Ecuador - Current economic position and prospects (Vol. 3 of 3) : Annexes

f- ~ ~ ~ - - i~~~~~~'~V 11 i; , -_ DOCUMENT OF INTERNATIONAL BANK FOR RECONSTRUCTION AND DEVELOPMENT INTERNATIONAL FINANCE CORPORATION INTERNATIONAL DEVELOPMENT ASSOCIATION Not For Public Use Report No. 14a-EC CURRENT ECONOMIC POSITION AND PROSPECTS OF ECUADOR (in three volumes) VOLUME III ANNEXES June 1, 1963 Latin America and the Caribbean Regional Office 1This report was prepared for official use only by the Bank Group. It may not be published, quoted or cited without Bank Group authorization. The Bank Group does not accept responsibility for the accuracy or completeness of the report.  CURRENCY EQUIVALENTS Current Unit = Sucre (S/) s/1.00 = US$o.oo4 US$1.00 = S/25.oo S/1 million = us$ho,ooo GLOSSARY OF ABBREVIATIONS BEV - Ecuadorian Housing Bank BNF - National Development Bank CENDES - Industrial Development Center CEPE - Ecuadorian State Petroleum Corporation CFN - National Finance Corporation CIDA - Inter American Center for Agricultural Development, COFIEC - Development Finance Company EEQ - Electric Company of Quito ENPROVIT - State Enterprise regulating trade in key commodities ENTEL - National Telephone Company EMELEC - Electricity Company of Ecuador IEOS - Ecuadorian Institute of Sanitary Works IERAC - Ecuadorian Institute of Agrarian Reform and Colonization IESS - Social Security Institute INECEL - Ecuadorian Institute for Electrification INERHI - Ecuadorian Institute for Water Resources INIAP - National Institute for Agricultural Research IPPF - International Planned Parenthood Federation GOVERNMENT OF ECUADOR FISCAL YEAR January 1 to December 31  CURRENT ECONOMIC POSITION AND LONG-TERM PROSPECTS OF ECUADOR LIST OF ANNEXES A. TECHNICAL NOTE ON NATIONAL ACCOUNTS B. TECHNICAL NOTE ON BASIC STATISTICS C. THE PETROLEUM SECTOR D. LONG-TERM MACRO-ECONOMIC PROJECTION MODEL   CURRENT ECONOMIC POSITION AND LONG-TERM PROSPECTS OF ECUADOR TABLE OF CONTENTS VOLUME III Page No. ANNEX A - TECHNICAL NOTE ON NATIONAL ACCOUNTS T. Introdiicti.on ........ . II. Summarv of Major Findinps and Recommendati.ons oeoooesoooo 1 III. Estimation of GDP at Current Market. Prices .. 3 IV. Major Components of the Expenditure on GDP ..**.0**......... 5 Private Consumption * 5 ....................... *0 , General Government Expenditure ...................... 7 Gross Domestic Fixed Investment .1....................., 0 Depreciation .................0........................ 10 Increase in Stocks ................. 0............... ... 10 Exports and Imports ...1*... 05500090 n 0 ^ g of ..... 10 V. Estimation of GNP at Market Prices and the National Incomp 13 Net Factor Payments .... ... 00....................0...... 13 Indirect Taxes and Subsidies 06..........-......0 00.0 17 VI. Industrial Origin of GDP at Factor Cost .*caoo.. ao.oe*. o 19 Agriculture GDP oo0oo..........o..oo ... .... 19 Mining GDP ................. ........ ..... 36 Manufacturinc rYDP o ....... .0000096*000...... 42 nonstruction GDP ...o 0 * 0 9 .... ... 0 0 7 Electricity, Wntir Supply,and Sanitation Services GDP .. 50 Transportation, Communications, and Storage GDP o.... . 1X Wholesale and Retail Trade GDP ...... o.00 S .0..S......... 56 Finance, Tnsuranree. nd Real Estate ............... 56 Ownership of Dwellings ............... ........... 59 Services GDP .............. ..0.. 0.0 ......*...... 59 Public Administration and Defense ...09060060.09 61 VII. The Deflation Problem .......................... . 61  ANNEX A Page ii LIST OF TABLES Table Page No. A-1 GDP at Current Market Prices, Ecuador, Estimating Procedure for 1970-1972 ..... ......................... A-2 General Government and Public Sector Consumption Expendi- ture at Current Market Prices, Ecuador, 1960-1970 ..... . 6 A-3 Public Sector Gross Domestic Fixed Investment at Current Market Prices, Alternative CBE and NEPCB Estimates, by Type of Asset, Ecuador, 1964-1970 .00......... ....... 8 A-4 Estimation of Total and Private Sector Gross Domestic Fixed Investment at Current Market Prices, Ecuador, 1964-1970.. 9 A-5 Gross Domestic Fixed Capital Formation at Current Market Prices, Ecuador, Estimating Procedure for 1969 and 1970.. 11 A-6 Gross Fixed Capital Consumption Allowances, Ecuador, Esti- mating Procedure for 1968-1970 ....o.............. .. 12 A-7 Exports and Imports of Goods and Nonfactor Services, Factor Payments, and Capital Inflow, Comparison of Alternative Estimates, Ecuador, 1960-1971 ................... 14 A-8 Net Factor Payments, Estimating Procedure, Ecuador, 1968-1970 .........**... ...... ...... ......*...... 0* 15 A-9 Net Factor Payments, Comparison of Alternative Estimates, Ecuador, 1960-1970 ....... *.*o......*.o..**..........*. *.* 16 A-10 Net Indirect Taxes, Ecuador, 1960-1970 . 0.................. 18 A-1l Major Agricultural Crops--Volume, Prices, and Value of Pro- duction, Ecuador, Estimating Procedure for 1969-1972..... 20 A-12 Bananas and Plantains--Volume, Prices, and Value of Pro- duction, Consumption, and Exports, Ecuador, Estimating Procedure for 1969-1972 0.... ..... ....... ................ 21 A-13 Cocoa--Volume, Prices, and Value of Production, Consump- tion, and Exports, Ecuador, Estimating Procedure for 1969-1972 .. . . .... *..........o. .o*....* 23 A-14 Coffee--Volume, Prices, and Value of Production, Consump- tion, and Exports, Ecuador, Estimating Procedure for 1969-1972 ....o.oo..o.......o.............o......... ....... 24  ANNEX A Page iii Table Page No. A-15 Minor Agricultural Crops--Volume, Prices, and Value of Production, Ecuador, Estimating Procedure for 1969 ...... 25 A-16 Selected Agricultural Crops--Volume, Prices, and Value of Production, Consumption, and Exports, Estimating Proce- dure for 1969 ....*.... 27 A-17 Gross Value of Meat Production, Ecuador, Estimating Proce- dure for 1969-1972 ............ *........... .... 29 A-18 Production of Milk, Eggs, and Wool, Ecuador, Estimating Procedure for 1969-1972 ........ ... ....... . 30 A-19 Output of Livestock and Kindred Products, Ecuador, Summary of Estimates for 1969-1972 ............. 0..* ......... 31 A-20 Gross Value of Agricultural Production, Ecuador, Summary of Estimates, 1960-1972 ............ ...... 32 A-21 GDP at Current Factor Cost Originating in Agriculture, Hunting, Forestry, and Fishing, Ecuador, Estimating Procedure for 1960-1965 ...... .*.........* 34 A-22 GDP at Current Factor Cost Originating in Agriculture, Forestry, Hunting, and Fishing, Ecuador, Estimating Pro- cedure for Calendar Years 1969- 1972 .................... 35 A-23 GDP at Constant 1960 Factor Cost Originating in Mining, Ecuador, Estimating Procedure for 1965 and 1966 ......... 37 A-24 Independent Variables Used in Regressions for Estimating Mining and Quarrying GDP, Ecuador, Selected Years 1957-1970 ......0...39 A-25 GDP at Current Factor Cost Originating in Mining and Quarrying, Ecuador, Estimating Procedure for 1969 and 1970 40 A-26 Independent Variables Used in Regressions for Estimating Manufacturing GDP, Ecuador, Selected Years 1957-1970 44 A-27 GDP at Current Factor Cost Originating in Manufacturing, Ecuador, Estimating Procedure for 1969-1972 ....0..... 45 A-28 Independent Variables Used in Regressions for Estimating Construction GDP, Ecuador, Selected Years 1957-1970 48  ANNEX A Page iv Table Page No. A-29 GDP at Current Factor Cost Originating in Construction, Ecuador, Estimating Procedure for 1969 and 1970 ...**..5.... A-30 Independent Variables Used in Regression for Estimating Electric Energy, Water Supply, and Sanitary Services GDP, Ecuador, Selected Years 1957-1970 ........ .-............ 51 A-31 GDP at Current Factor Cost Originating in Electric Energy, Water Supply, and Sanitary Services, Ecuador, Estimating Procedure for 1969 and 1970 ....... *. o ** .......*.......*. 52 A-32 Independent Variables Used in Regression for Estimating Transportation GDP, Ecuador, Selected Years 1957-1970.... 5L A-33 GDP at Current Factor Cost Originating in Transportation, Ecuador, Estimating Procedure for 1969 and 1970 .... 55 A-34 Independent Variables Used in Regressian for Estimating Domestic Trade GDP, Ecuador, Selected Years 1957-1970.... 57 A-35 GDP at Current Factor Cost Originating in Domestic Trade, Ecuador, Estimating Procedure for 1969 and 1970 ...... .. 58 A-36 Industrial Origin of GDP at Current Factor Cost, Alterna- tive Estimates of the CBE and NEPCB, Ecuador, 1969, 1970, and 1971 oo.....o ........................ 