Updating report on the World Bank research program (Vol. 3 of 6) : Report of the World Bank advisory panel on commodities

20310 Vol. 3 Report of the World Bank Advisory Panel on Commodities May 1979 VILE COPY PREFACE The members of the Panel were: Dr. Ali Ahmad Attiga Secretary General Organization of Arab Petroleum Exporting Countries, Kuwait Professor Alan Brown University of Oxford England Dr. Eric M. Ojala Senior Research Fellow Massey University Palmerston North, New Zealand Dr. Affonso Celso Pastore Minister of Finance, State of Sao Paulo, Brazil The Panel met at Bank headquarters, July 10-14, 1978. After this session Dr. Attiga was obliged to withdraw from the Panel owing to the pressure of his responsibilities, but he contributed written observations and suggestions which are attached as Annex 3. A second session of the Panel was convened at Bank headquarters in March 1979, but in the event Dr. Pastore was unable to attend because of the demands of his new Cabinet responsibilities. Thus, Dr. Attiga and Dr. Pastore did not attend any meetings of the Panel after the first; however, Dr. Ojala visited Kuwait and 55o Paulo to consult with them. Professor Brown and Dr. Ojala worked at the Bank from March 5-9, 1979. The Panel members wish to express their thanks to Dr. Leonard L. Fischman, Consultant on Mineral Commodities, for his valuable comments on early drafts of the report. Members of the Panel also greatly appreciated the cooperation extended to them by the many Bank staff whom they consulted. TABLE OF CONTENTS Page No. 1. INTRODUCTION AND SUMMARY..................................... 1 1. Terms of Reference....................................... 1 2. Summary of Conclusions and Recommendations............... 2 I COMNODITY WORK AND OUTPUT IN TIHE BANK........................ 6 3. Location of Commodity Work............................... 6 4. Research Committee Projects.............................. 11 5. Types of Research Output................................. 13 6. Dissemination of Commodity Documents..................... 15 III. PANEL'S EVALUATION OF WORLD BANK'S ANALYSIS AND RESEARCH ON COMMODITIES...................................... 17 7. Objectives............................................... 17 8. Technical Aspects of Bank's Commodity Work............... 18 9. Supporting Bank Operations............................... 29 10. Orienting Commodity Lending Policy....................... 33 11. Improving the Bank's Capacity to Give Policy Advice to its Members............................. 41 12. Broadening the Understanding of the Development Process...................................... 43 13. Assisting in Developing Indigenous Research Capacity in Developing Regions........................... 44 IV. INSTITUTIONAL MATTERS........................................ 45 14. Commodity Steering Committee............................. 45 ANNEX 1: Recent World Bank Output in the Commodity Field ANNEX 2: Brief Reviews of Selected Individual Commodity Models ANNEX 3: Some Observations and Suggestions Concerning Energy and Commodities in the Research Program of the World Bank I. INTRODUCTION AND SUMMARY Terms of Reference The terms of reference given to the Panel had three main aspects. The primary purpose indicated was "to evaluate the quality and effectiveness of the economic support given to the World Bank's operations and country economic analysis in primary commodity trends." This evaluation was to include: "an assessment of the overall approach to primary product/price/volume medium-term forecasting"; and "a review cf the appropriateness of the dissemination of the results to member countries." This aspect of the terms of reference brings into consideration aill Bank work relating to primary commodities, wherever located. Some this work entails research. The other two aspects specified in the terms of reference relate only to research. The Panel was required to review the research undertaken and sponsored by the Bank in the area of primary commodities in relation to the Bank's research objectives. The Panel was then to make recommendations to the management of the Bank regarding research priorities for the next five years. In considering these aspects of its terms of reference, the Panel decided that it could not take a purist view of research, but should take into account all analytical studies, as well as the data base upon which the support to the operating units Is built. Since many of the Bank staff who provide support in the area of primary commodities also give attention to current and short-term market reviews, the Panel did not exclude this latter type of work from consideration. This report is thus concerned in principle with evaluating the thole range of the World Bank's economic analysis and research on primary comodity issues, and submitting recommendations. - 2- 2. Summary of Conclusions and Recommendations (i) Commodity work in the Bank is conducted by many units, serving different objectives in different environments. (ii) For operational purposes, the most decisive work is that done in central and regional project and program units, at project and country levels, in connection with project appraisals and balance of payments assessments respectively. (iii) A large degree of uniform discipline for this operational commodity analysis is provided by the Commodities Division's output of global commodity forecasts and through--to a lesser extent--individual commodity studies. (iv) The central commodity forecasts, for which the operational staff contribute data, are competently and efficiently prepared, mainly by conventional market judgements, assisted to a small but growing extent by relatively simple commodity models developed by the Commodities Division's own staff. The forecasting methods used compare favorably with those in use for comparable purposes in other organizations. (v) The Panel accepts the advantages of one central location for global commodity forecasting services in the Bank, while recognizing the need for a continued inflow of specialized commodity knowledge from certain central and regional units, especially as regards fertilizers, metals, minerals and energy. (vi) There are some problems of consistency in the adaptation by the operating staff of the central forecasts of international prices to the equivalent prices to be used at project and country levels. These problems should be largely resolved by the ongoing preparation, jointly by commodity and operational staff, of a Manual on the Use of Price Forecasts. (vii) Continuous improvement of the commodity forecasting should be based more on acquiring a deeper understanding of markets and their functioning than on further sophistication of formal modelling. The models being developed and improved by the Commodities Division should continue to be exposed periodically to review by commodity and modelling specialists outside the Bank, with a focus on improving the forecasts rather than the models alone. (viii) Long-term forecasts (five to ten years), however rationally based at the time of forecasting, are bound to be upset by subsequent unforeseeable events. Nevertheless, the reliability of the Bank's commodity forecasts should be continuously and routinely monitored to assess ranges of accuracy and to identify the most cost-effective ways in which to seek feasible improvements in reliability. -3- Short-term commodity forecasting (up to two years) is required o.th the Bank and the IMF. Since present arrangements for this work .ot seem to be satisfactory, it is suggested that the management of sank might examine at an appropriate time, with the management of the possibility of establishing a joint service in this area. The Commodities Division cannot, without a very considerable sion of staff, give effective assistance to all country and project who need market prospect information for specific types of products ,:-.cential exports from specific countries to specific market areas. *: -.culd be most cost-effective to purchase such information, when from outside sources, such as the UNCTAD/GATT Trade Centre. The full potential of the commodity data bank of the Commodities rson as a source of information for operating staff should be realized 41 rapidly as possible. Historical and current series on production, :a-e and prices should be added for commodities for which forecasting _ zcct considered worthwhile, including selected processed products not :read covered, and for selected representative types of project inputs. ::;uterizaticn should be extended and terminals installed throughout te Bank, so that one single set of consistent data in the system can be readily accessed directly. i) Since there is currently no comprehensive source of consistent --cernational data on metallic and mineral commodities, it is recommended the Bank request that the United Nations develop such a system of .ata as soon as practicable. If such a request cannot be met, the Bank :c consider undertaking this task itself. (xiii) It would be highly desirable if the commodity and country infcrmation assembled by operating units could be assimilated into the central data base--methods for achieving this on a continuing basis sculd be explored as a matter of urgency. (xi:) The commodity research initiated under Research Committee lseices, though highly professional and sophisticated, has not yet tace any significant contribution to the operational and policy roles cf the Bank, and the direction of the work still underway is hardly Zcre promising in this respect. However, the results of the commodity cdelling to date under RPO 671-09 "Natural Resources and Planning" Should be examined by central and operational staff to develop the pctential applications to Bank forecasting and development policy analysis. O.V) The Commodities Division should ensure servicing of all relevant Bank units with current market reports issued by the authori- tative commodity bodies outside the Bank, without attempting to duplicate -heir work, but drawing attention to developments significant to the Bank. -4- (xvi) With some promising exceptions, the Bank's in-depth studies of individual commodities and commodity groups tend to focus too exclusively on overall market factors and global supply, demand and prices, and thus do not yet display the world development orientation appropriate to the Bank's status and role as a major international lender for commodity production. (xvii) To promote the coherence and operational/policy value of the whole range of commodity services and research in the Bank, it is recoz- mended that an in-house commodities steering group be established. (xviii) Comparative advantages, requirements and opportunities in the Bank point to the following evolution of its commodity work: (a) the assembly and systematization of the Bank's own operation-derived data on commodity production and processing, including productivity, costs, competition, and trends,with a view to making them available for manipulation and use; and (b) utilization of such data, along with conventional international data - to improve commodity projection and forecasting work, especially on the production side; - to start a new generation of Bank commodity studies that concentrate, not on current market reviews, but on longer term factors and development issues relevant to . commodity lending policy thinking in the Bank; and . international commodity policy, including Bank advice to international commodity bodies, as already provided for by Executive Board decisions. (xix) In addition to continuing improvement and better operational and policy orientation of commodity work, four research projects are recommended: (a) the linking of cyclical and trend elements in forecasting the behavior of commodity markets; (b) the treatment of inflation in price forecasting; (c) the study of comparative costs and productivity in production of selected agricultural and mineral commodities, starting from available Bank information and research projects, with a view to feeding into Bank commodity policy and international commodity policy; and -5- (d) the study of factors affecting the location of processing of selected agricultural and mineral commodities, with a view to orienting Bank and international development policies. (X) It is estimated that the above research projects, which should be monitored by the proposed Bank commodity steering group, would require one additional staff research coordinator in the Commodities Division, plus some $70,000 of external funds per year for five years. (xxi) It is recommended that the proposed commodity steering group, or a similar representative group, should develop a more coherent and open system of publication for the Bank's output in the commodity field. -6- II. COM2MODITY WORK AND OUTPUT IN THE BANK 3. Location of Commodity Work Analytical studies and research with a commodity component are commonly carried out at the following locations in the Bank: - Commodities and Export Projections Division; - Other Development Policy units, including . other Divisions of the Economic Analysis and Projections Department . Policy Planning and Program Review Department . Development Economics Department . Development Research Center - Agriculture and Rural Development Department; - Industrial Projects Department; - Energy, Water and Telecommunications Department; - Regional Departments. (a) Commodities and Export Projections Division The specialist commodity staff of the Bank are concentrated in this Division of the Economic Analysis and Projections Department. 1/ The principal role of the Division--with the collaboration of other units-- is to develop and disseminate short-term (up to two years), medium-term (three to five years), and long-term forecasts (up to 15 years) of export prices and global export volumes for main agricultural commodities, metals and minerals, including petroleum and fertilizers. 2/ For some commodities, especially non-agricultural and fertilizers, essential data and analyses are contributed by more specialized units of the Bank, as indicated below. These long-term forecasts are used by the Projects Departments in assessing the commodity production loans of the Bank; and the short- and medium-term forecasts by the Programs Departments in assessing the 1/ The Division has a staff of 15 professional and 15 non-professional officers. This compares with 50 professional and 50 non-professional staff in the Commodities and Trade Division of FAO, and 48 professional staff (plus non-professionals) in the Commodities Division of UNCTAD. 2/ Analytical work is focussed mainly on the medium- and long-term. An informal understanding with the IMF recognizes the role of the Fund's research staff in sort-term projections. -7- ~aue of merchandise exports and balance of payments prospects of individual ember countries and hence their creditworthiness and eligibility for regu- lat or IDA financing. The forecasts are also used in the global model and the country models underlying the preparation of the Bank's World Development Report. A principal output of the Division is its annual report on Price ?rospects for Major Primary Commodities, issued in June. The 1978 report (814/78) contains, for each of 15 agricultural commodities or commodity groups, plus fertilizers, and for 8 metals and minerals (plus petroleum), historical and current data on world prices and export volumes, and annual orice and export volume forecasts up to the year 1990, together with a brief market analysis and methodological note for each commodity covered. Revisions of the price forecasts are undertaker. and disseminated as necessary. The staff of the Commodities Division maintain an ongoing aware- ness of developments in the commodity situation through the assembly of current information from trade literature, other organizations and personal contacts, and provide briefing for Bank management and project or sector MisSions as required. As a by-product of the computerized data base built up to serve its forecasting function, the Division distributes within the Bank, and outside on request, a monthly printout of computerized data on current commodity prices, For general reference, an annual summary of historical statistics of Commodity Trade and Price Trends (e.g., Report No. EC 166/78) is issued. The chief sources for information stored in the commodity data bank are FAO, UN, UNCTAD, Metallgesellschaft A.C., US Department of Agricul- ture and US Bureau of Mines. The range of commodities covered by the data files greatly exceeds those routinely monitored by the Commodities Division. The system can respond quickly to requests for basic commodity data and pro- vides a basis for analytical studies which require standardized data for :any commodities and countries. Many such studies are produced by the Division, to serve a variety of purposes. Some are designed to enlarge, for a particular commodity, the brief explanatory reference given in the annual report on price prospecta, They provide for Bank staff an analysis, in greater depth, of demand, supply and/or other factors affecting the market and the market prospects of the commodity or commodity group in question. Other studies are internal policy papers prepared specifically for the Bank's Executive Board, for the purpose of eliciting or orienting an operational policy decision by the Bank with respect to commodity lending, or simply as background information for the Board. Other reports again deal With general questions or problems in the international commodity field, for the information of Bank staff and/or for general circulation. -8- An important continuing activity of the Division is research aimed at the improvement of methodology in operational price forecasting and in commodity analysis generally, Currently the Division is not undertaking any research funded by the Bank's Research Committee. However, it does engay in methodological research out of its own budget. (b) Other Development Policy Units Other units within the Development Policy Staff may undertake studies of special aspects of commodity markets, e.g., economic rent in metals and minerals (Economic Analysis and Projections Department); or development issues emerging from commodity analysis, e.g., a perspective on the foodgrain situation in the poorest countries (Policy Planning and Program Review Department 1/), food security (Development Economics Depart- ment) and a study of the weakest commodities produced by the poorest regions (Development Economics Department). Most of the com.odity-related research in these units is funded through Research Committee allocations. Some is contracted to outside institutions. Much of it is not geared to current operational requirements, although some research in the Economic Analysis and Projections Department funded through the Committee is designed to develop methodologies for linkirz commodity forecasting with macroeconomic analysis for purposes of the Bank's world development reporting. (c) Agriculture and Rural Development Department The main focus of this Department--as of other central projects staff--is on projects in particular countries. It is a major user of the world price forecasts for agricultural commodities, from which it derives I prices appropriate to the country and project level, further differentiated by type and grade of commodity produced and transport costs. This adapta- tion is often assisted by the engagement of consultants in the context of project appraisals, to assess market prospects for the specific commodities to be produced in the particular project. Such commodity studies may be attached as market annexes to the projects as submitted for approval. Long-term market analysis is done for a few commodities not covered in the annual price prospects report of the Commodities Division, e.g., dairy products. The staff have recently prepared a major development policy study on tropical root crops. This unit also prepared sector policy papers, e.g., on forestry and fisheries, designed to influence the content and orientatios of project proposals and of Bank lending in the particular sector. 