc: X^' .-3K- ' ^** ' .*' 1 °» o C« ■• .■ft/" %fc • V *-TT7' .A ^X v^ % .' «> ^5 ^ 1 ' • - ^ V t ^' .«* ^ ^^ <%*.....«*», .4> .t... .1... V t > _ T * ft *>■* A V * * • » *^^ « > u^ -'£M2>~\ ''^ < .*, .* ..•• .A *&.**' ; . av ^ .y^vr . «? v, ♦.«§?/ ^ ^ 7 ' ♦ V 'V o > -w ^°- * A v »bf o-' . *L 4 o * A v ^^ v 4 ^ v oV* >°^K r *d* »bv" > ^ j ^^ o « . **i A <, -»•> .g v V *7^T» A *?x *-.,,.* A o ^* •bv* •fc. G' G* % ♦TTf** A <- 'bV °^. .♦•■I' a V -.vT- ^,^ P vL^> > 5>^ "v> '"•* a & ^ *^T» A <. '».»• .G* V *TtTT» A <, '..?* .6^ ^ ' - '^d« *bl? * -P. 4 .'- v-^-'V ^■■wt-y v--^-v v-^-v v-r^-> v^' .V-* ^ *«