60 A-37 Deflation Bias in GDP and Its Major Components, Ecuador 1962-1969 ............... 63  ANNEX A Page 1 TECHNICAL NOTE ON NATIONAL ACCOUNTS I. Introduction 1. Annex A presents the findings and recommendations on Ecuador's national accounts; the related basic statistics are reviewed in Annex B below. The discussion focuses on the estimating procedures and alternative estimates produced by the Central Bank of Ecuador (CBE) and the National Economic Planning and Coordination Board (BEPCB). The summary of major findings and recommendations focuses on major problems, the details being given and supported by a more comprehensive discussion and tables in the subsequent sections of this Annex. The recommendations relate to various improvements in the estimating procedures of national accounts. The basic statistical sources and methodology, however, also require a thorough review to establish more meaningful priorities for data collection and processing 2. The Mission is aware of the extremely limited human and budgetary resources available for the statistical work in Ecuador, and especially for the national accounts. The various methodological shortcomings encountered by the Mission are frequently the result of shortcuts and expediency rather than lack of the necessary expertise. Although Ecuador procides little professional training in statistics, it nevertheless has several competent experts in national accounts and basic statistics with varying degrees of foreign training who are aware of the major deficiencies and who would be able to bring about considerable improvements if the necessary resources were made available. II. Summary of Major Findings and Recommendations 3. This summary outlines the methodologies used by the CBE and the NEPCB for deriving the major aggregates of national accounts. The CBE prepares most of the original estimates; the NEPCB updates, projects, and adjusts them for long-range planning and policy purposes. Both agencies use the old system of the UN national accounts. In view of considerable data problems, the new UN system may not be adopted for several years. The immediate improvement of the existing system of national accounts is of greater importance.  ANNEX A Page 2 L4. The CBFE makes global estimates of the GDP at current prices, aerives the major expenditure components of the GDP, and estimates thei GDP by industrial origin. It also prepares in considerable detail the revenue and expenditure accounts of the general government and the external trans- actions, showing in condensed form the other accounts of the old UN system -- all at current prices. Prior to 1968, the CBE used to calculate a general price index for deflating all GDP components. In recent years, apparent deficiencies in consumer price indexes and a complete absence of wholesale price indexes compelled the CBE to suspend temporarily the deflation of national accounts. The NEPCB has carried on the deflation work, even without the most inadequate price deflators, because it needs the national accounts at constant prices for planning purposes. 5. Tn the preparation of national accounts, the CBE has tried to allow and make necessary adjustments for the deficiencies in the basic' statistics. Whenever feasible, the ORE conducted its own surveys and made numerous inquiries every year to collect the data necessary for the national accounts. It employed up to 18 full-time economists in the 1950's for evaluating and incorporating all the basic relevant data into an almost complete set of national accounts. With a gradual deterioration of basic statistics, the burden of providing all the necessary data for the national accounts increased and priorities shifted to other areas. The CBE gradually reassigned its national accounts experts, reducing their number from 18 to only one part-time specialist. 6. Such a drastic staff reduction accompanied by a gradual deter- ioration of basic statistics compelled the CBE to modify its approach to the work on national accounts in the middle 1960's. Instead of aggregating the various components of national accounts from basic data, the CBE started regressing them on several readily available time series. Moreover, to cut corners, the OBE consolidated many detailed breakdowns and less impor- tant series into broader aggregates, finally abandoning all attempts to deflate the national accounts. While the use of regressions may produce sufficiently meaningful extrapolations for a few years, this shortcut method often results in misleading and estimates -- of which the Ecuadorian experience has contributed several examples pointed out below. Their number wouldi have been larger if the CBE experts had not employed various tests of consistency and reasonableness by deriving some estimates from alternative sources. 7. The National Institute of Statistics (NIS) has recently been reorganized into an autonomous agency associated with the NEPCB although it still appears to be weak in modern sampling techniques and data prccessing. At present, it has no staff for constructing the estimates for the national accounts. However, after the NIS has been sufficiently strengthened to produce meaningful, relevant, and timely statistics, it should establish a National Accounts Division (NAD), which would assume the responsibility for  ANNEX A rage 3 the national accounts work of the CBE and the NEPCB. As a major user of basic statistics, the NAD would exert a beneficial influence on the NIS priorities and standards with respect to its basic statistics (see Annex B for further details). 8. Price and production indexes should be improved and better adapted to the deflation needs of national accounts. The present deflation based on CPI's is inadequate. Each component of national accounts should be deflated with the most appropriate component of the CPI or the WPI. The price indexes of various cities should be aggregated, by components with proper regional weights, into national CPI and WPI. Whenever appropriate, industrial and agricultural production indexes should be computed and used for extrapolating the corresponding components of the national accounts at constant prices. The estimates at current prices would then be obtained by means of specific price indexes. 9. The following discussion presents a general picture of the CBE and NEPCB methodologies with selected illustrations drawn from recent experience. The Mission tried to use the same data and methodology with a few changes whenever they produced significantly improved results. The primary objective of the Mission in the area of national accounts has been to ascertain the basis for the present estimates rather than to develop a new set of national accounts. The latter undertaking would not have been feasible with the time and basic data at the disposal of the Mission. III. Estimation of the GDP at Current Market Prices 10. For 1969 and 1970, the CBE has estimated the GDP at current market prices by simple linear regression from the average monthly means of payment at the disposal of the public. The latter variable is readily available, and it has shown a remarkably stable relationship -- even for the years when the two variables were estimated independently -- of about 11 percent of GDP in the 1950's and 12 percent for most of the 1960's (see Table A-1), implying a slight decline in the velocity of the circulation of money. Such a secular decline has been observed in many countries. The expansion of the monetized sector tends to replace the barter economy of the subsistence sector. This has a double effect. More money is needed: (1) for transactions -- as a medium of exchange, and (2) for value transfers in time and place -- as a store of value. Instead of goods, means of payment are used for accumulating wealth. Thus, relatively, more money is needed per value unit of the GDP. Therefore, the increasing ratio of the money supply to the GDP is very plausible for Ecuador. 