1/ Jointly with the Agriculture and Rural Development Department. -9 Industrial Projects Department Staff of this Department contribute to and utilize the long-term forecasts disseminated in the annual Bank report, and make the necessary aptations to specific project level. Consltants are often used to under- mke market and price analyses in connection with important loan projects, id a market annex is prepared and attached to each project submitted for Ipproval. For project purposes, market analyses are kept up internally for *Cent, steel and pulp and paper. The Department has a Research Unit on fertilizers, which maintains to overvfew of the world fertilizer situation and short- and long-term prospects, 'n collaboration with 1NIDO and FAO through the UNIDO/FAO/World Bank Ferti- -:er Wcrking Group. The specialist staff of the Department contribute basic material t tne Ccmmodities Di7vision for use in tb pparation of the short- and :cng-ter forecasts for fertilizers and a number of mineral commodities, ircluding copper, nickel, iron ore, steel and coal. (e) Energy, Water and Telecommunications Denartment In connection with its operating requirements, the Energy and ,;els wing of this Department maintains analtical studies on energy supplies ad forecasts for developing countries and con_zibutes these data to the odities Division for its work on energy. A more detailed discussion h the activity of the Energy wing is ava4il1be in the 1-yort of the External Research Panel on Energy, Water and Telecommunications. (f) Regional Departments Regional economists dealing with countries that depend heavily o exports of particular agricultural or mineral commodities have to acquire fairl% detailed knowledge of the world markets for those commodities in -er to handle their operational responsibilities. This is especially so ':en the country's or region's exports are a large proportion of world exports of a ccmmodity. Instances are Ghana for cocoa, Jamaica for bauxite, Chile and Z--bia for copper, Morocco for phosphates, Brazil for coffee, Saudi Arabia ICt petroleum, and South Asia for rice and jute. Such staff engage in short-term price and output forecasting for the commodities concerned. They develop an expert point of view in interpreting Using the forecasts of the Commodities Division and can make an important tribution of data to the central commodity analysis, especially on the "PPly side. Staff of regional departments also undertake commodity-oriented Studies at the national level, in connection with sector economic reviews Ot project development. Some such studies are prepared by specialists of Qe CoModities Division, seconded to participate in project or country ssios. Such studies are occasionally distributed as Bank documents, e.g., h Asian report on the world jute economy, but are more commonly to project and country mission reports. - 10 - (g) Overall Consistency With so many staff concerned with commodity data and their uses in so many different units, the problem of ensuring consistency is ever- present. The purpose of the centralized commodity forecasts distributed by the Commodities Division is to provide all the Bank staff with the basis for a uniform view of the prospects for internationally traded commodities. This means that all balance of payments assessments and loa proposals for expanding the producticn and export of any commodity in any region of the world can be tested against the same set of market forecasts. In the absence of such a global view of commodity prices, the Bank could not function convincingly as a world development agency. It is recognized that the kinds of commodity analyses and market studies that have to do with project evaluation and country program and balance of payment assessments in the substantive work of the Bank are the responsibility of project or regional staff, since generalized forecasts and generalized commodity knowledge cannot substitute for the specific knowledge acquired and needed at country and project levels. To promote two-way exchanges between commodity and other economists, a formal review process takes place in the early stages (April) of the preparation of the annual central commodity forecasts. Inconsistencies are most likely in the adaptation of global fore- casts to local conditions and in the evaluation of short-term market behavior. In the case of conflicting forecasts, the onus has been placed administratiVt on the project or country analyst to specify how his estimates are reconcil- able with the globally-oriented forecasts of the commodity specialist on the basis of differences in detailed commodity specifications, transport costs$ marketing arrangements or other such factors, - 11 - Research Committee Projects Since its inception in 1971, the central research program of the Bank, which is administered by the Research Committee has allocated total of some $350,000 to projects which can be said to be commodity- related. This sum amounts to three percent of the Research Committee's total allocations to date. This is not a large proportion. It should be related, however, to the fact that the Commodities and Export Projections Division, which carries out much of the commodity analysis and research, is fully funded by the regular budget of the Bank. Operational studies with a commodity character conducted by the central projects staff and the regional staff. as described above, are also done by regular Bank staff or else by con- sultants engaged under project budgets. The average annual expenditure by the Research Committee on comodity-related projects has amounted to about $50,000, exclusive of 1,orld Bank staff time. Quantitatively, the central research activities have so far been marginal to the Bank's work on commodities. The Research Committee's allocations in this field since 1971 are summarized below: Started Title Allocation Originator Complot on in $ 1972 Nov. Prospects for Jute and 15,000 EPDCE /1 Jai Competition from Synthetics 1973 July Comparative Analysis of 22,000 EPDCE /1 July 1974 Resource Allocation in Cocoa Production 1973 Oct. Oxford Seminar on Inflation 4,500 EPDCE /1 Oct. 1973 and Commodity Prices 1974 Natural Resources and Plan- 165,000 DRC,DED /2 to be ning - Issues in Trade & determined Investment 1975 Feb. Agricultural Commodity 50,000 EPD /3 Projections 1975 July Project Link 100,000 EPD /3 1976 Dec. Simulations of Buffer 38,000 DED /2 Dec. 1978 Stocks of Foodgrains /l Commodities Division. /2 Development Research Center and Development Economics Department. 3 Eo-nomic Analysis and Projections Department. - 12 - The Commodities Division has been modest in its claims for central research funds. It has not initiated any proposals since 1973 although it has been involved in the supervision of the Agricultural Commodity Projections and Link projects. No Research Committee funda have been allocated to operating units for commodity research. Most of the commodity-related research of the Development Economics Department and the Development Research Center in recent years, as referred to above, has been based on Research Committee allocations. All the projects originated in 1974 and subsequently have had a strong component of model-building methodology. Three of the project on Cocoa, Projections and Link, and parts of the one on Natural Resources--were contracted out to universities. -13- -voes of Research Output A listing of the main titles produced on commodity-related ects in all parts of the Bank over the last few years is shown in 1. These titles exclude the commodity and market annexes to r reports and project proposals, as well as the internal analyses of aerting units. Such output probably provides the most decisive kslytical support to actual Bank operations. The following categories of documents have been distinguished: Commodity Trends and Forecasts This is the smallest category but contains the most significant item--the annual report on Price Prospects for Major Primary Commodities. (b) Commodity Modelling/Methodology Studies 1/ In order to expose the working models being developed by the Co=mcdities Division to experts both inside an outside the Bank, three vol=zes of workshop discussion drafts on the World Bank Commodity Models have been prepared and circulated. There are also separate papers. describing the models being developed for various individual commodities. A considerable number of modelling studies have emerged from some of the Research Committee-financed projects, especially those on 'arUral Resources and Planning and or. Agricultural Commodity Projections. (c) Individual Commodity Studies 2/ These constitute the largest output group. Such studies mostly describe or update the individual commodity market situation, usually with a current or short-term focus, although a few are concerned with projections, market structure or development. Most of the documents in this group were prepared by the Commodities Division, but important papers have come forward from the Industrial Projects Department on fertilizers, from the Energy Division on energy supply, from the Agriculture and Rural Development Department and from Regional Departments. (d) Papers on Bank Lending Policy for Commodities This output 3/ consists of internal policy papers prepared for 1/ See Annex 1 B. 2/ See Annex 1 C. 3/ See Annex 1 D. - 14- submission to the Executive Directors of the Bank. On some subjects papers are intended to provide a basis for discussion and decisiont other cases, the purpose is simply to inform the Board so that it has background of information against which to assess policy issues and ic,.. proposals. Market impact was not a major factor in the appraisal of projects when the Bank's commodity loans were few and small. But fro the 1970s the relevant project appraisal reports carried a "market annex" outlining the staff's assessment of the market prospects for the commodity in question and the probable impact of Bank lending on the world market. In 1972 a practice was introduced of submitting to the Executive Directors a yearly review of the market prospects for particuI,, commodities. This review was to consider the probable impact of pro- jected Bank Group lending over the next two or three years, including t incremental output from projects previously financed. In this way the Bank moved into the phase of "commodity lending policy"--a phase which, despite some important initiatives, has so far been rather weakly developed. Not many such reviews have been put to the Board. Some of the commodity policy papers prepared for the Board have not been concerned with market factors, but rather with changing the focus of Bank lending, as in the case of the important papers on the * fertilizer requirements of developing countries, minerals and energy in developing countries, and the forestry sector. (e) General Papers on Commodity Policies and Problems This category comprises studies and documents dealing not with individual commodities or countries, but rather with general questions and problems in the international commodity field. 1/ Such output draws upon the individual commodity analyses and forecasts of the Bank staff to discuss and illuminate general commodity issues. Those papers dealing with policy subjects closest to the commodity area, such as commodity agreements and price stabilization, are prepared by staff of the Commodities Division. Commodity policy studies oriented toward general development issues have originated from staff in other parts of the Bank, particularly in the Development Economics Department, the Policy Planning and Program Review Department and the Agriculture and Rural Development Department. 1/ See Annex 1 E. - 15 - 6. Dissemination of Commodity Documents (a) To Operational and Policy Users in the Bank The output most specifically designed to be of use to opera- tional staff--the annual report on "Price Prospects for Major Primary Commodities"--has been disseminated as a printed "document of the World Bank," with a restricted distribution and for official use only. It is readily available to all operational users in the Bank. The Panel believes that on the whole the output is widely appreciated and widely used by Bank staff, subject to the qualifications mentioned in the next section. The decision has recently been taken to make the annual Price Prospects report a declassified publication. The commodity policy papers prepared for top management or the Executive Directors are disseminated internally as restricted papers. Even after the policy decisions have been taken, the policies as deter- mined do not seem to gain wider distribution in any formal way. The other types of commodity studies are widely disseminated in the Bank and are therefore adequately available for use. Because they do not often originate from requests from operational staff, their assimi- lation will therefore depend on whether staff have a felt need for the information at the time it comes to hand or soon afterwards. (b) To Research Colleagues Inside and Outside the Bank This purpose is served by the circulation of preliminary drafts, informal notes or staff working papers. The comments received contribute to the preparation of more definitive texts, which may qualify for more formal methods of dissemination. (c) To Governments, International Bodies and Technical and Academic Institutions Disseminated to this group are World Bank Documents, World Bank Staff Commodity Papers or World Bank Staff Working Papers. Member governments of the Bank automatically receive copies of such papers through their Washington representatives, but it is unlikely that the documents move into the hands of all potential users in those governments, especially in the developing countries. Distribution to international technical and academic institu- tions is usually at the initiative of the producing unit, which has a selective list. For the most part the lists are too small. Such institutions and the interested public may obtain such documents on request if they are included in the World Bank Catalog of Publications free list. Thirteen of the 60 papers listed in Annex 1 are referred to in the 1978 issue of the Catalog. - 16 - (d) To the General Public Access of the general public to the analytical studies and research done in the Bank is limited to those papers included in the World Bank Catalog of Publications, under the free list or as publicr- tions for sale. Only three commodity studies are as yet listed as publications for sale. These enjoy the commercial promotion efforts of the publisher. (e) General Comments It is evident that the dissemination of commodity documentatio is an unsolved problem. The impact of the Bank's commodity analysis and research in the outside world, including member governments, is much =ore limited than is warranted by the quality of the work. Moreover, there is no easy way to discover the full range of the Bank's output in this :ield. The Commodities Division has made progress in clarifying its own policy by establishing a Commodity Note series for papers of a pre- liminary or exploratory nature, and a Staff Commodity Paper series for documents in more finished form. This framework does not bring any order to the identification or dissemination of commodity-related papers prepared in other units. Those that pass the draft stage are distributed as Staff Working Papers, whether designated for circulation only inside the Bank, or to outside institutions and the public as well. As a guide to dissemination policy, it would seem desirable to establish the main purpose to be served by each paper. If the aim is to inform Bank staff and seek reactions from them, restricted internal circulation would appear appropriate. Where the objective is to influ- ence governments and inform public opinion or contribute to international debate, a much more open and extensive system of distribution is called for including commercial publication on a more liberal scale. Also, much more coherence would be attained if the organization of Commodity Notes and Papers could be applied to all commodity-related documents, wherever prepared in the Bank. An important method of dissemination for some purposes is the workshop or seminar, usually based on a preliminary draft text. This method is useful for obtaining the reactions of a wide range of partici- pants, with a view to the preparation of a more definitive piece of analysis or research. National or regional seminars based on an accepted Bank position paper could in some instances be a useful way of dissemina- tion to the appropriate government circles. - 17 - III. PANEL'S EVALUATION OF WORLD BANK'S ANALYSIS AND RESEARCH ON COMMODITIES 7. Objectives The Panel was informed that the overall objectives of Bank research were defined as follows: - to support all aspects of Bank operations, including the assessment of development progress in member countries; - to broaden understanding of the development process; - to improve the Bank's capacity to give policy advice to its members; - to assist in developing indigenous research capacity in member countries. Although the above objectives are specified in relation to re- search, the Panel has used them as points of reference when considering the whole range of commodity work. Since the Bank's commodity work is predominantly a regular staff function, it seems appropriate to give most weight when evaluating this work to the first and third objectives, namely supporting Bank operations and improving the Bank's capacity to give policy advice to its members. However, all of the above general objectives are relevant for assessing the Bank's economic analysis and research in the commodity field, especially when account is taken of the centrally funded research projects. Before examining the commodity work in relation to the objec- tives, the Panel presents its comments and recommendations regarding the technical aspects of this work. Evaluation by an outside Panel has to be based on the circulated documents, supplemented by a limited amount of information obtained from personal interviews with Bank staff. The crucial internal commodity analyses undertaken by operational staff and consultants at project and country level, based on Bank and non-Bank sources of data, could not be evaluated by the Panel. - 18 - 8. Technical Aspects of Bank's Commodity Work (a) The Data Base For agricultural commodities and forest products, the Bank's data base is probably the best that can be assembled at present. The co- operative arrangement established with FAO feeds in the historical and current statistical series of that Organization on production, area har- vested, exports and imports (both volume and value) of all agricultural commodities and forest products for all countries, the series having already been checked in FAO for internal consistency. Yields where relevant, as well as export and import unit values, can easily be deri-ed from the basic FAO data. Basic data are also obtained from specialized international commodity bodies and the US Department of Agriculture. For minerals and metals there is no clearly authoritative inter- national source. The Bank draws statistical material from many agencies and has great difficulty in trying to establish a usable set of compre- hensive data, consistent as far as possible internally and capable of being fed with current data from the best outside authorities. This task is of fundamental importance, for the international community as well as for the Bank. Since the United Nations has the responsibility for primary data collection in this field, the Panel recommends that the Bank should request that body to undertake the basic task of constructing an inter- linking system of consistent and comprehensive international statistical data on metallic and mineral commodities. If the United Nations is unable to respond to such a request within a reasonable time, the Bank should consider doing the work itself. In the energy field, the Bank derives historical data for the developing countries and the centrally planned economies from the UN energy tape, using the UN conversion factors. For industrial coun- tries, OECD energy statistics are used. The Bank (Energy and Fuels wing) generates projected production data for developing countries. The organization of a comprehensive set of consistent statistical data is not yet completed. As regards computerization of the commodity data base, much progress has been made, but much more remains to be achieved. So far only the FAO tapes, the UN energy tapes and some USDA tapes can be in- corporated directly into the Bank's computerized system. However, by August 1979 a compact, UNSO Geneva version of the UN Series D trade data is expected to be available in machine-readable form. 1/ 1/ This will provide detailed value and quantity data for all reporting countries annually for the period 1962-1978. - 19 - The Panel recommends that more priority should be given--in the Commodities Division and other concerned units--to speeding up the establishment of a complete and fully computerized commodity data system in the Bank. Similarly, more emphasis should be placed on providing ter- .inals for all user units in the Bank, so that they may have immediate direct access to the data in the system. Within the context of this program, means should be found to integrate into the central data base such suitable data as can be made available on a regular basis from the operating units. As far as possible there should be only one set of consistent commodity data available on computer terminals to all Bank staff. Efforts are being made to link the commodity data and projec- tions to the country macroeconomic data and projections used by Bank regional staff. The Panel believes that this is a necessary development. (b) Knowledge of the Markets The Panel considered that the efforts undertaken by relevant Bank staff to maintain an up-to-date knowledge of the commodity markets were successful. Commodities Division staff had, perforce, to concentrate on world market developments by maintaining current statistics, following the trade and professional literature, keeping contact with the secreta- riats of the international bodies specializing in particular commodities, and attending the periodic sessions of all the intergovernmental co-modity councils and study groups. This concentration on the world markets necessarily meant that Commodities Division staff had rather superficial knowledge concerning the markets facing individual countries, apart from the major producing and exporting countries. With respect to smaller and less important trading countries, country economists and project staff often need more detailed market information than the Commodities Division is geared to supply. A considerable increase in staff would be needed to provide, within the Bank, for a full service of information on potential market prospects for specific types of products as potential exports from specific countries to particular market areas. The Panel considers that it would be most cost-effective to purchase such information from outside agencies, such as the International Trade Center of UNCTAD/GATT. On the whole, the Panel felt that the commodity documents of the Bank reflected an adequate knowledge of world markets. At the same time, it stresses the importance for the specialized commodity staff to deepen progressively their understanding of the markets in areas such as the structure of the markets, the nature of the competition, the dynamic aspects, the policy constraints, and the development implications of these factors. The convenience of computerized statistical data series should not be allowed to induce an undue dependence on mechanical com- Pilations, to the detriment of a constant search for better understanding - 20 - of the statistical series, which is the key to the market and develc-e: judgements which the Bank recuires. (c) Problems of Price Forecasting In its annual report on price prospects the Bank presents medium- and long-term price forecasts in both constant and current TS dollars. It has been generally accepted that the short- and medium-term projections, covering up to five years ahead, should take into acco,.n- cyclical and other short-term influences incl-ding market disequilibria, whereas the long-term, covering up to 15 years ahead, should be a m::=: solely of trend projections, reflecting assessments of the fundamental structure of commodity markets. There is a general impression that Bank's short-term forecasting--as distinct from its long-term analysis-- is rather an ad hoc exercise, dependent mainl on market judgements :ase: on the work of the international commodity secretariats, with some e in from operational staff and some unorganized consultation with the research staff of the IMF. It is a question whether the Bank should attempt short-tern price forecasts at all and rely entirely on the IMF in this field, in which the I1F is the recognized international specialist, or purchase needed reports from commercial sources. The fact that the theoretica division of labor between the Bank and the IMF does not seem to be effective operationally in this area suggests the need for clarificatf:n at the policv le-els of the two agencies. If the operational requiret of the two bodies for short-term commodity forecasts do not coincide t are nevertheless complementary, it is possible that both agencies could be more effectively served by a joint staff unit working in this field. The Panel therefore recommends that the management of the Bank consider consulting with the management of the IMF to examine the case for establishing a joint service. Serious attention should be given to the methodology of linking cyclical and trend movements in the forecast price series. This is a matter calling for basic research, and the Panel strongly recommends that it should be initiated by the Bank as soon as possible. Another major methodological problem is the treatment of in- flation in commodity price forecasts, or the relation between forecasts in constant and current dollars. In the Bank the index of international inflation is used as the deflator for the prices of primary commodities exported from developing countries. This index, which is calculated in and projected by the International Economy Division of the Bank, reflects movements in the CIF US dollar prices of developed countries' manufac- tured exports to developing countries. If short-term prices are forecast in current dollars and con- verted into real terms by such a projected deflator, any inaccuracies in the original forecasts are accentuated. The same result obtains if--as - 21 - in long-term projections--the commodity price is forecast originally in constant-currency terms and then "inflated" by the index. There is the additional problem that the chosen inflation index is not equally appro- priate for all commodities. With Research Committee assistance, the Bank sponsored a conference at Oxford, England in 1973 on the impact of inflation in industrialized countries on commodity prices. The Panel recommends that the suggestions of this Conference be made the starting point for further research in this area. Basic issues to be explored are the differential behavior of various commodity prices under different rates of inflation at different times and places, and what better inflation index or indexes to use and how to project them. From the operational point of view in the Bank, a further technical problem is the relation of the projected international indicator prices to the prices of particular grades, in different markets and at different stages of processing, which are required by users. In general this is an appropriate subject for econometric analysis using established techniques. The relationships are currently arrived at by operating staff, with some assistance at times from commodities staff. Some of the rela- tionships may indeed be highly specific to individual cases, especially for mineral commodities. A further question is how to value the output of non-traded commodities in the economic analysis of production projects. This whole group of problems is extensively discussed in a manual 1/ currently being drafted in the Bank. The dialogue among commodities and operating staff involved in the finalization of this manual should greatly assist the development of consistent methodologies and identify any issues calling for further analysis or research. (d) Reliability of Price Forecasts Commodity forecasts and the techniques embodied in them can best be judged by comparing them with the reality after the event. This too, is a good method through which to determine the most cost-effective ways for improving reliability. The question of reliability raises others of operational significance. Have the operating decisions of the Bank staff based on the forecasts proven correct? In other words, have the assessments that certain countries were creditworthy and others not, based on balance of payments prospects, turned out to be good operating decisions? And have the production loans justified by reference to the forecasts of commodity prices and markets proved in hindsight to be sound? 1/ Manual on the Use of Commodity Price Forecasts, Robert Bacon, Commodities Division (draft). - 22 - The Panel has not seen any systematic review or evidence on these points. However, since the first meeting of the Panel, the nrk has commissioned a study which makes a beginning in the development of techniques for the measurement of reliability, although so far it reviews only forecasts up to four years ahead. It is too early to give = evaluation of this study, but the general findings leave no room Zor com- placency. 1/ The Bank's own recent audits of project performance indicate that, in general, price forecasts for agricultural commodities an some minerals have turned out to be conservative in that actual prices realized in production projects exceeded those anticipated at the app'raisal stage. 2/ For some loans this had meant that the rate of return achieved had been greater than that on which the project had been approvec. In other cases the higher level of actual prices had floated the prcfe:ts above the negative effects in other components, where actual res2:s proved more unfavorable than projected, e.g., slower increase in crop yields, higher cost of inputs, less effective management. On this evidence, the conservative price forecasts have favored the Bank as a lender. On the other hand, they may have caused the abandonment of some loan proposals which could have proved beneficial to the potential borrowing countries. One would expect that during certain periods price forecasting would tend to be optimistic instead of c.n- servative. The Bank staff recognize that the price forecasts have :o 'e revised periodically, and, if the elements of change are considered to have long-term effects, revisions are prepared and disseminated. Occasional subsequent revisions in long-term forecasts that appear rational at the time of projection are inevitable. The Panel strongly recommends that the Bank should continue the review begun in the Ghose study, with a view to instituting as soon as possible a routine and continuing monitoring system for its commoditV forecasts, designed to point to the most effective and efficient improve- ments in market analysis and projections methodology. 1/ Price Forecasts for Primary Commodities: An Evaluation of Past Experience, A.K. Chose, October 1978. This study indicates that the Bank's near-term price forecasts are of doubtful reliability, but better than forecasting no change. 2/ Annual Review of Project Performance Audit Results, World Bank, November 1978, paragraphs 2.11, 3.41 and 3.196. - 23 - (e) The Use of Formal Models (i) Recent History The Bank's work on commodity models has developed in recent years in two main locations--the Development Research Center and the Commodities Division. 1/ In the last two to three years, the Commodities Division has been trying to evolve models desLgned to systematize the data and rela- tionships in individual commodity markets. So far, models have been built by this unit for eleven commoditis or commodity groups. This is essentially research work, but it has been funded predominantly within the ordinary budget of the Division, except that the jute model has been taken over from RPO project 671-09. With central research funds Mde available by the Research Committee, under RPO 671-09, the Development Research Center sponsored the preparation of a world copper modeL, a world aluminum model and a dynamic simulation model of the world jute economy. Work was also initiated on energy and pulp and paper industry models. The objectives of this research were to "provLde the analyt[cal tools and data format to analyze global supply and demand conditions of important commodities... to determine globally optimal output, inve!stment and trade patterns... to assess the desirability and FeasLhtlity of collusive action by developing country producrs...." This RPO project waS very amibLtious. The anticipated amount of consultant contributions did not materialize, and as a result, after several years of research, none of the models seems to have proceeded beyond initial versions, and final reports are not yet available. Two other research projects funded by the Research Committee have been concerned with commodity models, namely RPO 671-23: Agricul- tural Commodity Projections, and RPO 671-28: Project Link. Both projects were initiated in 1975 and are overdue for completion. Project RPO 671-23 is being conducted at the University of Illinois, with the support of the Ford Foundation as well as the World Bank. The project aims at construct- ing a world model of grains, soybeans, cotton, beef, pork and poultry for 20 countries or regions, designed for rational planning of the world food economy. Project RPO 671-28 was mainly designed to provide Bank financial support to Project Link, which is an econometric system of the world economy based on an international cooperative research program centered at the University of Pennsylvania. The Link Model encompasses three general commodity models provided by the Bank; project RPO 671-28 will expand those to twenty. 