1L The preliminary GDP estimate extrapolated by the regression on the means of payment is then adjusted on the basis of changes in foreign trade, current and capital expenditure of the public sector, construction permits, output of selected industrial products, and various  ANNEX A Page a Table A-i: GDP AT CIJRRENT MARKET PRICES, ECUADOR, ESTIMATING PROCEDWURE FOR 1970-1972 GDP at Current Means of Payment Means of Market Prices at the Disposal of the Public Payment (Millions (Monthly Averages in (Percent Year of Sucres) Millions of Sucres) of GDF) Y x % 95o 7,2I5 803 11 1951 7,761 866 11 1952 8,854 947 11 1953 9,349 1,080 12 1954 10,447 1,205 12 1955 11,049 1,215 11 1956 11,266 1,239 11 1957 12,007 1,311 11 1958 12,357 1,362 11 1959 13,009 1,474 11 1960 14,140 1,627 12 1961 15,075 1,731 11 1962 16,104 1,840 11 1963 17,437 2,054 12 1964 19,414 2,362 12 1965 20,787 2,411 12 1966 22,834 2,615 11 1967 25,128 2,962 12 1968 27,466 3,398 12 1969 30,717 3,693 12 1970 37,575 4,5h9 12 1971 64,673 5,435 12 1972 53,205 6,500 12 The above variables produced the following estimating equation for GD]? at current market prices: Yi = 1,130.33 + 8.01147Xi This equation produced the following estimates: Y1970 = 3O,574 5 Y1971 = 44,672.6 Y1972 - 53,204.9 The CBE computed a slightly different estimating equation (Yi = 1,130.13 + 8.011311Xi)1 Source: Central Bank of Ecuador, National Income Division; Mission estimates.  ANNEX A Page 5 other indicators. In addition, the CBE takes into consideration expert appraisals produced by various government agencies and international organizations. The integration of all this information in a GDP estimate cannot be objective. It usually results in several tentative hypotheses, the most plausible of which is finally accepted. This estimating procedure can be illustrated by GDP estimates prepared for recent years. 12. Given the inflation rate by the Quito CPI of 5.0 percent in 1970 and the increase in the GDP at current market prices of 21.6 percent implied by the regression on the means of payment, the real rate of growth of 15.8 percent appeared to be on the high side. Making allowance for higher price changes, the CBE estimated the price deflator at 7.1 per- cent for 1970. However, even with this higher price deflator, the real growth of GDP comes to 12.4 percent for 1970. In order to bring it down to a more reasonable level of about 8 percent, the CBE reduced the GDP at current market prices from 37,574 (as given by the regression) to 35,7h3 million sucres. For 1969, the preliminary GDP estimate of 30,317 million sucres implied only 1.85 percent rate of growth in the per capita national income. This estimate appeared on the low side, and four other alternative estimates with higher growth in the per capita income were considered (30,843; 3110h8; 31,282; and 36,112). The lower of the latter four estimates appeared most plausible, and it was finally adopted for 1969. Similar considerations determined the estimates of GDP at current market prices in 1968, 1967, 1966, and 1965. For 1964 and earlier years, the GDP was estimated by the aggregation of value added by sectors. These procedures are discussed below. 13. The NEPCB estimated the GDP by converting exports and imports from the balance of payments dollar values with different exchange rates, and although it used essentially the same consumption and investment estimates as the CBE, its estimates of imports became larger and those of GDP somewhat smaller. The details of these NEVCBestimates are discussed below. IV. Major Components of the Expenditure on the GDP 14. Private consuption-. This GDP component is derived as a residual by subtracting gross domestic capital formation from the available resources. In the 1950's, the CBE tried to check this component by estimating independently private consumption expenditure. The absence of necessary data made such estimates of private consumption expenditure almost impossible, and their preparation has been suspended. 15. General goverrment expenditure. The Nntiennl Income Division (NID) of the CBE estimates general government consumption expenditure from the accounts of Central, provincial, and municipal governments. The consumption expenditure of the Ecuadorian Social  Table A-2: GENERAL GOVERNMENT AND PUBLIC SECTOR CONSUMPTION EXPENDITURE AT CURRENT MARKE.T PRICES, ECUADOR, 1960-1970 (Millions of Current Sucres) Consumption Expenditure 1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 Central Bank of Ecuador Estimates Employee Compensation General government 1,158 1,166 1,252 1,373 1,576 1,766 2,008 2,091 2,751 3,234 n.a. Plus: other entities -48 91 74 48 55 59 66 81 97 110 n.a. Public sector 1,110 1,257 1,326 1,421 1,631 1,825 2,074 2,172 2,848 3,344 n.a. Purchases of goods and servicesi/ General government 655 901 903 859 1,014 1,072 1,211 1,311 1,308 1,326 n.a. Public sector 576 739 751 761 896 906 830 902 1,259 1,312 n.a. Consumption expenditure General government 1,813 2,067 2,155 2,232 2,590 2,838 3,219 3,402 4,059 4,560 4,980 Public sector 1,686 1,996 2,077 2,182 2,527 2,731 2,904 3,074 4,107 4,656 n.a. National Economic Planning and Coordination Board Estimates Public sector consumption *- *- 2,387 2,628 2,809 3,158 3,743 4,171 5,115 1/ Purchases of goods and services exclude purchases from within the specified sectors Therefore, the general government purchases (and consumption expenditure) may exceed those of the public sector as a whole. Sources: Central Bank of Ecuador, National Income and Fiscal Studies Divisions; National Economic Planning and Coordination Board. a>4  ANNEX A Page 7 Security Institute is also included while the consumption expenditure of government enterprises and public corporations is excluded from general government and implicitly included in the private sector. These classi- fications are consistent with the UN system of national accounts. 16. The Fiscal Studies Division of the CBE and the NEPCB h we also estimated public consumption which includes over 500 autonomous entities, government enterprises, and public corporations. According to the CBE estimates, general government accounts for about 97 percent of the employee compensation of the public sector. The general government purchases of goods and services exceed those of the public sector. This is possible because the intrasector transactions are excluded and the general government agencies make purchases from the rest of the public sector. The latter are included in the general government but are excluded from the public sector (see Table A-2). The NEPCB has also estimated public sector consumption which tends to be lower (except in 1967 and 1970) than that estimated by the OBE (see Table A-2). The methodological details of the latter estimates are not available. 17. Gross domestic fixed investment (GDFI). The National Income Division of the CBE estimates first the public GDFI from the data compiled by the Fiscal Studies Division for the public sector as a whole (which includes government enterprises and public corporations). The latter Division classifies annual expenditure of each entity, ageregating new GDFI by major components. The National Income Division adds financial investments of the public sector in land, buildings, and other real estate to the new GDFI to obtain its own estimate of gross domestic fixed capital formation which it uses in the national accounts (the UN system of national accounts excludes financial investment). 18. The NEPCB makes further upward adjustments in the public GDFI, including several additional autonomous entities as well as the deferred payments for financinp public GDFI of the praceding year (see Table A-3). The methodological details of these adjustments are not available. A comparison of the CBE and the NEPCB estimates of GDFI by type of asset shows that the NEPCB estimates tend to be higher than those of the CBE for other construction although they tend to be lower for machinery and buildings (see Table A-3). These differences may partly be attributed to the different coverage of autonomous agencies although an exhaustive comparative study of the two sets of data for all of the 500 autonomous entities could not have been undertaken within the scope of the present Mission. 19. The CBF. estimates the private rDFI as a residual by subtracting the public GDFI from the total GDFI. The NEPCB adopts the CBE estimate of the total GDFI and derives the rrivate GDFI by subtracting its own estimate of public GDFI (see Table A-h). Thus, both agencies use the same total GDFI and different estimates for the private and public sectors.  