1/ For documents produced, see Annex I i, -24- (ii) Technical Evaluation of the Models Formal commodity models as such play a relatively small role in the routine work of the Bank. Out of 34 commodities for which the Bank publishes regular price forecasts, there are not more than eight for which a model exists which is written up in final form and could, in principle, be used for generating price forecasts. These eight models and a number of others at an early stage of develooment, have been constructed as a sideline activity by the staff members of the Comm:i Division. The table below summarizes :he situaticon as of May 1979. will be noted that as yet only two have been cleared for publication. COMM2ODITY MODELS IN THE COMODITIES DIVISION Status Model Final Runs In Final Published as Incomplete Complete Draft Working Paper Coffee X X Sugar I X X Sugar II X Bananas X Cocoa X X Fats and Oils X X -ea X Rubber X No. 3 Jute X X Iron/Steel X X Tin X No. 1 TOTAL: 11 3 8 6 2 The other models produced from Research Committee funds as referred to above were not specifically aimed at producing price fore- casts and, in contrast to the Commodities Division's models, are the out- come of major research efforts. The Panel has reviewed the majority of both groups of models individually, 1/ and in this section of the report sets out its general comments and conclusions. The Commodities Division group of models is the simpler to consider first, since these models are all aimed at the same general objective, namely, to raise the quality of the operational work of the Bank, particularly the routine work of price forecasting. This general 1/ See Annex 2. -25- objective can be broken down into the following particular objectives: - to summarize the essential relationships which have existed to date in a particular commodity market, i.e., supply, demand and price relationships, and to test the stability of these; and - either to create a mechanism capable of generating a future time series of world prices--a mechanism which can easily be updated; or - in certain cases, to provide a facility for studying the effects of future policy alternatives, such as bufferstocks or commodity agreements. Most of the Commodities Division models constructed in the last two years are designed to meet the first two of the above three objectives, and on average each has taken about three man-months to construct. They have all but one (iron and steel), the same basic structure. Supply equations are constructed from annual time series for major producing countries or regions; demand equations are similarly constructed for the main consuming countries or regions; and the model is closed by a price equation, in which an important element is the level of world stocks. The supply equations usually contain lagged values of the dependent variable on the right-hand side, so that estimates of both short- and long-run elasticities can be made. These models are fairly standard and contain no innovative elements. They are at least as good, in general, as comparable mcdels used in UNCTAD or FAO, thou:h the two latter agencies are not involved in long-term price projections. Being rela- tively simple in character and not very demanding as regards data, they meet the criterion that they can be easily updated by staff members who have not too much time to spare. In fact, the Division must be compli- mented on its overall productivity. From a technical point of view, the main weaknesses of the internal model-building are: (a) the models contain no cyclical elements; even though the intention is to project long-term trends, it is often necessary to take account of cyclical movements during the historic period over which parameters have been estimated; (b) there is too easy an acceptance of a particular hypothesis (Nerlove) about the lag structure in supply. Since this leads to the inclusion of the lagged value of the dependent variable on the right-hand side, high correlations are obtained which tell one little about the structural relations; (c) there is not always sufficient recognition of previous work done on the commodity outside the Bank; the Panel recommends that a literature survey paper should always precede the construction of a model; - 26 - (d) the parameter estimates have not been sufficiently critically examined to explain differences between countries and to look for possible changeover time; and (e) more work should be done (as has been done for meat 1/) on the relationship between the chosen indicator prices and other important prices (e.g., for different grades of the same commodity). The Panel is glad to note that it is planned to expose all the models at an international seminar in 1979. This type of exposition should take place, say, at two-year intervals. Two of the models (jute and sugar) are aimed at policy simulation, that is at the first and third of the above objectives, though the struc- ture of the models is not dissimilar from those described. The jute model is innovative in that it attempts to allow for the effect on supply-of a reduction in uncertainty caused by any success attained by price stabili- zation schemes. This device could profitably be tried for other commodi- ties. One model (iron and steel) is entirely unconventional in that it uses a programming technique for the purpose of price projection. This approach has to be regarded as experimental, as it raises both con- ceptual and empirical problems, and its forecasting power has not yet been demonstrated. In general, the Fanel's view is that model building by the Commodities Division on about the current scale should continue to be encouraged, even though there is no chance that a formal model can ever reliably take over the complete task of price forecasting. It is an educative process and a good way of keeping upcontacts with professional opinion outside the Bank. It is also a way in which subjective hypotheses of past or future market phenomena can be tested in a consistent quanti- tative framework. Aluminum and copper models have been substantially completed on Research Committee Funds. These are aimed primarily at the problem of optimum investment distribution between producing countries. The effort has been put almost entirely into the supply side, and the models contain a wealth of data on technology and costs. These models are difficult to test by their nature, and in any case little attention has been given to the demand side, something which will be essential if the models are ever used operationally. Another problem is that models of this kind have a high rate of obsolescence if not kept in use, and it is unfortu- nately easy for them to become museum pieces if there are not sufficient 1/ Reference: Annex 1 C, 11. - 27 - internal resources to keep them up-to-date. Particularly in the field of metals and minerals, however, this type of model building is likely to be necessary if commodity work is to be used in forming Bank policy. In view of their high cost, the Panel has reservations about some of the other projects which have been supported by Research Committee funds, if the criterion is their usefulness to the work of the Bank. For example, it is very difficult to see any likely opera- tional payoff in commodity work to the $100,000 granted to the study of the linkage of commodity and country models (Project Link). The project "Natural Resources and Planning" ($165,000) has so far produced the studies on jute, aluminum and copper already dis- cussed, but the prognosis for work on a world food model (second report, 1977 SR 671-23) is very poor. A number of quite different studies have come under the general heading of "Natural Resources and Planning" which seem to the Panel to be much too wide and not well-defined. The Panel recommends that studies financed by the Bank but carried out by consultants should be confined to: (a) commissioned papers on the state of the art, literature and institutions active in a particular field; and (b) the establishment of new types of methodology in commodity market analysis not so far developed in the Bank (the copper and aluminum models are examples), provided that the terms of reference of the study are closely defined; that the progress of the study is monitored regularly; and that the Bank is satisfied that at the conclusion of the project, it possesses the resources to take over the new methodology for operational purposes. (iii) More General Comment In the view of the Panel there is no set of obviously superior techniques relevant to the Bank's specific needs on which the Bank's staff are seriously missing out. It is obviously possible to cite cases where more mathematically sophisticated techniques have been used else- where, without, however, guaranteeing that their use in the Bank would produce more reliable forecasts or policy simulations. The Panel considers, therefore, that whatever the specific techniques used for forecasting, the principal basis must be not methodological excellence, nor necessarily the use of the most "advanced" techniques, but an in-depth understanding of the relevant relationships within the commodity sector and the economic environmeht which affects it. If, because of resource constraints, a choice has to be made between the attention the commodity specialists give to the development of methodology and to their maintaining familiarity with the changing - 28 - economics and institutions that characterize their industries, the Panel would give preference to the latter. In practice, it is hoped that improvements are attainable in both respects. To this end, the Panel recommends that efforts be continued in the Bank, designed to improve the commodity models available, both formal and non-formal, and adapt them to the operational and policy re- quirements of the Bank. For some time to come, the major use will be in forecasting, and most of the improvements and modifications will be made by che Commodities Division staff in the course of the regular fore- casting exercise. However, the jute and sugar models, which are aimed at policy simulation, and the aluminum and copper models, which are aimed at optimal investment allocation, make an extension of the Bank's model building from the traditional task of price forecasting. The Panel welcomes this extension as a move toward the exploration of the capability, of formal models in the area of general commodity policy, particularl- development policy. It also welcomes the move toward the integration of commodi:v models with global macroeconomic models. Though it is a wider issue which cannot be treated in this Report, the Panel wishes to note that there are many aspects of development policy in which commodity models are potentially useful. These include such issues as income distribution and employment. - 29 - 9. Supporting Bank Operations This objective is seen to have two aspects. The first is the bread-and-butter task of serving the daily operational requirements of the Bank. The other and perhaps increasingly significant task is orienting the Bank's operations in a policy issue, in the light of global trade and development issues that are or can be illuminated by commodity analysis and research. These questions of commodity lending policy will be taken up in the next chapter. An essential requirement for the Bank, as already indicated, is that all its operational perspectives at any one time--for all countries and projects--in which commodity exports or imports feature, should be based on one view of the world market for each (and preferably all) of the commodities concerned, and that this view should reflect the most authoritative judgements available inside and outside the Bank. By and large the Bank's commodity analysis, undertaken in various units, enables this essential requirement to be met, but with many qualifications and reservations. These relate to: - the technical problems of price forecasting; - the coverage of commodities; - the definition of commodities; - processed commodity exports; - lack of comparable data on input costs; - the "spot" nature of the price forecasts; and - the aggregate nature of central commodity studies. (a) Technical Problems of Price Forecasting These have already been discussed in the previous chapter, in- cluding their impact on operational work through effects on the reliability of price forecasts. The Panel considers that the routine monitoring of price forecasts in hindsight, which is strongly recommended, should be extended to cover the impact on operational decisions of fore- cast prices which turned out to be above or below the eventual realiza- tions (b) Commodity Coverage The traded commodities not dealt with in the 1978 Price Pros- pects report include dairy products, sheep meat, hard fibers, pepper, hides and skins, wool, forestry products other than tropical hardwood, - 30 - fishery products, plus a number of mineral commodities including cobalt, chromium, uranium and coal. For these commodities, the staff of the regional or program units concerned have to make their own market fore- casts, when required. For some of these commodities, the volume of Bank business may not warrant centralized analysis and forecasting. However, to provide centralized assistance to operational staff concerned, the Panel recom- mends that the coverage of the central commodity data bank be gradually extended, with the collab oration of interested units, to include a miniiniu of historical and current statistical material on a selected number of such commodities, to be accessed through computer terminals in the Bank. (c) Commodity Definition The adjustment of the standard international price forecasts to the actual qualities and types of commodities exported or imported' by a particular country is not necessarily done uniformly by all groups of operational staff, causing the rigor of the Bank's standard commodity price forecasts to be weakened in practice. A similar problem arises in converting the world market prices, as forecast, to realizable prices at the border of the country under study or at the specific project site. For this conversion, all the operational units have to make their own estimates of transport costs, in the absence of centralized data in the Bank on international freight rates. This is a very difficult area for systematic data collection, but the Panel recommends that consideration be given to initiating the assembly and dissemination in the Bank of a minimum amount of transport cost estimates for major commodity trade flows. As mentioned above, a manual on the use of commodity price forecasts is now under preparation, which should help to improve the consistency of the operational adjustments. (d) Processed Commodity Exports Countries which have generated a certain momentum of economic development based on the export of primary commodities normally seek to diversify their economies by industrialization. This leads usually to the expansion of exports of commodities in processed or more refined forms. To the extent that the Bank issues standard price forecasts for primary commodities only, the Bank's operational staff working in such countries derive limited assistance. For the most part, this is not - 31 - a problem for the metallic minerals, for most of which the price fore- casts relate to the refined metal, a major exception being iron ore. 1/ The Panel recognizes the major problems involved in trying to establish standard price series for processed commodities, and feels that pricing in this area should continue to rest with regional or central operating staff, until such time as Bank studies in the field of proces- sing enable some practical possibilities for additional central services in this area to be identified. (e) Lack of Comparable Input Cost Data The costs of many internationally traded inputs into production projects have been rising rapidly in recent years. For the operational staff concerned with such projects, price forecasts for the commodities to be produced are only a part of their data requirements. Estimates and forecasts of prices of major imported inputs are increasingly signifi- cant in project appraisals. If there is serious lack of consistency in such estimates by different regional units, the uniformity of price forecasts for the corresponding output commodity may become irrelevant. The Bank does make and disseminate forecasts of fertilizer prices. Other important inputs that are commonly imported for projects include insecticides, pesticides, tractors, earth-moving machinery, cement, reinforcing steel, structural steel and construction equipment. Many of these items are highly specialized, more so than refined com- modity products, but markets analyses are maintained for a number of such projects in connection with industrial project operations. The Panel suggests that the Bank explore the possibility of establishing some useful standard cost estimates and projections. In any case, historical and current price series of some standard descriptions could perhaps be introduced into the central data base, for access through terminals in the Bank, if there is a widespread need. Beyond this, it might be feasible for appropriate units in the Bank to prepare and disseminate monographs on the major input industries, describing world structure and functioning of the industry, major pro- ducts, developments and price trends. Such monographs could be extremely useful, not only for project design and appraisal in the Bank, but also in other development agencies and national authorities. 1/ Iron ore presents special complications, since the processed forms (pig iron, steel ingot, etc.) usually do not move much in interna- tional trade before conversion into rolling mill shapes and forms. For these, a single price series is almost meaningless, but some sort of substitute international indicator might usefully be developed. - 32 - (f) "Spot" Nature of Price Forecasts Some users have reservations about the realism of a single "spot" price forecast for a commodity some 10 years ahead, and find difficulty in applying it to operational decisions. On the other hand, forecasts of a price range would open up too wide a band of interpreta- tions and practical adjustments. The Panel feels that it would be a useful advance, not involv- ing the development of econometric technique, to increase the awareness of users as to the sensitivity of the forecasts to the major assumptions made in their formulation. In particular, it would be helpful to indicate the sensitivity of demand for agricultural raw materials and minerals to assumptions concerning future growth of GDP or industrial production in OECD member countries. (g) Aggregate Nature of Central Commodity Studies It has been mentioned earlier that the Commodities Division can hardly be geared up to handle all types of detailed market surveys, in addition to covering world market developments adequately. The re- quirements of the operating staff for such specific surveys which cannot be handled by the central commodities unit will have to be purchased outside the Bank. - 33 - 10. Ori_ntinLg Commodity Lending Policy The previous discussion has been concerned with the use in the ank of commodity price and export volume forecasts in assessing the ,reditworthiness of countries and the economic viability of individual :roduction projects. This section raises another set of issues which 3re relevant for the production operations of the Bank, but in terms of broader development orientation and policy. (a) The Project Approach The traditional Bank approach is that any production project that is viable in its national context should be assisted by the Bank. With the strong ?ressure within the Bank to expand its operations as fast and effi- ciently as possible in order to play a full role in speeding up the develop- rent of countries and people who are lagging behind in the advance of mankind, this approach of supporting individual projects has many supporters. In fact, it is probably the dominant approach and helps to account for the strong forward drive in the operational activities of the Bank. In the earlier years of the Bank, the choice of prices for valuing the commodity output of projects was fairly haphazard, and market prospects were assessed by project analysts. The analysis of projects has been evolv- ing toward a "world market" approach, with the establishment of centralized long-term commodity market forecasts, including price forecasts. This service has recuired all operating staff to appraise production projects against price forecasts that reflect the best estimates of the future evolution of world markets. If the forecasts show stable or rising world prices due to strong or excess demand, this favors the search for and justification of new pro- jects for increasing production of the commodity concerned. On the contrary, a forecast of falling world prices due to prospective oversupply, directs the attention of operating staff (and governments) to new production pro- jects only under conditions where the investment can be justified even at the low prices forecast. Despite this important evolution, the project approach remained basically the same, with operational interest focussed on the national impact of the project. This approach, particularly appropriate in the early years of the Bank's work, remains appropriate so long as the impact of the national Project on other member countries or on the world market can be regarded as negligible. (b) Development Policy Approach for Commodities The fact is that this price-based method of evaluation has a number of drawbacks. While adequate for individual projects, it ignores - 34 - the potential significant cumulative effect Bank projects for a single commodity may have on the world price, especially if