ANNEX A Page 8 Table A-3: PUBLIC SECTOR GROSS DOMESTIC FIMBD INVESTIC3NT AT CU-RRENT iVARKET PRICES, ALTERNATIVE CO.E AND N&PCB ES?TIMATES, BY TTP- OF ASSET, ECTIADOR, 196h-1970 (Millions of Current %ucres) 1964 1965 1966 1967 1968 1969 1970 Public new GDFI 875 886 844 998 1,2242 1,522 1,874 Plus: land, etc. 71 47 38 130 26 75 n.a. GDFI in CBE national accounts2/ 946 933 882 1,128 1,268 1,597 n.a. Plus: adjustment 3/ 7 111 245 154 243 38 n.a. GDFI inNEPCB national accounts/ 953 1,044 1,127 1,282 1,511 1,&35 2,013 Public new GDFI, by type of asset, total: CBE 875 886 844 998 1,242 1,'22 n.a. NEPCB 908 1,009 1,093 1,246 1,371 1,L453 1,846 Machinery and equipment: CBE 144 188 171 166 206 239 n.a. NEPCB 123 170 204 219 159 05 394 Buildings: CBE 150 113 94 134 206 '245 n.a. NEPCB 87 78 49 91 151 120 169 Other construction: CBE 581 585 579 698 830 1,038 n.a. NEPCB 698 761 82o 936 1,061 1,128 1,283 1/ Includes CBE estimates of land, buildings, and other real estate financial investments of the public sector. 2/ CBE estimates of public GDFI used in CBE national accounts. 3/ Includes deferred payments for financing public GDFI of the preceding yea' and additional autonomous entities. 4/ NEPCBestimates of public GDFI used in NEPCBnational accounts. Source: Central Bank of Ecuador; National Economic Planning and Coordination Board; Mission estimates.  ANNEX A Page 9 Table A-4: ESTIMATION OF TOTAT, AND PRIVATE SECTOR GROSS DOMESTIC FIXED INVESTMENT AT CURRENT MARKET PRICES, ECIJADOR, 1964-70 (Millions of Current lucres) 1964 1965 1966 1967 1968_/ 19691/ 1970a/ Central Bank of Ecuador Estimates Regressed total GDFI 2,339 2,406 2,515 2,969 3,272 3,963 4,296 CBE adjustment - - - - 502 781 n.a. CBE total GDFI 2,339 2,406 2,515 2,969 3,774 4,744 n.a. Less: CBE public GDFI 946 933 882 1,128 1,268 1,597 n.a. CBE private GDFI 1,393 1,473 1,633 1,841 2,506 3,147 n.a. National Economic Planning and Coordination Board Estimates CBE total GDFI 2,339 2,406 2,515 2,969 3,774 4,744 6,298 Less: NEPCBpublic GDFI 953 1,044 1,127 1,282 1,511 1,635 2,813 NEPCBprivate GDFI 1,386 1,362 1,388 1,687 2,263 3,109 3,485 1/ Preliminary. Source: Central Bank of Ecuador, National Income Division; National Economic Planning and Coordination Board  ANNEX A Page 10 20. The CBE estimates the total GDFI by regression. It first converts the GDFI from the current into constant prices, expresses the latter as an index, and regresses it for 1969-70 on three related indexes: (1) GJDFOF of the general government (which is not available and is replaced with public sector GDFCF), (2) unweighted volume index of cement production, and (3) quantum index of imported capital goods (see Table A-5). The latter index has not been available since the middle 1960's and it had to be estimated from other data on imports. It was apparently lagged one year although this could not be established with certainty because the computational details of this quantum index are not available. It was computed from the volume of imported capital goods and the 1960 unit price. The cement production was similarly calculated from the metric tons and the average 1960 cement price. The results at constant prices were converted to index numbers which the CBE used in the regression for deriving GDFCF estimates as far back as 196h. 21. Depreciation. The CBE estimates the provision for fixed capital consumption as a percentage of preceding year's GDFCF. The latter are further adjusted by the statistical discrepancy in the reconciliation of the account showing the finance of gross capital formation. In the latter account, the CBE reconciles the depreciation with the surplus or deficit of the nation on current account, gross domestic saving, and the gross capital formation. The depreciation allowanceshave thus been sig_lnificantly adjusted, especially for recent years (see Table A-6). 22. Increase in stocks. The CBF has estimated the change in stocks from the data of the Finance Ministry for 1950-53. It first estimated increases in stocks at current prices, deflated them with the GDP deflator, and calculated an average annual increase per capita. In the 1960's, the NID has used the changes in the volume of exports and imports rather than the population growth. The changes in stocks have been further adjusted in recent years to show smooth and gradual increases. 2 3 Exports and imports. The CBE National Income Division derives exports and imports of goods and nonfactor services from the balance of payments data in US dollars. It converts each entry at the exchange rate applicable to each foreign exchange market. Thus, in 1968 and 1969, the CBE distinguished between exchange rates in the official market (17.82 sucres for exports and 18.18 sucres per US dollar for imports) and those in the free market (22.15 and 22.26, res- pectively). For the first half of 1970 (through June 21), it used 21.55 and 21.71 sucres per US dollar for the free market. From June 22 till Aupust 17, in the so-called "parallel market", the CBE used 23.20 as a buying and 23.43 sucres as a selling rate, and from August 17 till the end of 1970, the CBE used again two sets of rates: 2h.75- 25.25 sucres for the "unified" and 27.00-28.00 sucres for the "black" market. Aggregating separately the dollar and the sucre amounts, the  ANNEX A Page 11 Table A-5: GROSS DOMESTIC FIXED CAPITAL FORMATION AT CURRENT MARKET PRICES, ECUADOR, ESTIMATING PROCEDURE FOR 1969 AND 1970 Gross Domestic Fixed Capital Formation Independent Variables (Millions of Sucres) (Indexes, 1965 = 100) Current At 1965 Index Year Prices Prices Y X1 X2 X3 1957 1,561 1,920 79.8 61.0 52.2 47.7 1958 1,516 1,853 77.0 64.2 53.9 52.2 1959 1,734 2,117 88.0 81.1 52.9 47.8 1960 1,897 2,272 94.4 97.6 61.6 60.1 1961 2,047 2,337 97.1 102.8 67.0 58.4 1962 1,959 2,186 90.9 84.8 65.3 46.6 1963 2,146 2,300 95.6 89.0 79.1 58.0 1964 2,339 2,426 100.8 101.4 89.2 80.8 1965 2,406 2,406 100.0 100.0 100.0 100.0 1966 2,515 2,393 99.5 94.5 115.8 107.0 1967 2,969 2,726 113.3 120.9 131.6 126.5 1968 3,272 2,893 120.2 133.0_/ 1h6.1 140.2 1969 3,963 3,289 136.6933 171.2 153.5 153.3 1970 4,296 3,330 138.4013 176.3(p) 154.2 158.7 1/ The index value computed from the CBE data for 1968 is 135.9. Independent Variables: X1 = GDFCF of the general government X2 = Cement production X3 = Quantum index of imported capital goods. The above variables produced the following estimating equation for GDFCF at constant 1965 factor cost: Yi = 45.O670 + 0.4331X1 * 0.2375X2 - 0.1216X3 This equation produced the following final estimates: Y1969 = 3,289 Y1970 = 3,330 The CBE computed the same estimating equation. Source: Central Bank of Ecuador, National Income Division.  Table A-6: GROSS FIXED CAPITAL CONSUMPTION ALLOWANCES, ECUADOR, ESTIMATING PROCEDURE FOR 1968-1970 (Millions of Current Sucres and Percent) 1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 GDFCF 1,897 2,047 1,959 2,149 2,339 2,406 2,515 2,969 3,774 4,744 6,298 Unadjusted GFCCA (as 43% of the GDFCF in preceding year) - 816 880 842 924 1,005 1,035 1,081 1,277 1,623 2,040 Plus: reconciliation adjustment- -53 -98 -25 103 62 78 69 -67 -398 -440 Adjusted GFCCA 684 763 782 817 1,027 1,067 1,113 1,150 1,210 1,225 1,600 GFCCA in percent of adjusted GDFCF of the preceding year - 40.2 38.2 41.7 47.8 45.6 46.3 45.7 40.8 32.5 33.7 Sources: Central Bank of Ecuador, National Income Division; Mission estimates.  ANNEX A Page 13 CBE has derived for 1970 the average weighted exchange rates of 21.33 for exports and 21.h9 for imports, with an overall weighted average of 21.h2 sucres per US dollar. 24. The Foreign Trade Section of the NEPCB has used somewhat different exchange rates and made other adjustments in converting the balance of payments data from US dollars into sucres. Further discre- pancies arise from the different treatment of services. In addition to investment income, the NID classifies services and commissions on loans and donations as factor payments; the balance of payments and the NEPCB include these as nonfactor services with imports. To this extent, the NID tends to understate the imports and to overstate the net factor payments. The net capital inflow data show relatively smaller discre- pancies than its two major components (see Table A-7). Using the balance of payments data and the NID weighted average exchange rate for imports, the Mission estimated the total value of imports in sucres which approxi- mates the NEPCB estimates (see Table A-7). The same procedure produced also the NEPCB estimates of exports. V. Estimation of the GNP at Market Prices and the National Income 25. The CBE and the NEPCB estimate the GNP at current market prices by deducting net factor payments to abroad from the GDP at current market prices. Subtracting the net indirect taxes from the GNP is obtained (see Tables 2.1 and 2.2 of the Statistical Appendix to the main report). The factor payments are overstated and the indirect taxes understated by items which are classified differently by the IBRD, IMF, and the UN system of national accounts. Tho major discrepancies are discussed below. 