demand for the commodity is inelastic as for coffee and tea. In such cases the project will redistribute trade shares and therefore income between countries. The income redistribution may be desirable or undesirable in itself or on grounds of comparative advantage, and these could be regarded as factors to be considered in evaluating a production project facing inelastic world demand for the product. There is another type of situation where price in the inter- national market has some drawbacks as a criterion for the Bank's lending operations in developing countries. This can be argued for instance, in the case of sugar and dairy products, for which the international prices are depressed--sometimes below the cost of production of low-cost exporters --by overproduction in protected markets of some developed countries and the export of their surpluses at almost dumping prices. If such low international prices are judged as likely to continue over the long term, it may be considered sound lending policy to refrain from investing in the production of these commodities in developing countries, if the investment could not be justified against imports at such prices. But would it be good international development policy for the Bank? Is this a relevant question? The Panel believes that it is, while recognizing the almost intractable nature of the problem which it poses for the Bank, especially regarding commodities of which some of its major donors are important producers or subsidizing exporters. The issues raised by these drawbacks of the project approach can be comprehended under commodity lending policy, but they are so basic as to be considered part of development policy. Since it is well accepted that the Bank should, in the future, seek to justify projects in the wider framework of development, where world market impact is only one factor, this shift of emphasis seems both appropriate and hopeful. Other criteria in Bank lending which are currently in use in evaluating projects and which should be viewed in this broader context include: - employment promotion effect; - income effect on poverty sectors; - foreign exchange earning effect in low-income countries; - diversification of production; - institution-building, e.g., marketing, cooperatives, credit. In such a wider calculus, the world market effects of commodity production elements in individual projects could be brought into the evaluation of the project. It is beyond doubt that the Bank's production loans to certain countries do, under certain circumstances, have important development implications for other member countries. Thus, Bank loans for expanding - 35 - steel production in developing countries have led to adjustments in older producing countries. Likewise, the assistance given by the Bank group in the 1960s for the shift in coffee and tea expansion in African countries benefited these countries at the expense of Latin American ad Asian producers respectively. Similarly, the policy of restricted ending for sugar and citrus expansion favors the existing producers and exporters at the expense of potential exporters. This is indeed a problem area for the Bank. But with the rapid growth of the Bank's operations, both in commodity production and commodity ,rocessing, and especially in agriculture, the problem will not go away. in fact, it is bound to assume greater proportions. Since the Bank is a world development agency committed to the application of objective economic criteria in all its judgements, it cannot ignore the need to evolve well- considered commodity and development policies in conformity with its role and status. "It is a fact that the international economy is growing more interdependent. That evolution can and should benefit developing and developed countries alike, but if it is to do so there must be adjustments in the global pattern of trade to reflect shifts in comparative advantage. these adjustments will not be easy, but the alternative to a more rational economic framework can only mean, in the end, greater penalties for all." 11 There is an important commodity dimension to interdependence a-d international adjustment, to which the Bank will be called upon in the future to give consideration. At the very least, it should be under- taking the analyses and research which will enable it to justify what it is doing, and provide a foundation upon which to build its policies for the future. (c) Bank's Commodity Lending Policies The issue of global commodity interdependence was first brought to the Board's attention in January 1973, in a paper on Development Policy for Countries Highly Dependent on Exports of Primary Products. 2/ Starting from market analyses, the Commodities Division paper focussed on the slow rates of growth in value of exports from developing countries of commodities Zacing inelastic demand. For such commodities--in particular coffee, cocoa, tea, bananas, bauxite/alumina/aluminum and tin--the staff recognized clear dangers that Bank lending to increase production might result in depressing prices. However, there were some countries where such commodities offered the best or even the only opportunity to increase export income. W/ World Development Report, 1978; World Bank: Foreword. 2 Annex 1 D, 10. - 36 - On the basis of this analysis, the President proposed that e Bank should limit financing of new production of primary commodities facing inelastic demand, in general, to countries with few export alter_ natives. The paper was discussed by the Board, and its eventual ap,rc,. constituted the Bank's lending policy for commodities facing unfavorable market prospects. The first application of this general policy to a specific commodity--tea--came a few months later, in August 1973, when the President submitted a memorandum on "Bank Group Financing of Tea" to the Executive D-irectors. The need for a Bank policy on tea was preciDf- tated by eight IDA credits made between 1964 and 1973 for financing new tea production, mainly in East Africa. It was estimated that these cred:- wculd generate 20' of the total increase in tea exports to be expected in the decade 1975-84. This increase in total supplies would exceed the slowly growing demand, resulting in a decline in world prices. Nevert-e- less, as the memorandum pointed out, some individual countries and their impoverished smallholders could gain from producing for export, as had happened in East Africa. Under such circumstances, it was claimed that the conditions under which the Bank should contemplate financing tea projects ought to be strictly limited. The Bank staff recognized that a policy of limitatic should ideally be based on the analysis of comparative advantage in tea production, but in view of the complexity and uncertainty of such analysfs considered that simpler approaches could be used. The memorandum chere- fore proposed, and the Executive Directors accepted, the following polity: that as a general rule, the Bank should undertake no further financing of tea production; that any exceptions should be strictly limited to financing (i) increases in output in countries with no investment alter- natives yielding an acceptable rate of return; and (ii) rehabilitation involving no increase in output; and that projects for diversification out of tea production should be encouraged where economically feasible. In the absence of any change in the long-run market outlook for tea and of an effective international tea agreement, this tea-lending policy is still in force. In hindsight, it would appear that the Bank was.slow to take policy cognizance of developments in the tea market and may have thereby been forced to overreact by withdrawing from the production side of the industry almost completely. Clearly the Bank had a responsibility to take some action. The Panel strongly supports the Bank's current decision to consult with other main international and bilateral lenders on tea pro- duction policy. Such a consultation, based on the most authoritative market analyses available, should be concerned with establishing joint guidelines, if possible, first for future tea production operations and, second, for assisting adjustments in tea-producing countries where necessary- - 37 - The Bank has been exercising similar restraint in the financing of coffee projects, in consultation with the International Coffee Organi- zation. But there may not be many other cases where the Bank has a need at present to evolve a commodity-lending policy akin to that on tea. Nevertheless the Panel considers that the Bank, as a major and increasingly dominant international lender for commodity production, cannot for long sidestep such basic development issues as have arisen in this case, and could well develop in the future for other commodities. It is noteworthy, for instance, that the central price forecasts identify a number of commodities subject to long-term decline in prices attributable mainly to likely oversupply. It is important that the implications, if any, of such trends for the Bank's own commodity-lending policies, and for policy consultations with other lenders, be assessed in good time. It is equally important for commodities facing favorable world market prospects to be identified and appropriate operations encouraged. Commodity papers should also be geared in this direction. For instance, policy papers might well examine factors responsible for the slow response of developing countries in the expansion of production and exports of commodities for which favorable world market prospects are widely perceived. It is understood that such a paper is currently being prepared in relation to rubber by Central Projects Staff and Development Policy Staff jointly. This could be a good example to follow. Other papers with policy orientation 1/ have been presented to the Executive Directors since the initiatives of 1973. A paper on cocoa lending in 1974 was discussed by the Board, which agreed to the foreseen investments in this crop. Two policy papers on palm oil and one on beef were not discussed by the Board, but cleared the way for the Bank to con- tinue supporting projects in these fields. A policy paper on sugar in 1972 was discussed by the Board. It opened the way for somewhat more liberal lending by the Bank because of shorter supplies in the world market. Another paper on sugar, reverting to the earlier more restricted role and fitting Bank lending for production into the operational framework of the new Sugar Agreement, was put to the Executive Directors for information in 1978. A more long-term policy cn this commodity is needed in the Bank. Papers on natural rubber, the world jute economy, and the long- term problems of jute, hard fibers and tea did not seem to raise policy issues for the Bank. The important forestry sector paper, which completely re-oriented Bank lending in this sector, was likewise presented for infor- mation only. 1/ See Annex 1 D. - 38 - A policy paper on minerals and energy in the developing countries opened the way for the Bank to expand its program in fuel and non-fuel resource exploration, which was a significant innovation. These illustrations show that commodity papers have played an important part in orienting the Bank's lending policy, as reflected in decisions or approvals of the Board. However, the papers have, for the most part, been geared to the project approach. The Panel feels that the new orientation which it is proposing :, for the regular in-depth commodity studies of the Bank would help to identify development policy issues and prepare the way for the evolution of appropriate Bank policies with respect to its own operations and its initiatives in the international community. (d) Bank Commodity Papers In earlier sections the Panel has reviewed the methodological studies of the Bank, the Price Prospects report and the commodity policy papers submitted to the Board. In this section it records some further comments and recommendations in the light of the considerations presented in this chapter. (i) Long-term Commodity Forecasts As already indicated, such forecasts, whether disseminated centrally or purchased commercially by operating units, provide essential support to the project approach to Bank operations. The short commodity notes accompanying these forecasts could be modified, where necessary, to draw attention to policy or development aspects. (ii) Current World Market Reviews It is important, in addition, that policy and operational staff of the Bank should be regularly informed of the current market situation for all commodities of interest. But unless a sudden need arises, there seems to be no need to prepare special market reports for this purpose. It should be sufficient to obtain and distribute within the Bank copies of the periodic market reports prepared by the secretariats of the authori- tative international commodity bodies, with a covering note, if necessary, highlighting any significant developments for the Bank. (iii) In-depth Commodity Studies 2/ The Panel feels that the Bank's individual commodity studies tend to focus unduly on overall market factors and global supply, trade 1/ In section (d) (iii) below. 2/ See Annex 1 C. - 39 - and prices. To this extent they serve well the project approach in the Bank. However, such aspects are, for the most part, adequately covered by similar studies outside the Bank. In this respect, although there are promising initiatives, 1/ the Panel feels that the Bank's commodity analysis as a whole does not yet display the development orientation appropriate for the evolution of the Bank's operations and role in the future. The Panel recommends that the Bank's own in-depth studies of individual commodities should have a special character, different from that of other agencies and from that currently prevailing. In particular, such studies should focus more strongly on the broad development issues inherent in global commodity situations which Bank loans and policies will increasingly have to take into account, such as: - the status of food deficits in poor countries; - investment requirements in metals and minerals; - the political and geographical structure of production and trade; - progress of trade among developing countries; - development opportunities for specific developing countries in the increase of commodity production and processing for export; - progress toward accepted international development objec- tives, including those of the Bank; and - implications of the evolving commodity situation for the role of the Bank. In other words, commodity analysis should move more strongly into the stream of policy thinking in the Bank. This will usually entail more cooperative studies with other operational and development policy staff of the Bank and with other international bodies with specialized commodity or production knowledge. (e) Research Needs (i) Commodity Lending Policv The Panel recommends that a research project be designed and implemented in the Bank for the study of comparative costs and comparative advantages in the production of a selected number of agricultural and 1/ Especially some recent papers on fertilizers (Annex 1 C, 1), tropical root crops (Annex 1 C, 2), rubber (Annex 1 C, 7), and Near East meat situation (Annex 1 E, 1). - 40- mineral commodities, following up both the RPO-funded projects on resour. allocation in cocoa production and on copper and aluminum modelling 1/ and the study on a "stages" approach to comparative advantage. 2/ The research should be designed to lay the foundations for a more comprehensive and systematic commodity lending policy. It should take into account the cumulative effect of Bank lending for particular commodities, particularly those facing inelastic demand in world markets or artificially low prices due to subsidized exports from developed countries. The project should examine the various options associated with, on the one hand, the possibly undesirable revenue effects of lendin; where elasticities or artificial prices are unfavorable and, on the other, the need for investments to permit adjustments in the production of such commodities or alternative commodities within and among countries. (ii) Processing of Commodities In view of similar development issues arising in connection with Bank operations in the field of processing, the Panel recommends that the Bank undertake a study of factors affecting the location of pro- cessing of selected agricultural and mineral commodities. The research should start from an examination of studies already available with the Bank and be designed with a view to contributing to Bank policy and to international development thinking. 1/ Annex 1 B, 5, 12, 14. 2/ Annex 1 E, 4. - 41 - 1.. 9i.proving the Bank's Capacity to Give Policy Advice to Its Members All the current and potential commodity analysis and research activities discussed above do or would produce results which could improve the Bank's capacity to give policy advice to its members, as well as supporting or orienting its operations. However, very little of the commodity analysis and research undertaken by the Bank seems to have been designed with this purpose in view. Various general papers on commodity policies and problems 1/ by the Bank staff can be regarded as contributing to this objective. The subjects include: - commodity price stabilization and the developing countries; - compensatory financing; - possible gains to developing countries from the elimination of trade barriers in primary and semi-processed products; - food insecurity: magnitude and remedies; - should developing countries carry grain reserves? - stabilization, adjustment and diversification policies for jute, hard fibers and tea. The papers perused by the Panel have been well-researched, well- written and make an authoritative contribution to international discussion of the topic chosen. However, the scope of subjects for general commodity studies of this kind is very wide. It is important for the Bank to con- tribute to international debate, but if priority is to be given to long- term forecasting and a development approach to commodity lending, some constraint may be needed in the field of general studies. Papers published by Bank staff will be most appropriate and valuable in cases where there is a specific contribution of Bank knowledge, experience, research or point of view to make on an issue not far removed from the Bank's role. The RPO research projects on comparative analysis of resource allocation in cocoa production on natural resources and planning, and on simulations of bufferstocks were apparently intended as exercises in methodology, but could be regarded as improving the Bank's capacity to give policy advice to the extent that the projects were completed. Only the bufferstocks simulation project was completed. The resulting publications introduced an alternative approach to food security based on food import bill insurance. While a useful contribution, the subject of bufferstocks was already very much researched. There is con- Siderable international interest in bufferstock approaches to market management, in view of the decision to establish a Common Fund. Studies 1/ See Annex 1 E. - 42 - designed to feed into this international initiative could be useful. It is important in bufferstock modelling, however, to recognize that the existence of a national or international bufferstock may greatly change the market behavior of the private sector. As regards advice to member governments on international commo- dity policies, the Bank Group was authorized by the Executive Board in 1969 I/ to play a more active role than it had been doing. In particular the Bank Group is empowered to assist in facilitating the establishment and operation of appropriate international commodity arrangements by a series of measures which entail advice to such commodity bodies. In recognition of the world development issues inherent in commodity lending, the Bank has a practice of consulting the secretariats of intergovern- mental commodity bodies with respect to the Bank's production projects for the respective commodity. If the Panel's recommendations above are accepted, so that'the Bank's commodity staff specialize on the development policy issues inherent in the commodity markets, bringing into such analysis the Bank's own data on supply conditions and costs, the Bank's commodity policy papers and the results of its research could make an important objective contribution to the evolution of international commodity policies. I/ Stabilization of Prices of Primary Products, Part II; Report of the Executive Directors of the IBRD, Washington, D.C., 1969, page 5. - 43 - 12. Broadening the Understanding of the Development Process This objective calls for basic research rather than applied research or servicing. The limited amount of such research done in the comodity field has been that funded by the Research Committee. The main item was the natural resources and planning project (RPO 671-09), which attempted to describe and explain, through the preparation of models, the functioning of the world energy, pulp and paper, copper, aluminum and jute sectors. Although this work has not been finalized, the interim results are of value. Another relevant RPO project was the Seminar at Oxford in 1973, which studied the link between inflation in industrialized countries and the movements in the price of selected commodities. Following up this seminar with further research, as recommended by the Panel, should add to the understanding of this aspect of the development process. Another research project recommended by the Panel, namely on the factors influencing the location of processing industries, should also contribute to knowledge of development, as well as helping to orient Bank operations. - 44- 13. Assisting in Developing Indigenous Research Capacity in Developing Regions Commodity research by its very nature calls for global data and global analysis. International agencies, therefore, have a marked advan- tage in this field. The US Government has some of the few national agencies doing international multi-commodity research. There are a few other natic: agencies, mostly in developed countries, which have built up competence in research on one or several important export commodities. Most such agencies utilize the reports of international commodity bodies as starting points for nationally-oriented research. It would help in the formulation of sound national commodity policies and investment loan requests if more developing countries had a higher level of research competence in this area. But for supporting its own operations, the Bank has no real alternative but to do its own global market analysis. However, efforts by the Bank to promote the establishment of general economic research capacity in developing countries should encompass commodity research capacity where appropriate. In addition, there may be critical gaps in knowledge affecting assessments of market prospects for some commodities, e.g., the age struc- ture of tree crops in main producing countries and the assessment of mineral reserves. Assistance by the Bank to promote research and data- gathering on such problems would give a high return for commodity policy formulation in both the countries concerned and in the Bank. If the Panel's recommendations for systematic Bank research in the fields of national production systems and costs and factors affecting the location of processing are taken up, it would be highly advantageous to involve national institutions in the work in selected countries. - 45 - IV, INSTITUTIONAL MATTERS 14. .Cyomodity Steering Committee As indicated in Chapter III, the Panel has taken note of the participation of many units of the Bank in commodity analysis and research_ for an array of different purposes in many different environments. It is hard to identify any units who could stop what they are doing without the vork of the Bank suffering. The exchange of complementary information that takes place among the different units in the commodities field is commendable. However, the total work of the Bank in commodity analysis and research does not add up to a coherent whole. The situation is better in the servicing of opera- tions than in research or dissemination, but there is nevertheless scope for mcre exchange of market studies and of information relevant to commo- dity studies and forecasting. The Panel accepts the advantages of one central location for global commodity forecasting services in the Bank, while recognizing the need for a continued inflow of specialized commodity knowledge from certain central and regional units, especially as regards fertilizers, metals, minerals and energy. To promote the coherence and operational/policy value of the whole range of commodity services and research in the Bank, the Panel reco=mends that an in-house commodities steering group be established. The Panel recommends that the proposed group, in addition to promoting the coherence of the operational services in commodities, should also monitor the whole research program in commodities, including proposed new projects, and develop a more coherent and open system of publication for the Bank's output in the commodity field. - 46 - ANNEX 1: RECENT WORLD BANK OUTPUT IN THE COMMODITY FIELD A. COMMODITY TRENDS AND FORECASTS 1. Price Prospects for Major Primary Commodities; June 1978 and June 1977: EPDCE., 1/ Report Nos. 814/78 and 814/77. 2. Commodity Trade and Price Trends; August 1978 and August 1977: EPDCE, Report Nos. EC-166/78 and EC-166/77. B. COMMODITY MODELLING/METHODOLOGY STUDIES 1. An Econometric Forecasting Model of the World Sugar Economy; March 1979: EPDCE (draft). 2. Prospects for the World Iron and Steel Economy: The WISE Model; February 1979: EPDCE (draft). 3. An Econometric Model of the World Cocoa Economy; January 1979: EPDCE (draft). 4. An Econometric Model of the World Rubber Economy; November 1978: EPDCE (draft). 5. A World Aluminum Model; September 1978: Development Research Center. 6. The World Tin Economy: An Econometric Analysis; June 1978: EPDCE, World Bank Staff Commodity Paper No. 1. 7. A Dynamic Simulation Model of the World Jute Economy; May 1978: Development Research Center. 8. World Bank Commodity Models - Workshop Discussion Drafts - Vol. 1, Purpose and Scope, Basic Structure; April 18-19, 1978: EPDCE. 9. World Bank Commodity Models, Vol. 2 - Tin, Iron and Steel, Rubber, Jute, Energy - Demand (Raw Materials); April 18-19, 1978: EPDCE. 10. World Bank Commodity Models, Vol. 3 - Food Products (Fats and Oils, Sugar, Cocoa, Coffee, Tea); April 18-19, 1978: EPDCE. 11. Tea Model; April 1978: EPDCE, Status Report. 1/ This abbreviation identifies the Commodities and Export Projections Division, Economic Analysis and Projections Department. -47 - ANNEX 1 Page 2 12. A World Copper Model; September 1976: Development Research Center. 13. Dynamic Market Oriented World Food Projection and Planning Models; July 1976: RPO SR 671-23, Report No. 2. 14. Comp.arative Analysis of Cocoa Production in Selected Countries; July 26, 1974: EPDCE, SecM74-528. C. INDIVIDUAL COMKODITY STUDIES 1. World Fertilizer Review and Fertilizer Requirements of Develop- ing Countries, 1979; May 1979: Fertilizer Unit, Industrial Projects Department (draft). 2. Tropical Root Crops and Rural Development; February 1979: Agri- culture and Rural Development Department. 3. World Grain and Rice Situation and Outlook; December 1978: EPDCE, Commodity Note No. 9. 4. World Potash Survey; September 1978: Fertilizer Unit, Industrial Projects Department, World Bank Staff Working Paper No. 293. 5. International Cotton Market Prospects; June 1978: EPDCE, World Bank Staff Commodity Paper No. 2. 6. Optimal Hedging under Price and Quantity Uncertainty: The Case of a Cocoa Producer; June 1978: EPDCE, World Bank Commodity Note No. 6. 7. The World Rubber Economy: Structure, Changes, Prospects; June 1978: Joint Staff Study by World Bank (EPDCE) and FAO. 8. Long-Term Projections of Steel Consumption; May 1978: EPDCE, World Bank Commodity Note No. 4. 9. Copper: Current Situation and Short-Term Outlook; May 1978: EPDCE, World Bank Commodity Note No. 3. 10. Structures and Trends in the World Tobacco Market; December 1977: EPDCE, World Bank Commodity Note No. 2. 11. Toward Selecting a World Indicator Price for Beef; December 1977: EPDCE, World Bank Commodity Note No. 1. 12. Fluctuations in Invisible Stocks: A Problem in Copper Market Fore- casting; October 1977: EPDCE, World Bank Commodity Paper No. 27. 13. Market Structure of Bauxite/Alumina/Aluminum: Prospects for Develop- ing Countries; March 1977: EPDCE, World Bank Commodity Paper No. 24. - 48 - ANNEX 1 Page 3 14. World Phosphate Survey; 1979: Fertilizer Unit, Industrial Projects Department, Ref. No. R403 (draft). 15. Energy Demand Forecasting: A Critical Review of Current Approaches; February 1977: Energy, Water and Telecommunications Department, P.U. Report No. RES 10. 16. Energy Supply Demand Outlook, 1980-1985; July 18, 1974: Energy, Water and Telecommunications Department, Background Paper No. 2 for IBRD Report No. 477. 17. The World Jute Economy, Vol. I; July 12, 1973: South Asia Regional Department, Report No. 114a-BD. 18. Prospects for Palm Oil; July 1976: EPDCE, World Bank Commodity Paper No. 23. 19. Tropical Hardwood Trade in the Asia-Pacific Region; 1974: EPDCE, World Bank Staff Occasional Paper No. 17. D. PAPERS ON BANK LENDING POLICY FOR COMMODITIES 1. Bank Financing of Tea - A Reappraisal; February 1979: EPDCE. 2. The World Sugar Economy: Review and Outlook for Bank Group Lending; February 2, 1978: EPDCE, Report No. 1894. 3. Forestry; February 1978: Central Projects Staff, World Bank Sector Policy Paper. 4. World Grain and Rice Situation and Outlook - 1977; July 27, 1977: EPDCE, for information of Executive Directors, SecM77-608. 5. Tin Buffer Stock; May 23, 1975: EPDCE, memorandum from President to Executive Directors, SecM75-378. 6. Minerals and Energy in the Developing Countries; May 4, 1977: Office of Vice President, Finance, Report No. 1588. 7. Fertilizer Requirements of Developing Countries; March 1974: International Finance Corporation. 8. The World Cocoa Market - Review and Outlook for Bank Lending; February 5, 1974: EPDCE, for Executive Directors Meeting, R74-36. 9. Financing of Tea; August 17, 1973: EPDCE, memorandum from Presi- dent to Executive Directors, R73-206. - 49 - ANNEX 1 Page 4 10. Development Policy for Countries Highly Dependent on Exports of Primary Products; January 4, 1973: EPDCE, memorandum from President to Executive Directors, R73-3. 11. Market Prospects for Rice - Framework for Commodity Lending in 1973-74; October 16, 1972: EPDCE, Background Paper for Executive Directors, SecM72-524. GEN7RAL PAPERS ON COMMODITY POLICIES AND PROBLEMS 1. The Outlook for Meat Production and Trade in the Near East and East Africa; December 1977: Joint Staff Study by FAO and the World Bank (EPDCE). 2. Food Insecurity: Magnitude and Remedies; July 1977: Development Economics Department, World Bank Staff Working Paper No. 267. 3. Commodity Price Stabilization and the Developing Countries: The Problem of Choice; July 1977: EPDCE, World Bank Staff Working Paper No. 262. 4. A "Stages" Approach to Comparative Advantage; May 1977: Office of the Vice President, Development Policy, World Bank Staff Working Paper No. 256. 5. Possible LDC Export Gains from Elimination of Trade Barriers in Major Primary and Semi-Processed Agricultural Products; April 1977: EPDCE (draft). 6. A Perspective on the Foodgrain Situation in the Poorest Countries; April 1977: Policy Planning and Program Review Department and Agriculture and Rural Development Department, World Bank Staff Working Paper No. 251. 7. Coffee, Tea, Cocoa - Market Prospects and Development Lending; 1977: EPDCE, World Bank Occasional Paper No. 22. 8. Export Incentives and Export Performance in Developing Countries: A Comparative Analysis; January 1977: Office of the Vice President, Development Policy, World Bank Staff Working Paper No. 248. 9. Statement made before the House of Lords Select Committee on Commodity Prices; December 8, 1976: EPDCE (mimeograph). 10. Developing Country Foodgrain Projections for 1985; November 1976: EPDCE, World Bank Staff Working Paper No. 247. - 50 - ANNEY, Page 5_ 11. Stabilization, Adjustment and Diversification: A Study of the Weakest Commodities Produced by the Poorest Regions; November 1976: Development Economics Department, World Bank Staff Working Paper No. 245. 12. The International Dialogue on Commodities; June 1976: EPDCE, World Bank Reprint Series No. 39. 13. Energy and Petroleum in Non-OPEC Developing Countries, 1974-80; February 1976: EPDCE, World Bank Staff Working Paper No. 229. 14. Compensatory Financing: A Quantitative Analysis; December 1975: EPDCE, World Bank Staff Working Paper No. 228. 15. Raw Materials: US Government Stockpile As Compared with World Production, 1950-1974; June 9, 1975: EPDCE, Office Memoraidum. 16. Agriculture and International Trade: Some Issues of Interest to the Developing Countries: Address to the Third Annual International Trade Conference, Seattle, May 12, 1975: EPDCE. - 51 - ANNEX 2: BRIEF REVIEWS OF SELECTED INDIVIDUAL COMMODITY MODELS The World Tin Economy: An Econometric Analysis World Bank Commodity Paper No. 1, June 1978 This is an orthodox econometric model, with supply equations for 7 producing countries and 1 for the rest of the world; demand equations for 6 regions; and a price equation. The supply equations are usually fur.ctions of the previous year's supply, the LME spot price, deflated, and time trends and dummy variables. The demand equations are divided for each region, where possible, between tin consumption for tinplate and other uses. For tinplate use the explanatory variables are usually the lagged price of tin, the ratio of electrolytic to total tinplate production, and sometimes the price of aluminum and dummies. For non- tinplate uses the explanatory variables vary more from region to region, including, for example, indices of industrial production, of machinery and construction production, and of transport equipment production. For the price equation the explanatory variables are the price lagged one year, world stocks and the US wholesale price index. On the supply side, the main criticisms are the very simple lag structure and the lack of differentiation between investment in capacity and production. The demand equations are successful over the historical period since they are tied rather closely to the uses of tin. From a prediction point of view, however, the problem is merely shifted from the projection of tin consumption to the projection of such variables as tinplate production, the production of machinery and construction goods, etc. Even the secular substitution against tin hangs on the proxy variable--the ratio of electrolytic to total tinplate production--which can also be affected by the macroeconomic situation. It seems that the demand equations should be supplemented by equations linking the end- uses of tin to the macroeconomic variables on which assumptions can be made which are common to all commodities. The inability of the price equation to reproduce sharp changes in tin prices reflects the difficulties for a commodity which can be stocked and for which there may be strong speculative factors influencing the market. From the point of view of long-term projections, such a weakness in the price equation is perhaps tolerable, but it does raise the problem of linking the shortrun, in which expectations and speculation can exacerbate the effects of a disequilibrium situation, with the long- run. It also creates difficulties for the use of the model, as it is used in this paper, for the simulation of possible bufferstocks policies. The price equation needs to be much more sophisticated than the present One, which superficially solves its problems by the use of dummy variables (variants 2 and 3), and probably ought to take account of US sales from strategic stocks. - 52 - ANNEX 2 Page 2 One has to question the final conclusion with regard to future prices: "...it would appear... that tin producers can look forward to a period of slow, but steady, rise in import demand for this metal as well as prices.... However, as a result of lags that characterize the response of demand and supply... the current price boom will be followed by a per4cd of much lower real prices which is likely to begin in the early 1980s and to last until the mid-1980s." The second half of this statement does not appear to be derived from the model, nor is the view of lag structure implied in the quotation reflected in the model. An Econometric Model of the World Cocoa Economy January 1979 (draft) This again is an orthodox econometric model, with demand equa- tions for 8 regions, supply (lagged adjustment) equations for 6 countries, a world price equation, and 2 producer price equations. Although cocoa is a tree crop with a gestation lag of 4 years and an interval to maximum production of 12 years, in only two countries was this explicitly recog- nized in the supply equations. It is curious that in the section dis- cussing the supply equations, there is reference to only one piece of previously published work (D.L.L. Shu, 1975), when in fact a great deal of previous research has been carried out, especially with reference to conditions in individual countries. The implied (price) supply elasticities range from 0.2 to 1.7, but there is only cursory reference to this range and no explanation is offered. Consumption, proxied by grindings (it may be true that no recent consumption data are available in published form, but are there no un- published data available?) is usually explained as a function of GNP and price, but the presence of the lagged value of the dependent variable as usual confuses the issue and makes it difficult to assess the role of the economic variables. What, however, seems so remarkable as to be unacceptable without detailed justification is the projection that con- sumption (grindings) will increase by a factor of 4 during the 12 years from 