26. Net factor payments. The CBE derives factor payments from the balance of payments data, converting the US dollars into sucres at various exchange rates (similarly to the foreign trade data). The items classified as investment income payments include dividends and interest on direct investment as well as other dividends and interest (IMF charges, interest on external debt, bank interest, and net portfolio investment income). Moreover, real estate transactions  Table A-7: EXPORTS AND IMPORTS OF GOODS AND NONFACTOR SERVICES, NET FACTOR PAYMIENTS, ANJ NET CAPITAL INFLOW, COMPARISON OF ALTERNATIVE ESTIMATES, ECUTADOR, 1960-71 Alternative S2UceA. _ 1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 Exports of Goods and Nonfactor Services in Millions of Current Sucres NID (CBE) exports l 2,530 2,524 3,081 3,024 3,245 3,618 3,726 4,041 4,258 4,183 5,437 Plus: discrepancy -6 +106 -34 -28 -34 -45 +10 -44 -37 -19 -43 NEPCB exportsl/ 2,524 2,630 3,047 2,996 3,211 3,573 3,736 3,997 4,221 4,164 5,394 6,-i22 Mission estimates 2,524 2,630 3,047 2,99(, 3,211 3,573 3,736 3,997 4,227 4,390 5,463 6,422 Imports of Goods and Nonfactor Services in Millions of Current Sucres NID (CBE) importsl/ 2,476 2,581 2,929 2,940 3,334 3,574 3,604 4,209 5,134 5,842 7,043 Plus: discrepancy -21 +167 +9 -10 +130 +104 +127 +172 +214 +486 +552 NEPCB importal/ 2,455 2,748 2,938 2,930 3,464 3,678 3,731 4,381 5,348 6,328 7,595 11,642 Mission estimates 2,455 2,748 2,938 2,930 3,46h 3,678 3,731 4,381 5,362 6,301 7,760 11,642 Net Imports orf Goods actor Services in Millions of Current Sucres NID (CBE) -54 57 -152 -.54 59 -44 -122 165 876 1,659 1,606 Plus: discrepancy -15 61 43 18 164 149 117 216 251 505 595 NEPCB -69 118 -109 -66 253 105 -5 384 1 ,127 2,164 2,201 Mission estimates -69 118 -109 -66 253 105 -5 384 1,135 1 , 911 2,297 5,220 Net Factor Payments i n =r Sucres NID (CBE) 3-9 460 439 333 455 566 613 662 751 9d5 1,432 Plus:discrepancy -3 5 -4 0 -131 -101 -119 -172 -177 -392 _666 NEPCB 395 465 435 333 357 465 494 490 574 593 766 Mission estimates 395 465 435 333 357 465 494 490 572 593 727 865 Netj ai nffo-nM iTillions of Current Sucres NID (CBE) 344h 17 287 249 577 522 491 530 1,627 2,644 3,038 5 s Plus:discrepancy -1 8 66 39 18 33 48 -2 44 74 113 -71 . NEPCB 326 583 326 267 61o 570 489 874 1,701 2,757 2,967 Mission estimates 326 583 326 267 610 570 489 874 1,707 2,504 3,024 6,085 1. Includes goods and nonfactor services Sources: NEPCB, _orqi?n Trade Yearbook (for 1957 through 1969), Central Bank of Ecuador; Mission estimates.  Table A-8: NET FACTOR PAYMENTS, ESTIMATING PROCEDURE, ECUADOR, 1968-1970 (Millions of Current Sucres) 1968 1969 1970 1. Investment payments 570.3 592.6 727.2 a. Dividends and interest on direct investment 441.4 45.2 513.8 b. Other dividends -10.7 -7.5 15.8 c. Interest on external debt 138.2 141.8 194.3 d. Bank interest 7t3 e. IMF charges 3.6 3.6 5.5 f. Portfolio investment -2.2 -2.2 g. Net lease, rent, and income of the US embassy in Quito 2.2 2. Other factor payments 1/ 1.8 3. Net factor payments: (1) Mission estimates 572.1 592.6 727.2 (2)NEPCB estimates 57h 593 766 4. Nonfactor payments 21 179.3 392.6 704.8 a. Services and commissions on direct investment .. 219.5 460.2 b. Services and commissions on public debt 86.8 c. Services and commissions on loans 83.6 43.6 122.5 d. Services and commissions on donations 129.5 119.9 e. Film rentals 8.9 2.2 5. Net factor payments: CBE estimates 751.4 985.2 1,432.0 1/ Comprises payments to foreign technicians. 2/ The CBE classifies these non-factor services as factor payments. X I Sources: Central Bank of Ecuador, Department of Economio Research; National Economic Planning and Coordination Board.  Table A-9: NET FACTOR PAYMENTS, COMPARISON OF ALTERNATIVE ESTIMATES, ECUADOR, 1960-70. Alternative Sources -r55il --l96 -1972r 5-13 1 96 196- 19-5 1967 1968 1969 1970 Millions of US Dollars Balance of payments investment income 22.8 23.2 19.9 16.7 19.3 25.0 25.2 24.9 27.1 28.4 33.5 Millions of Sucres Balance of payments investment income2/ 345 422 362 304 351 455 458 453 493 516 609 NEPOB net factor payments 395 465 435 333 357 465 494 490 574 593 766 CBE net factor payments 398 460 439 333 488 566 613 662 751 985 1,432 / Based on the official selling rate of 15.15 sucres per US dollar for 1960 and 18.18 sucres for 1961-70. Sources: IMF Balance of Payments Division, Balance of Payments Yearbook Vols. 17-22 (for 1966-1970); OBE7 Menoria (issues from 1962 through 1786)T ad unpublished esi t'es for 1968-1970; National Economic Planning and Coordination Board; Mission estimates.  ANNEX A Page 17 such as the lease, rent, and income of the US embassy in Quito are also included in investment income. Other factor p 'ments include payments to foreign technicians. No estimate is included for %luadorian labor income earned abroad because it is believed to be negligib e. The sum of the above net amounts approximates net factor payments isee Table A-8). 27. The CBE has included certain other items as factor payments which appear in the unpublished statement of the balance of payments. A detailed breakdown and description of these additional items is not available, and their classification as factor payments may not be established with certainty. They include services and commissions on direct investment, public debt, loans, and donations. To some extent these finance charges are akin to other returns to capital, and they could be classified as factor payments. However, the IMF and the IBRD normnally classify them with non-factor services rather than with factor payments. Payments for film rentals belong also to nonfactor services (see Table A-8). 28. The net factor payments of Ecuador thus consist mostly of investment income which appears in the balance of payments. Relatively higher amounts in sucres and a higher implicit exchange rate for the NEPCB factor payments indicate that they are more inclusive than those shown in the balance of payments (see Table A-9). Prior to 196h, the CBE and the NEPCB made relatively small adjustrnerts and they differed insignificantly. Starting in 1964, however, the National Income Division of the CBE began including considerable additional amounts into net factor payments. This dicrepancy increased from 37 percent in 1964 to 87 percent in 1970 (see Table A-9). The NEPCB estimates are, on the other hand, close to those prepared by the Mission for 1968-70, and in the absence of more detailed data, they have been adopted also for ea-rlier years. 29. Indirect taxes and subsidies. The UN system of national accounts includes as indirect taxes all taxes assessed on producers for the production, sale, purchase or use of goods and services which they charge to production expense. The indirect taxes estimated by the CBE comprise import duties (including consular rights and import permit stamps), consumer taxes (including net value of consumption for monopoly products), taxes on transactions (excise taxes and registration fees), and other indirect taxes (see Table A-10). The UN system of national accounts requires that the operating surplus of fiscal and similar government monopolies be reduced by the normal (or average) profit margin earned by private business firms in the same industrial branches. It appears that the available data do not permit such adjustments in Ecuador. Moreover, the CBE classifies the export taxes as direct, following apparently tlle argument that Ecuadorian exporters face perfectly elastic demand in the world markets for their products. Being unable to raise prices by the amount of export taxes, the exporters  Table A-10: NET INDIRECT TAXES, ECUADOR, 1960-1970 (Millions of Current Sucres) Taxes and Subsidies 1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 Central Bank of Ecuador Estimates Import taxes 6h2 655 669 861 1,060 1,018 1,201 1,568 1,703 1,739 n.a. Consumer taxes 2/ 437 43h 418 h67 557 454 h32 h62 533 686 n.a. Transaction taxes - 35 36 38 41 h7 45 53 6h 75 75 n.a. Other indirect taxes 1914/ 172 193 198 2h8 265 285 325 379 370 n.a. Total indirect taxes 1,305;-/ 1,297 1,318 1,567 1,912 1,782 1,971 2,h19 2,690 2,870 n.a. oess: subsidies 20 60 106 43 96 117 139 97 90 124 n.a. 6 Net indirect taxes 1,28 5 1,237 1,212 1,52h 1,816 1,665 1,832 2,322 2,600 2,7h6 n.a. NEPCB Estimates Net indirect taxes 1,285 1,237 1,212 1,52h 1,816 1,665 1,832 2,322 2,600 2,7h6 3,030 Mission Estimates Export taxes 175 212 245 267 360 290 276 282 307 301 n.a. Total adjusted net indirect taxes l,4O 1,449 1,457 1,791 2,176 1,955 2,108 2,604 2,907 3,0o7 n.a. l/ Includes consular rights and import permit stamps. 2/ Includes net consumption value of state monopoly products. 3/ Includes certain excise taxes and registration fees. 4/ For 1960, the FSD of the CBE estimated other indirect taxes at 171 mil. sucres and total indirect taxes at -l 1,285 mil. sucres but the NID raised the latter to 1,305 mil. sucres, apparently to keep the net indirect z taxes of 1,285 mil, unchanged after 20 mil. sucres of subsidies had been estimated. C J Sources: Central Bank of Ecuador and National Economic Planning and Coordination Board.  ANNEX A Page 19 charge these taxes to profits rather than to production expenses. This controversial issue has been raised in several countries, and it remains open for discussion in Ecuador. The exclusion of export taxes understates the GDP at market prices. 30. The NE?PC1 uses essentially the CBE estimates of indirect taxes and subsidies (see Table A-10). Subsidies are relativelv small in Ecuador. They include tax credit certificates which can be credited to future tax payments. Subsidies also include the cost of pesticide fumigation of banana plantations. VI. Industrial Origin of the GDP at Factor Cost 31. The CBE estimates the industrial origin of the GDP by regressing each GDP component on selected indicators. The NEPCB adjusts these estimates further for consistency with its estimates cf exports, imports, and the GDP. The NEPCB adjustments are relatively minor. The following discussion focuses on the CBE methodology for deriving the GDP by industrial origin. 32. Agriculture GDP. Agriculture is still by far the largest economic sector in Ecuador, employing more than half of its people. Its relative importance in total GD? has declined, however, from 37 percent in 1960 to 31 percent in 1970. Agriculture GD? includes estimates for 13 major and 59 minor crops, livestock products, poultry, and eggs. Rough estimates for h'intinv-, forestry, and fishing are also included. The volume of crop production, which accounts for about 70 percent of the gross value of output in this sector, is estimated from the 195h Census of Agriculture and the Ministry of Production data based on the annual reports of agricultural extension workers. For the 1950's and most of the 1960's, the CHB has estimated first the gross value of output and then derived the value added by subtracting the nonfactor inputs. Since 1969, the CBE has used regressions for deriving the agriculture GDP. 33 . The COB estimates the gross value of agricultural production for 13 major crops, which account for over 80percent of the total value of all agriculltural crops, multiplying the volume of output by estimated prices of each commodity (see Table A-Il). Special adjustments are made for major exports. Thus, for banana and plant6ins, which account for about one third of the total value of all crops, their gross value is estimated separately for exports and for domestic consumption (see Table \-12). Inasmuch as the export permit declarations are understated, the C3E has used the estimates of the National Banana Bureau (NBB). The CBE has also adjusted upwards the banana export prices for most of the 1960's. Given the volume of domestic production and exports of bananas  Table A-i1L: MAJOR AGRICULTURAL CROPS--VOLUiSE, ?tICES, AND VALUE OF PRODUCTION, ECUADOR, ESTDhIATING PROCEDURE FOR 1969-1972 1 _69 1970 __971_ 1972 Agricultural Metric Prices PMetric rices Crops Tons Sucres per Million Tons Sucres per Million Million Million (000) Quintalj M. Ton1 Sucres (000) QuLotal M. Tor Sucres Sucres Sucres 13 ljaior Crops 1. 3antna.s2 4,316.0 b/ - b/ 3,314.5 bl b/ b/ 3,740.7 3,834.9 3,910.3 2. 3arLe. 77.7 90.98 1.977.9 153.u 110.0 90.3 1-973.0 215.9 250.0 280.0 3. 3eans 37.9 307.48 6,3684.6 253.3 41.3 307.6 6,687.0 276.2 310.0 340.0 4. Castor beans 23.8 b/ b/ 45.2 16.5 51 .3 1.995.7 32.9 35.0 40o.o 5. Cocoa 48.0 b/ b/ 614.8 b/ b/ b/ 5-9.6 7D1.6 715.5 6. Coffee 55.9 7/ b/ 643.3 I b/ 7/ b/ 1,239.3 1,175.3 1,222.0 7. Corn 222.5 7/ 533.8 271.5 107.C 2,377.8 637.4 720.0 810.0 8. Cotton (raw) 23.6 176.70 j3,841 .5 90.5 7.6 193.7 4,210.9 32.0 46.0 60.0 9. Onions 96.0 111.23~~~~~~~~~~~~~~~~~~~5 4 2,418.69495320 9. Onions 28.0 151.25 3288.2 232.1 94.9 118.8 2,582.6 245.1 280.0 320.0 i'. Padd4C/ 288.0 151.25 13,288.2 947.0 407.2 161.4 3,508.7 1,428.7 1,500.0 1,700.0 11. Potatoes 45t.7 78.aS '1,727.2 783.8 541.8 85.8 1,865.2 1,010.6 1,130.0 1,270.0 12. Sugar cane 11,886.6 .. 87.8 1,043.6 (9,500.0 .. (90.0) 855-0 1,060.2 1,200.0 13. Wheat 94.1 118.08 2,567.1 241.6 81.0 .. (2,700.0) 218.7 250.0 300.0 Major crops 17,626.8 .. .. 8,902.1 .. .. 10,532.1 11,353.0 12,173.8 59 Minor Crops Minor crops 1,130.8 .. .. 2,163.5 *- . . .. 2,136.7 2,425.0 2,800.0 _'i2Agricultural Crops Total 18,757.6 _0 .. 11,065.6 .. .. . 12,668.8 13,788.0 14,973.8 L/ Includes plantains. V See detailed calculations in tables below. N) c/ Outpit in metric tons adjusted to the level of output reported by the National Committee for Rice. Sources: Central Bank of Ecuador, National Income Division; M4inistry of Production, General Bureau of Planning, Division of Statistics; Mission estimates.  TableA-l1 : BANANAS AND PLANTAINS--VCLUME, PRICES, AND VALUE OF PRODUCTION, CONSUMPTION, AND EXPORTS, ECUADOR, ESTIMATING PROCEDURE FOR 1969-1972 Jan. -July Aug.-Dec. Items 1969k 1970 1970 1970 1971 1972 Volume in Thousands of Metric Tons Domestic production a/ 4,316.0 o. . 4,136.7 3,805.8 3,800.0 Less: '13B, exports 1,189.7 .. 1,364.0 1,351.0 1,319.0 Domestic consumption 3,126.3 *- * 2,772.7 2,454.8 2,481.0 Domestic and Export Prices Domestic price per 30 kgs. bunch (sucres) 12.71 .. 11.50 12.60 13.80 Domestic price per metric ton (sucres) 423.67 .. .. 383.33 420.00 460.00 Export price per metric ton, unadjusted (US$) 56.03 .. 69.12 74.87 75.82 Plus: export price adjustment (US$) 30.84 .. .. 20.89 8.15 8.33 Export price per metric ton, adjusted (US$) 86.87 .. .. 90.01 83.02 84.15 -~ ~ I Value of Production, Consumption, Exports, and Balance of Payments Adjustment Exports (thousands of US$) 66,653 44,532 49,745 94,277 101,154 100,000 Exchange rate (sucres per us$) 17.82 17.82 24.75 21.48 25.00 25.00 Unadjusted exports (mil. of sucres) 1,187.7 793.5 1,231.2 2,024.7 2,528.9 2,500.0 Balance of payments adjustment (thousands of US$) 36,700 24,492 4,008 28,500 11,000 11,000 Exchange rate (sucres per US$) 21.86 21.86 24.75 22.92 25.00 25.00 Balance of payments adjustment (millions of sucres 802.3 554.0 99.2 653.2 275.0 275.0 Adjusted exports (thousands of us$) 103,353 69,024 53,753 122,777 112,154 111,000 Adjusted exports (millions of sucres) 1,990.0 1,347.5 1,330.4 2,677.9 2,803.9 2,775.0 Domestic consumption (millions of sucres) 1,324.5 . .. 1,062.8 1,031.0 1,141.3 Production (millions of sucres) 3,314.5 .. .. 3,740.7 3,834.9 3,916.3 a/ Includes plantains. CD Sources: Central Bank of Ecuador, National Income Division; Ministry of Production, General Bureau OI Plarming, H > Division of Statistics; Mission estimates.  ANNEX A Paga 22 and plantains, the CBE derives the implicit domestic consumption and values it at unweighted average consumer prices reported by the Ministry of Production (see Table B-15), supplemented by estimates of its own (see Table A-15)1/ The prices for bananas, plantains, and other agricultural crops are generally higher than the farm-gate prices. Thus, the gross value of production includes transportation costs and trade margins. 