1978-90, when over the observation period 1962-76, consumption in North America has fallen, in Western Europe has been approximately static, and the world total has increased only by some 30%. The implausibility of these results was recognized by the authors who have produced a new version of the model by revising the supply equa- tions for 4 producing countries, essentially by changing the assumed lag structure of the assumed supply response to prices (sometimes increasing and sometimes decreasing the lag). The fact that these changes result in a revision of the projections so that the increase in production and grindings over the period 1978-90 is reduced to about 50%, while "the - 53 - ANNEX 2 Page 3 simulations with the new model differed only slightly from the simulations with the original cocoa model," induce a certain despondency, rather than providing reassurance. That the "projections are roughly in line with currently available information about plans to expand cocoa production in major producing countries" again makes one doubt the whole methodological and epistemological approach. If an econometric model has to be manipulated until its projections come into line with information already available, we are dangerously near to practicing computerized black magic. An Econometric Model of the World Rubber Economy World Bank Staff Commodity Working Paper No. 3, November 1978 This is an orthodox econometric model of the supply, demand and price of natural rubber. The supply equations covered 3 countries plus the rest of the world. The demand equations were in two stages: demand for total elastomer was estimated for 6 regions; and the market share of actual rubber was estimated for the same 6 regions. to this was added an estimate of demand for natural rubber by China. The NY spot price for RSS 1 was estimated as a function of world stocks and the price of syn- thetic rubber. Simultaneity was avoided by time lags in the supply equa- tions. Judged by elasticities, the model performs reasonably. Supply price elasticities vary from 0.2 to 0.25 in producing areas; the price elasticities for natural rubbers' market share varied from 0.1 to 0.4 by region, short-term, and for the two regions for which estimates were made, from 1.2 to 1.8. Spot prices showed a strong inverse-relation with world stock and a positive relation with synthetic prices. A criticism of this model is that it conceals too much. In the demand for total elastomers, it is argued that, since a large part of the demand is derived from automobile production which is dependent on GNP, and most of the rest can also be linked to GNP, it is sufficient to use GN? as the main explanatory variable. This is true, but much more infor- mation would have been provided by a disaggregation in which these rela- tionships could have been seen more explicitly. The demand for new auto- mobiles and for replacement tires for the national automobile park, relates differently to GNP in different countries, and an understanding of this would be helpful for prediction. The present equations rely too much on the assumption that a number of concealed links between demand for elastomers and GNP will remain in.the same form as in the past. Since the mid-point of the data series is 1965, and we are interested in projecting into the 1980s, such an assumption is dangerous. - 54 - ANNEX 2 Page 4 In the market share equations for natural rubber, the lagged value of the dependent variable appears also on the righthand side. While this allows the interpretation that there is a distinction between short- and long-term elasticities, this is not the only interpretation, given the high inevitable correlations between values of a variable at times t and t-l. In such a case, it should require investigation and explanation, when such a formulation works for all regions except apparently Japan, and for the CPE countries it is also necessary to include a quadratic function of time. By such arbitrary means it is possible to get a good fit to t-e data, but we are left with the unpalatable result that when the model is used to project prices from 1977 (US$900/metric ton), the range of possibilities in 1982 is $846 to $1,549. A result spanning this range might have been arrived at more or less by verbal reasoning. The real criticism, however, is not so much the wide range of projections, but the fact that very little can be done about them. The model simply does not provide enough information to get much beyond this result. What it could provide it has not been asked to, nor is there any discussion of what lies behind the high and low assumptions of GNP growth and inflation (how does inflation enter the model in any case?). It is depressing to have to state that the final conclusion--"Rubber producers can, therefore, look forward to a period of steadily rising prices"--is not scientifically supported by the analysis as such. It seems rather to rest on "available information about the world rubber economy" referred to (in what seems a completely inadmissable way, as by sleight-of-hand) in the sentence which precedes the conclusion. Sugar: Econometric Forecasting Model of the World Sugar Economy March 1979 (draft) This is a substantial econometric model based on annual time series for the period 1951 through 1975. There are supply equations for 34 countries or groups of countries, demand equations for 131 countries, and a world price equation. This degree of disaggregation is greater than necessary for making forecasts and simulating at the level of the world market, but the model was also intended to be useful at the level of individual countries. The report on the model contains a careful description of the institutional characteristics specific to the sugar market, and the model takes these characteristics into account. It is possible that professional opinion would vary on the particular formation chosen, but the one used is at least as valid as any other and presents a more consistent global picture than any previously available. For example, the world price used - 55 - ANNEX 2. Page 5 in the model is a weighted average of the price in the US preferential market, the price in the Commonwealth preferential market, and the residual world price, whereas an alternative approach would have been to separate these markets, at the same time providing links between them. However, since the basic data set has been established, further experiments would always be possible. Table 1 of Annex 1 shows that there have been some significant diversions of prices in the three markets during the estimation period. The supply equations based on a lagged response of producers to changes in price are expected to differ between beet and cane producers and produce a range of supply elasticities, most of which can be plausibly interpreted. The short-run elasticities are small and all but one lie in the narrow range 0.0 to 0.3. Within this range the variation differs according to market conditions, particularly according to the degree of government intervention designed to insulate producers from the residual world price. The long-run supply elasticities are both larger and more variable, the range being from 0.1 to 4.9. Again, within this range, most of the differences between countries can be reasonably interpreted, there being, in addition to other explanatory factors, a positive rela- tionship between the size of the supply elasticity and the rate of growth of consumption. The demand equations are classical equations depending upon population, income and price. In this case the divergence of retail prices in individual countries from the world price is sidestepped by taking both income and price elasticities from FAO analyses which were carried out at the individual country level. There is, however, an implicit assumption that, in the cases of countries for which price elasticities were used (countries depending for a large part of their consumption on imports from the residual world market), retail prices moved proportionately to the residual world price. After testing the model for fit to the historical series, simu- lations are then performed under different assumptions related to the 1978 International Sugar Agreement. These simulations are very interest- ing and clearly worth taking seriously on the basis of a wide discussion. A very acceptable feature of the report is the large number of graphs and tables which enable the reader to follow the calculations as easily as possible. -56- ANNEX 2 Page 6 A Dynamic Simulation Model of the World Jute Economy May 1978 (draft) This report is a study of the world jute economy with the specific intention of studying the implications of alternative policies such as compensatory financing, tariff policy and acreage limitations. It is a classical econometric model in that it is based on supply and demand equations and a price equation. An important innovative feature is that it attempts, in the supply equations, to allow for uncertainty as well as a lagged response to actual price. In the simulations, first a deterministic model is run, in which rules are given to a bufferstock agency. Then a stochastic model is run, in which random stocks are applied to the supplv equations, for comparison with the deterministic runs. Broadly, the results show that attempts to stabilize prices at around their average historical levels would be feasible, and with a profitable bufferstock operation. Attempts to raise the price either once and for all or on a rising trend by inducing supply increases in face of an inelastic demand would lead to large losses by the bufferstock authority. These losses would be roughly offset by gains to the pro- ducing countries. These gains would be unequally distributed because of the differential response of supply to price in the producing ccuntries. These results are clearly worth wide discussion. From a tech- nical point of view, it is not clear whether the inclusion of terms in the supply equations to reflect producers' reactions to the diminished uncertainty created by stabilization policies is a significant element in the results obtained. Of course, it may be argued that even if it is not, it is helpful from a policy point of view to know that. Against which it may be said that, just as the Nerlove approach is only one specific approach to the problem of lagged response, so the present representation of uncertainty by the lagged variance of past observations is only one possible representation of uncertainty. The buestion must remain unresolved until the method has been tried for other appropriate commodities. This model has the virtue that it is not too difficult to apply and needs no new data. - 57 - ANNEX 2 Page 7 Prospects for the World Iron and Steel Economy: The WISE Model February 1979 (draft) This paper represents the current stage in the development of a model for the world iron and steel economy, based on the premise that steel producers formulate and carry out their investment plans on rational grounds with perfect foresight in a market environment which ensures that prices cannot widely diverge from costs for more than short periods. This is clearly a bold endeavor, when it is recalled that, for example, -anning ratios differ widely in different producers countries, even when similar technologies are used; that in some countries, such as the UK and Italy, the steel industry is publicly owned and employment policy has dominated market considerations on a number of occasions; and that in many other countries, -regulatory policies are followed for the purpose of protecting the domestic industry. Probably what makes the approach feasible is the dominance in the world market for Japan as a producer, especially the dominance of Japan as the major incremental producer and exporter. The increment of production in Japan comes from greenfield sites embodying the latest technology. However, the author recognizes the market imperfections referred to above and in fact presents two alternative models, Model A based on perfect rationality and foresight, and Model 3 constrained by actual investment decisions during the 1970s. The main technical advance in these dynamic models is the simultaneous solution of the dynamic invest- ment program for each year from the starting point over a planning horizon assumed to extend to 1990. As was to be expected, Model A failed to correspond with reality as a result of the optimism of the industry in the early 1970s and the consequent failure to foresee the depression of the mid-1970s. It, nevertheless, projects a fairly smooth progression of prices through the 1980s to 1990. Model B, due to its acceptance of over-investment in the 1970s, forecasts a steep decline in steel prices after 1985. Such models are, of course, very demanding of good technological information, but this is perhaps the smaller part of the problem when considered purely as a problem of price forecasting. Because of the importance of investment and the long gestation lags, dynamic models of this kind must be expected to produce fairly dramatic solutions, and relatively small errors in forecasting market conditions in the early part of the forecasting period will have sharp repercussions later on. What, therefore, assuming that the models can be further refined, can they be expected to contribute to the operational role of the Commodity Price Forecasts? It seems that they can be useful in probing alternative Possibilities some 10 to 15 years ahead, and particularly useful in high- lighting the fairly wide range of possible outcomes. For the short- to -58- ANNEX 2 Page 8 medium-term, it would seem wise to utilize more conventional models more closely tied to actual behavior and market imperfection, plus, of course, a knowledge of the detailed institutional structure of national steel industries and of world trade conditions. A World Copper Model RPO 671-09 September 1976 (draft) This model considers production decisions in 6 primary producing areas which utilize only primary copper, and 7 primary and secondary producing areas which process both primary and secondary copper at smelters and refineries. The world is divided into 13 market areas. The model, which is based on a mixed integer programming system, exploits the fact that much copper is produced by large transnationals and that much technological and cost information can be obtained, includ- ing international transportation costs. Essentially, the model as presented accepts the demand for refined copper in the 13 marketing areas of the world in 1973, 1980 and 1990 as estimated by a previous study (Fisher, Cootner and Bailey, 1972) and estimates the way in which the demands are likely to be distributed over the producers. Clearly the work is based on many previous essays in the development of similar models and on a great deal of information on supply conditions. To this extent, it can be considered as a more mature example of the type of model being developed by Mr. Hashimoto for the world iron and steel industry. From the point of view of the Bank, the crucial question will remain as to the projection of future demand conditions, here taken as given. These have an importance not only insofar as there may be uncer- tainty as regards the total demands in different regions, but also in the division of demand between producers of the primary metal and the recovery of scrap. As in the case of the WISE model, it is recommended that more conventional studies should supplement the further refinement of this type of model. A further crucial factor is the evolution of the substitution relations between copper and aluminum in electricity generation and distribution. It may be added that specialized studies such as "Fluctuation in Invisible Stocks: A Problem for Copper Market Forecasting," World Bank Commodity Paper No. 27, October 1977, M. Radetzki, provide both more insight, which can be used in a qualitative sense to check model pro- jections, and also a useful input into model design itself. - 59 - ANNEX 2 Page 9 A World Aluminum Model September 1978 This is a very detailed study of the world aluminum market, based methodologically on the approach developed for other markets by the author and others in the last 10 y:ars. The model falls into the class of cost minimization models using demand conditions estimated from the projection of econometric equations. Various scenarios are investigated on the basis of major strategic assumptions, with a time horizon extending to 1990. As in other cases, the greater part of the research effort is devoted to supply conditions and the solution of these by a cost-mini- mization assumption. The estimation of demand for aluminum in individual countries is very briefly discussed and amounts to little more than a presentation of the estimates. As far as one can determine, the paper contains little or no discussion of the correspondence between projection and reality, and from within the document it is very difficult to appraise its validity as an operational tool. Judged by the detail of the information collected on supply condi- tions and the obvious depth of knowledge from a producer's point of view which is displayed, one expects that the model is a highly polished one within its own terms of reference. An appropriate strategy for the Bank to adopt in such circumstances might be to leave the development of such sophisticated supply models to outside consultants, relying on the informed opinion of other experts in the field, and devote the Bank's own resources to an investigation of demand factors and the choice of scenarios for which to use the supply models. Since these models typically relate to metals for which the demand is derived, often at several stages separation from final goods, major errors are easy to make. The link through to final demand, with substitution possibilities threatening at each stage, itself presents formidable problems. Perhaps the Bank staff may be better equipped to investigate these problems than outside experts who view the world mainly from the point of view of the producing industries. It could be added, however, that an additional use of this type of model, apart from projections, is more directly in cost-benefit calculations. -60- ANNEX 2 Page 10 Dynamic Market-Oriented World Food Projection and Planning Models and Their Empirical Results for 1970-74 for World Food Situation Report 2, July 1976 RPO: SR 671-23 This report is a good example of work which the Bank should not allocate to outside consultants. It is well-known by now that gran- diose model building at the world level, which claims to have solutions to most of the world's food problems, is a non-starter. Such model building, complex enough for a world divided into 2 regions and contain- ing only 8 commodities, inevitably becomes more and more complex, until it sinks beneath its own weight. In this particular case, even the theoretical basis of the model is open to stringent criticism, since it makes so many strong assumptions, such as perfect competition and foresight, and is forced to adopt so many simplicications out of computational necessity, that it is incapable of throwing any light on the nature of the real world. And the more regions, the more commodities and the more time-periods are included in such a model, the greater its incapability will become. - 61 - ANNEX 3: SOME OBSERVATIONS AND SUGGESTIONS CONCERNING ENERGY AND COMMODITIES IN THE RESEARCH PROGRAM OF THE WORLD BANK by Ali A. Attiga The following are some brief observations and suggestions cncerning my participation in the first session of the Advisory Panel on Co=odities established by the Bank.