34. The CBE follows a similar estimating procedure for deriving the volume, prices, gross value of production, consumption, and exports of cocoa, coffee, castor beans, corn, and six minor selected crops with sizeable export shares (see Tables A-13, A-1h, and A-16). In each case, the volume of exports is subtracted from the domestic production to estimate implicit domestic consumption. The latter is multiplied by average domestic prices to obtain the total value of domestic consumption. The addition of the value of exports and the value of domestic consumption :-ives the value of total production. The remaining 8 major and 53 minor crops are valued at domestic wholesale prices (see Tables A-ll and A-15). Finally, the CBE aggregates the gross value of production for all 13 major and 59 minor crops (see Table A-ll). 35. Livestock and kindred products, which account for about a fifth of the total value of agricultural production in Ecuador, comprise meat, milk, eggs, and wool. The estimates of meat production are bui.lt up from the per capita consumption of beef, lamb, pork, and poultry 1/ The CBE tries to review the basic data cautiously and judiciously, rejecting or adjusting inadeouate statistical information. Never- theless, the overwhelming inadequacies and continuous revisions of basic statistics make it very difficult to avoid occasional slipups. Thus, estimating the 1969 GDP at factor cost originating in agriculture, the CBE used a preliminary estimate of 5,833,562 metric tons of bananas reported for 1969 by the Ministry of Product- ion and arrived at an estimate of 8,874 million sucres for the agricultural GDP in 1969. Subsemiently, the Ministry of ProductVon published a revised estimate of 5,388,099 metric tons which it later revised further down to 3,870,h99 metric tons of bananas for 1969 (see errata to the Ministry of Production, General Bureau of Planning, Department of Statistics, Estimation of Harvested Area and Agricultural Production of Ecuadr ear 9 no publica- tion date . Usinw the latest figure shown in the errata, the GDF in agriculture is reduced from 8,87h to 8,h62 million current sucres and its real growth drops from 8.7 to 4.0 percent in 1969. The CBE will probably lower its estimate in the next round of revisions.  Table A-13: COCOA--VOLUME, PRICES, ANID VALUE OF PRODUCTION, CONSUMPTION, AND EXPORTS, ECUADOR, ESTIMATING PROCEDURE FOR 1969-1972 Items Jan.-July Aug.-Dec. Iterns 1969 1970 1970 1970 1971 1972 Volume in Metric Tons Domestic production 47,993 *t 53,584 65,900 60,300 Less: exports 32,649 23,290 13,396 36,686 50,900 45,000 Domestic consumption 15,344 to to 14,606 15,000 15,300 Domestic and Export Prices Domestic price per 46 kg. quintal (sucres) 534.7 .. .. 467.1 390.0 400.0 Domestic price per metric ton (sucres) 11,625 .. .. 10,154 8,478 8,696 Export price per metric ton (US$) 750.10 618.94 586.37 607.04 499o00 517.00 Values of Exports and Production Exports (thousands of US$) 24,490 14,415 7,855 22,270 25,376 23,300 Exchange rate (sucres per US$) 17.82 17.82 24.75 20.26 25.00 25.00 Exports (millions of sucres) 436.4 256.9 19)4.4 451.3 634.4 582.5 Domestic consumption (millions of sucres) 178.4 .0 0- 148.3 127.2 133.0 Total production (millions of sucres) 614.8 *- * 599.6 761.6 715 5 . | \JJ> Sources: Central Bank of Ecuador, National Income Division; Ministry of Production, General Bureau of Planning, Division of Statistics; Mission estimates.  Table A-li : COFFEE--VOLUME, PRICES, AND VALUE OF PRODUCTION, CONSUMPTION, AND EXPORTS, ECUADOR, E9TTM-ATING PROCE1DUTRE FOR 1969-1972 Jan.-July Aug.-Dec. Items 1969 1970 1970 1970 1971 1972 Volume in Metric Tons Domestic production 55,893 .. .. 60,427 64,400 68,600 Less: exports 38,161 16,911 35,664 52,575 46,400 49,600 Domestic consumption- 17,732 . *. 7,852 18,000 19,000 Domestic and Export Prices Domestic price per pound (sucres) 4.3 *- 5.6 4.1 4.4 Domestic price per metric ton (sucres) 9,505.6 .. 12,362.0 9,050.8 9,713.0 Export price per metric ton (US$) 696.34 *- *- 960.84 786.64 837.00 Value of Exports and Production Exports (thousands of US$) 26,639 15,599 34,917 50,516 36,494 41,500 Exchange rate (sucres per US$) 17.82 17.82 24.75 22.61 25.00 25.00 Exports (millions of sucres) 474.7 278.0 864.2 1,142.2 912.4 1,037.5 Domestic consumption (millions of sucres) 168.6 .. 97.1 162.9 184.5 Total production (millions of sucres) 643.3 1,239.3 1,075.3 1,222.0 Ong Sources: Central Bank of Ecuador, National Income Division; Ministry of Production, General 3ureau of Planning, Division of Statistics; Mission estimates.  ANNEX A Page 25 Table A-15: MINOR AGRICULTURAL CROPS--VOLUME, PRICES, AND VALUE OF PRODUCTION, ECUADOR, ESTIMATING PROCEDURE FOR 1969-1972 Domestic Prices in Sucres Value in Metric per 46 Kg. per Metric Sucres # Minor Crops Tons Quintal Ton (000) 1. Abaca 3,500 6/ b/ 18,700 2. Anise 276 900.00 19,5 6.00 5,400 3. Annato 1,150 b b/ 8,776 4. Apples 4,546 281.00 6,108.94 27,771 5. Apricots 280 250.00 5,435.00 1,522 6. Avocados 23,768 251.00 5,456.74 129,696 7. Bananas (oritos) 8,015 10.00 217.40 1,742 8. Beans (lima) 10,963 161.62 3,513.62 38,520 9. Beets 3,142 85.70 1,863.12 5,854 10. Cabbage 92,906 72.00 1,565.28 145,424 11. Carrots (white) 2,760 84.00 1,826.16 5,040 12. Carrots (yellow) 3,477 78.60 1,708.76 5,941 13. Cassava 390,177 44.20 960.91 374,925 14. Cauliflower 4,745 212.00 4,608.88 21,867 15. Cherimoya 7,700 183.32 3,985.38 30,687 16. Chili 2,009 163.30 3,550.14 7,131 17. Citron 21,015 21.00 456.54 9,594 18. Coconuts 42,785 64.00 1,391.36 59,529 19. Flax 20 700.00 15,218.00 304 20. Garlic 3,830 404.00 8,782.96 33,639 21. Goosefoot 828 72.00 1,695.72 1,404 22. Grapes 1,285 1,040.00 22,609.60 29,053 23. Grapefruit 10,219 50.21 1,091.56 11,154 24. Lemons 15,324 181.20 3,939.29 60,365 25. Lentils 1,1142 536.61 11,665.90 13,322 26. Lettuce 10,982 61.00 1,326.14 14,563 27. Lupines 1,461 159.92aL 3,476.66 5,079 28. Mamney 2,189 135.00 2,935.00 6,424 29. Mandarins 11,058 248.00 5,391.52 59,619 30. Mangoes 12,007 30.00 652.20 7,831 31. Melons 715 b/ b/ 750 32. Oats 221 208.00 4,521.92 999 33. Oranges 154,866 61.00 1,326.14 205,374 34. Oranges (small green) 25,150 144.00 3,130.56 78,733 35. Papaya 29,738 76.40 1,660.94 49,392 See footnotes at end of table.  ANNEX A Page 26 Table A-15 -- Continued Domestic Prices in Sucres Value in Metric per 46 Kg. per Metric Sucres # Minor Crops Tons Quintal Ton (000) 36. Palm (African) 10,050 116.70 2,537.06 25,h97 37. Peaches 2,489 183.50 3,989.29 9,929 38. Peas (green) 11,103 179.30 3,897.98 h3,279 39. Peanuts 5,858 345.00 7,500.30 43,937 40. Pears 2,934 131.50 2,858.81 8,388 41. Pineapples 57,292 b/ b/ 107,014 42. Plums 1,840 280.00 6,077.20 11,200 43. Potatoes (sweet) 11,311 61.10 1,328.31 1,,024 L4. Pyrethrum 1,586 b/ b/ 9,548 45. Ranunculus 12,079 61.00 1,391.36 16,806 46. Rye 2,622 135.00 2,93h.90 7,695 47. Sapota 7,3L1 50.00 1,087.00 7,980 L8. Sesame 1,675 291.h1 6,335.25 10D,611 49. Sisal 21,839 253.00 5,500.22 120,116 50. Sour grass 7,820 65.00 1,413.10 1L,050 51. Soyac/ 472 90.00 1,956.60 943 52. Soya_/ 552 90.00 1,956.60 L,080 53. Squash 19,269 31.00 673-94 12,986 5h. Sweet lime 603 120.00 2,608.80 1,572 55. Tea 273 180.00 3,913.19 L,068 56. Tobacco 1,494 b/ b/ 19,190 57. Tomatoes 25,194 192.00 4,17h.08 10,,162 58. Watermelons 5,481 164.82 3,583.19 19,639 59. Other 15,407 .. .. 77,677 Total for 59 minor crops 1,130,831 2 . _ ,163,515 a/ The 1968 price for lupines (item 27) of 1h5.h6 sucres per quintal was raised by 9.94 percent which corresponds to the average increase in food prizes for the city of Quito. b/ See tables below. c/ Soya beans appear twice with the same prices and different quantities, Sources: Central Bank of Ecuador, National Income Division; Ministry of Production, General Bureau of Planning, Division of Statistics.  Table A-1 6: SELECTED AGRICULTURAL CROPS--VOLUME, PRICES, AND VALUE OF PRODUCTION, CONSUMPTION, AND EXPORTS, ESTIMATING PROCEDURE FOR 1969-1972 Exports Dom stic Consumr tion Production Price per U.S. Price per Metric M. Ton Dollars Exchange Sucres Metric M. Ton Sucres Metric Sucres Tons (us$) (00O) Rate (OOO) Tons (Sucres) (000) Tons (000) Selected Major Crops Castor beans 16,159 111 1,793 17.82 31,951 7,614 1,739 13,242 23,773 45,194 Corn 200 55 11 17.82 196 222,286 2,401 533,653 222,486 533,849 Selected Minor Crops Abaca 456 265 121 17.82 2,156 3,044 5,435 16,544 3,500 18,700 Annato 161 211 34 17.82 606 989 8,261 8,170 1,150 8,776 Melons 242 79 19 17.82 339 473 870 411 715 750 Pineapples 1,681 169 284 17.82 5,o60 55,611 1,833 101,953 57,292 107,014 Pyrethrum 143 699 100 17.82 1,782 1,443 5,382 7,766 1,586 9,548 Tobacco 18 3,556 64 17.82 1,140 1,476 12,229 18,049 1,494 19,190 Source: Central Bank of Ecuador, National Income Division. F.  