* Unfortunately, due to my late arrival, I was not able to join the Panel for the important briefing session which took place Monday, July 10. Also, it is regrettable that =v present duties prevent me from continuing my work as a member of the ?anel, but, as conveyed to the Bank, I shall be glad to read the final report of the Panel and offer my comments as appropriate. Enerv in the World Bank From the point of view of developing countries, perhaps one of the side benefits of the 1973-74 oil price adjustments was greater involvement of the World Bank in the energy prcblems and options of its .ember countries. Prior to the oil price adjustm-nts, the Bank confined its activities in the field of energy to studies and investments in elec- tric power, especially hydroelectric power in developing countries. The initial reaction of the Bank to the so-called energy or oil crisis was perhaps naturally directed to the impact of the increase in oil prices on the accumulation of foreign liquid assets held by the oil-exporting countries. But, unfortunately, the Bank overreacted in this regard with the result that its forecasts of so-called OPEC surplus funds were greatly overestimated in comparison with other forecasts pub- lished in 1974. Compared to actual figures, the Bank's initial fore- casts turned out to be about four times greater than the present accumu- lated surplus of OPEC countries. After that initial and rather hasty jump into the deep water of global oil and energy problems, the Bank undertook a series of useful studies dealing with the relationship between the price of energy and Potential growth in developed countries, the energy needs and potentials of developing countries, and it tried to develop a dynamic model of OPEC trade and production. At present, work on energy is divided between the Economic Analysis and Projections Department, including the Commodities Division, the Development Research Center and the Energy, Water and Telecommunications *The Panel held its first session from July 10-14, 1978 in Washington, D.C. - 62 - ANNEX 3 Department. The Department for Development Economics, along with regiona1 departments, may also get involved in energy studies and related issues. From the point of view of more efficient allocation of resources, a better division of labor in the energy field seems necessary and highly desirable at this early stage of the Bank's involvement in this area. There may be different ways of achieving such division of labor, but two main approaches are suggested here. One rather simple and clearcut approach is to consolidate all matters dealing with energy in the newly-created Energy Division. Since the main task of this Division for the foreseeable future is to promote development of domestic energy resources of energy-importing developing countries, it will need to undertake periodic studies of the energy situa- tion in these countries. Perhaps it would be more efficient to concen- trate all the staff needed for the preparation of such studies in the , Energy Division, where they could be in daily contact with the operational staff and maintain a continuous relationship with energy-oriented institu- tions in the developing countries and indeed the world at large. In this case, the research staff can also participate in the country missions and project studies undertaken by the Energy Division. Such a close rela- tionship would give them the advantage of planning their studies and making their estimateo and forecasts on the basis of the actual needs and expecta- tions of the energy authorities of the countries assisted by the Bank. It would also give the operational staff in the Energy Division direct and daily contact and interaction with the energy staff of the Bank, which should -facilitate better coordination and more efficient allocation of resources in the Bank's energy-related activities. Another approach is just to follow the usual Bank system of functional division of labor as it exists in other fields such as agri- culture and industry. In fact, this approach is already being practiced in the case of energy. Generally, it would allocate global data collec- tion and energy forecasts to the Commodities Division, perhaps energy plan- ning and programming to the Economic Development Department, some basic energy modelling and research to the Development Research Center and, finally, all operational matters to the Energy Division. Under this approach, one can see an active role for The regional departments, espe- cially those with large energy exporters and/or importers among their member countries. However, as the work of the Bank expands in the various aspects of energy supplies and options, this approach may compli- cate further the internal coordination problems of the Bank. It could also deprive it of the opportunity to create the necessary specialization in the energy needs and potentials of developing countries due to exces- sive fragmentation of energy discipline within the Bank. The Bank's management will have to decide on which of these approaches it wishes to adopt for pursuing its objectives in the energy field. Perhaps a general look at the nature of the work required for this purpose will help in selecting the most appropriate approach. - 63 - ANNEX 3 Page 3 At present, the Commodities Division collects global data on vergY supply and demand from published sources. It also includes oil erices in the tables of commodities contained in its annual report on Basic Primary Commodities. Forecasting the prices of oil and other sources of energy is not contemplated in any serious way due to the ad- -inistered nature of these prices. Moreover, there are many other re- W Ional, national and specialized energy research institutions which :eriodically make such forecasts. The energy research staff can readily ,ake use of these forecasts and adopt them to its own needs. In fact, e Commodities Division currently depends on a London-based consultant for h of its information on oil and energy. It is doubtful that the Bank vould be able in the near future to collect and produce its own primary energy data on a global basis. It is neither well-placed for this kind of work nor, in my opinion, should this be its primary concern. On the other hand, assisting the energy-deficit developing countries to develop their domestic energy sources would require systema- cic collection of primary data on energy consumption by sector of the economy and source of energy. It would also require data on energy pricing and domestic pricing policies in these countries. In order for the Bank to succeed in performing this task, it should provide adequate technical assistance to these countries in the organization and initial management of their energy programs, which should begin with the formula- tion of the national, and where appropriate, regional data base, needed for the formulation of such programs and policies. This kind of work is rather operational and can be readily tied to the project-lending activities of the Bank. As the Bank would need to assess the cverall energy situation and potentials of each country or group of countries before it could decide on its lending priorities in the field of energy, it would seem desirable to combine the technical assistance program in energy with the lending operation for specific energy development projects. Such a combination could help the Bank obtain primary data and design overall energy programs for the countries assisted, which would be rather difficult to do if approached in isolation. Indeed, the developing coun- tries need to formulate national and regional energy programs for meeting their development needs, and the Bank should be in a good position to assist them in this regard. As the energy needs of most developing countries consist of either creating or increasing their domestic energy resources, it is clear that project identification, evaluation, financing and implementa- tion should constitute the bulk of Bank activities in this field. Syste- matic data collection and adequate formulation of energy programs are essential for this type of work. This would argue for consolidating the energy activities of the Bank in the Energy Division, since that division is already in possession of abundant data on electric power and is responsible for the operation of all Bank lending in energy. - 64 - ANNEX 3 Page 4 Another policy consideration for the Bank is the kind and extent of its investment in the energy field. As explained to the Panel, the Bank has decided to act as a promoter of investment and to direct its lending to energy development projects, mainly for oil, gas and coal, but also for nuclear and unconventional sources of energy. Enhanced oil recovery could also obtain Bank lending. As far as exploration is concerned, the Bank would lend for development exploration, but not for exploration as such. It plans to lend about $500-600 million per year over tne next five years. While this policy represents a radical and progressive shift in the Bank's position on investment in the hydrocarbco1 induscry, it is still far short of meeting the basic needs of developinz countries in the discovery and development of their domestic energy rescurces. One such basic need is the provision of some risk capital for a:cploration. For understandable reasons, the Bank finds it too risky to use its own resources or borrowing facilities for this purpose. But it may be desirable to study the feasibility of establishing a separate exploration Fund for this purpose, either along the lines of the UN Fund for Natural Resource Explorations, or a different kind of Fund which would participate in the venture capital for oil and gas exploration and share proportionately in earnings from successful discoveries. The funding of such a scheme could come from annual contribution from the Bank's earnings and contributions from industrialized countries interested in expanding the energy base of developing countries to reduce their dependence on imported oil and thus leave more of it for exports to industrialized countries. These contributions may come from public appropriation for energy development in the oil-importing industrialized countrics. The OPEC Special Fund is already providing some hi*-:Ly favorable financial-assistance to developing countries in the development of their energy resources. The World Bank is well-placed to activate the interests of public, regional and international financing institutions in the pro- vision of essential risk capital for the discovery of domestic energy sources, especially hydrocarbons, in the developing countries. Once such resources are located it should be relatively easy for the develop- ing countries to obtain development financing through regular established banking and investment institutions. Perhaps for these reasons it would be more effective for the Bank to use its relatively limited resources for assiscance in financing exploration, rather than development activities. Other Suggestions Related to the Work of the Commodities Division 1. The Data Base In addition to the existing data sources now in use, it would be desirable and perhaps essential for the improvement of the data base to make better use of the vast country and project data concerning commodities assembled by the regional staff of the Bank in the course of its country -65 - ANNEX 3 Page 5 and project-oriented studies. At present it seems that little or no use is being made of this potentially rich source of relevant primary data. Commodity data collected by UN specialized agencies, such as the FAO, should provide the Bank with much of its global data needs for agricultural commodities. The adoption of such policy could help reduce wasteful dupli- cation and promote cooperation between the Bank and other relevant regional and international institutions. 2. Practical Use of Forecasting The regional staff of the Bank seem to have considerable difficulty in adapting the global forecasts of the Commodities Division to the specific case of a particular country or region. It would seem highly desirable to achieve greater contact and integration between the work of the Commodities Division and the regional staff in the formula- tion and analysis of field surveys, data analyses and market forecasts. At present there seems to be no organized forum of consultation or co- operation between the regional staff and the Commodities Division in the case of forecasting. I am sure Professor Brown of the Panel will cover this subject in detail and make recommendations for improving this situa- tion. There is also need for closer cooperation between the Bank and the Fund on balance of payments forecasting, especially in the case of medium-term forecasting which the Fund is making regularly in connection with the Extended Fund Facilities. 3. Scope and Orientation of Research It is assumed that research in the Commodities Division is carried out in response to the policy needs of the Bank. But it is not clear whether the Bank's policy is formulated in such a way as to facilitate the efficient division of labor between the departments involved in re- search. At present it seems that there is poor linkage between these departments in the case of research formulation, execution and exchange of results. There would seem to be an urgent need for more contacts and closer coordination between the policymakers of the Bank and the research staff on the one hand, and between the research staff of different depart- ments on the other hand. Some of the research papers shown to the Panel seem to place too much emphasis on the theoretical question of comparative advantage in the production of certain primary commodities in two or more developing countries, with the object of orienting Bank lending to projects and countries with the highest advantage. It would seem reasonable to suggest that the Bank, as a world development institution, should conduct its research more in the direction of identifying and emphasizing the potential and existing advantage of certain developing countries over industrialized countries in the production of certain basic commodities such as sugar. At present Bank research activities in commodities seem to reinforce the existing market structure and economic relationships between the developed - 66 - ANNEX 3 Page 6 and developing countries. What is needed, in my opinion, is more econc'ic research aimed at exploring the benefits of more efficient division of labor between developed and developing countries in the production of cer- tain basic commodities. The production of semimanufactured goods and certain manufactured products in which some developing countries either have or could be made to have clear advantage should become the subject of extensive research by the Bank. There could also be more coordinatica between the research staff of the Bank and the work of UNCTAD in focusing' attention on the benefits of giving the exports of developing countries greater access to the markets of advanced countries. More coordination with the UN ngencies, such as FAO, UNDP and regional development banks and institutions, would seem highly desirable and even necessary for improving the quality of research and increasing its relevance to the real problems and policy issues facing the developing countries. The lending power of the Bank can play a vital role in giving substance to such coordination on a regular and systematic basis. 4. The Need for Building Research Capability within the Developing Countries It is within the research objectives of the Bank to assist developing countries in the formulation and development of their research institutions. Yet there seem to be no specific project within the Bank that is designed to implement this policy. Obviously such an important and long-term objective cannot be left to ad hoc arrangements, usually made as a byproduct of regular Bank accivities. Rather it would seem necessary to establish within the Bank a specific unit charged with the responsibility of creating and strengthening economic research capabilities within the development banks, planning institutions and ministries of finance of developing countries. Priority should, in my opinion, be given to small countries with little or no research background; it is in such countries that the Bank could make the greatest impact. Emphasis should be placed on domestic data collection, storage and analysis, as well as on publication of results where appropriate. 5. More Efficient Dissemination of Research Results The question of wider and more effective dissemination of Bank research within the developing countries is highly related to the previous objective of building their domestic research capabilities. At present the research papers and reports distributed by the Bank have rather limited circulation and do not generally get the professional attention they deserve because of inadequate local research facilities and staff. Thus, to achieve adequate progress in the dissemination of Bank research results in developing countries, it would seem necessary to build upthe role of the Bank in the area of technical assistance designed to build up research capabilities in developing countries. It may be desirable to begin by a general survey of the existing situation and actual needs in this field, followed by a comprehensive program in which the role of the Bank would be clearly defined and supported by adequate resources. - 67 - ANNEX 3 Page 7 Another step which could improve the dissemination of informa- tion from the Bank would be to enlarge the list of institutions and agen- Cies receiving Bank publications. Universities, relevant government rnistries and departments and national and regional development institutions s,Ould all be included. Another suggestion in this regard is for the Bank to organize regional seminars and working groups in cooperation with regional organiza- ticns and agencies. The general objectives of such seminars should be to ;Ocus attention on the development problems and needs of each region selected and the role of economic research in helping to identify these problems, as well as to make suggestions for their solutions where appropriate. This suggestion can help mobilize the research staff and agencies in the developing countries and promote contacts between them. It can also in- crease the role of the Bank as a promoter of economic research in develop- ing countries.