ANNEX A Page 2 projected for 1969-1973. Meat prices also reflect transportation costs and trade margins. The per capita consumption multiplied by the mid-year population of Ecuador yield total consumption which i- assumed to be roughly equal to the volume of production. Multiplied by corresponding prices for each of the four meat types, the CBE obtains the gross value of meat production (see Table A-17). 36. The production of milk is estimated from the projected number of milking cows multiplied by the projected average annual milk production per cow. The former projection assumes an accelerating rate (from 2.8 percent in 1970 to 3.7 percent in 1972) and the latter is based on a decelarating rate (3.7 percent in 1969 and 2.6 percent in 1972), the production growing at about a constant rate of 6.2-6.h percent per year. Estimating the total human consumption of milk, the CBE adjusts these production data (provided by the Ministry of Production) for the milk consumption by calves and for small losses and additives. The total milk consumption is then multiplied by the average consumer price per liter of milk to estimate the gross value of milk production. The gross value of egg production is similarly estimated from the projected per capita annual consumption of eggs multiplied by the total projected population and by the consumer prices for eggs. The gross value of wool production is based on the projected sheep herds, the averaee annual production of wool per head projected by the Ministry of Production and adjusted upward according to the estimates prepared by the National Association of Sheep Growers (ANCO) and the consumer prices for wool (see Table A-18). 37. The above national accounts estimates of livestock and livestock products are projections which are based either on constant or on accelerating growth rates. The use of wholesale and retail instead of farm-gate prices tends to raise the level of the estimated production. Although the higher level of production is to some extent corrected by regressions, the projections nevertheless impart a continuous growth to the estimated livestock production, as the estimates for meat are aggregated with those of the other livestock products (see Table A-19). 38. The estimates for hunting, forestry, and fishing are very crude extrapolations based either on the assumption that the output failed to increase at all (e.g. hunting in 1966, 1967, and 1969 and forestry in 1969) or that the increase reflected upward price changes, or some real grou-th, as e.g. the fishing in 1967 and in 1971. Inasmuch as the forestry and fishing account for less than 10 percent of the total gross value of the agricultural sector, the limitations of these estimates add relatively little to the range of uncertainty which pervades the estimated total agricultural production (see Table A-20). 1/ Dr. Nelson Jaramillo, Projections of Livestock Popu ation,_1969-1973 (Quito: Ministry of Production, Department of Livestock, no publication date).  ANNEX A Page 29 Table A-17: GTROSS VALUE OF MEAT PRODUCTION, EOUAD1OR ESTIMATING PROCEDURE FOR 1969-1972 1969 1970 1971 1972 Consumption per capita: 1 Beef (kgs.) 7.11 7.21 7.38 7.58 Lamb (kgs.) 1.12 1.16 1.24 1.34 Pork (kgs.) 4.06 4.38 4.82 5.33 Fowl (kgs.) 1.50 1.60 1.70 1.80 Mid-year population ('000) 5,889.5 6,092.c 6,297. 6,508.3 Total consumption - production Beef (m. tons) 41,875 43,930 46,473 49,333 Lamb (m. tons) 6,596 7,068 7,809 8,721 Pork (m. tons) 23,911 26,687 30,353 34,689 Fowl (m. tons) 8,834 9,749 10,705 11,715 Total production (m. tons) 81,216 87,434 95,34o 10o4,458 Prices (sucres per kg.): Beef 15.79 16.53 18.53 'O.40 Lamb 13.00 13.22 14.82 16.30 Pork 14.59 14.99 16.80 18.50 Fowl 23.81 24.25 27.43 30.20 Gross value of production: Beef 661.2 726.2 861.1 1,006 4 Lamb 85.8 93.4 115.7 142.2 Pork 348.9 400.0 509.9 641.7 Fowl 210.3 236.4 293.6 353.8 Total (mil. sucres) 1,306.2 1,456.0 1,780.3 2,144.2 Sources: Ministry of Production, Department of Livestock; Central Bank of Ecuador, National Income Division; National Institute of Statistics, Division of Price and Labor Statistics; Mission estimates.  ANNEX A Page 30 Table A-18: PRODUCTION OF MILK, EGGS, AND WOOL, ECUADOR, ESTIMATING PROCEDURE FOR 1969-1972 Indicators 1969 1970 1971 1972 Raw Milk Production Milking cows (thousands of heads) 437.3 453.6 46L.7 476.8 Annual production per cow (liters) 1,280 1,310 1,360 1iblo Total gross production (mil. lit.) 559.7 59h.2 632.0 672.2 Less: consumed by calves (mil. lit.) 2.8 3.0 3.2 3.4 Unadjusted consumption (mil. lit.) 556.9 591.2 628.8 668.8 Losses and additives (mil. lit.) 0.2 -0.2 0.2 0.2 Adjusted consumption (mil. lit.) 557.1 591.0 630.0 6,70.0 Price (sucres per liter) 1.76 2.0 2.2 2.4 Total gross value (mil. sucres) 980.6 1,182.0 1,386.0 1,608.0 Egg Production Consumption per capita (Kg.) 4.0 .5 .8 5.0 Mid-year population (000) 5,889.5 6,092.9 6,297.2 6,508.3 Total production (m. tons) 23,558 27,418 30,227 32,542 Price (sucres per kg.) 1h.70 14.8 18.26 21.0 Total gross value (mil. sucres) 3h6.4 406.8 551.9 653.3 Wool Production Sheep herds (thousands of heads) | 2,230.9 2,307.6 2,388.8 2,h74.1 Wool production per head (grams) t439 464 507 5L6 Total wool productiona/ (m. "/ons) 980 1,070.0 1,210.0 L,350.0 Adjusted for undervaluationb/ (n. tons) 1,990.6 2,060.0 2,460.0 ',750.0 Price (sucres per kg.) 16.3h 19.LO 21.73 24.34 Total gross value (mil. sucres) 32.5 40.0 53-5 66.9 a/ Ministry of Production, Department of Livestock. b/ The National Association of Sheep Growers (ANCO) adjusted for the undervalua- tion in 1969, the Central Bank of Ecuador made the adjustment for 1970, and the Mission adjusted for 1971 and 1972. Sources: Ministry of Production, Department of Livestock; Central Bank of Ecuador, National Income Division; National Institute of Statistics, Division of Price aid Labor Statistics; Mission estimates.  ANNEX A Page 31 Table A-19: OUTPUT OF LIVESTOCK AND KINDRED PRODUCTS, ECUADOR, SUMMARY OF ESTIMATES FOR 1969-1972 (fross lialue of Production in Millions of current .ucres) Livestock and Kindred Products 1969 1970 1971 1972 Beef 661.2 726.2 861.1 1,006.4 Lamb 85.8 93.4 115.7 142.2 Pork 348.9 400.0 509.9 641.7 Fowl 210.3 236.4 293.6 353.8 Milk 980.6 1,182.0 1,386.0 1,608.0 Eggs 346.4 406.8 551.9 683.3 Wool 32.5 40.0 53.5 66.9 TOTAL 2,665.7 3,084.8 3,771.7 4,502.3 Source: Tables 17 and 18 above.  ANTEX A Page 32 Table A-20: GROSS VALUFE OF AGRICULTURAL PRODUCTTON, ECUADOR, SUMMARY OF ESTIMATES, 1960-1972 (MIllions of Current Sucres) 13 Major Anrual and 59 Minor Hunt- Growth Year Crops Livestock ing Forestry Fishimn Total Rate 1960 5,867.6 1,303.6 2.4 432.7 168.4 7,77h.7 1961 6,307.2 1,397.6 2.3 462.9 189.8 8,359.8 7.5 1962 7,3h2.3 1,507.4 2.3 481.6 175.8 9,509.4 19.8 1963 7,673.3 1,660.5 2.4 510.3 191.L 10,037.9 E.6 1964 7,980.9 1,656.4 2.5 551.0 176.1 10,366.9 .-3 1965 8,970.1 1,801.3 2.6 600.2 230.7 11,60h.9 -.9 1966 9,581.6 1,973.8 2.6 653.7 228.8 12,L40.5 7.2 1967 9,943.6 2,113.6 2.6 711.9 320.9 13,092.6 r 2 1968 9,661.9 2,178.6 3.0 861.3 325.7 13,030.5 -() 5 1969 11,065.6 2,665.7 3.0 861.3 330.1 14,925.7 1L45 1970 12,668.8 3,084.8 3.2 930.0 359.1 17,0o5.9 1 2 1971 13,778.0 3,771.7 3.3 1,000.0 575.0 19,128.0 1972 1h,973.8 4,502.3 3.4 1,100.0 660.0 21,239.5 1>0 Sources: Central Bank of Ecuador, National Income Division, Mission estimites.  ANNEX A Paage 33 39. Agricultural nonfactor inputs are subtracted from the gross value of agricultural production as a whole, that is, from the total value of crops, livestock products, hunting, forestry, and fishing. Such a global adjustment can only be very crude. For the 1950's, the CBE adopted average Latin American ratios for agricultural non- factor inputs, as reported at a national accounts conference in Rio de Janeiro in the miid-1950's. For the 1960's, the CBE replaced the average Latin American estimates by Ecuadorian nonfactor input ratios, as estimated by the NEPCB in 1963 (see Table A-21). The NEPCB esti- mates for 1963 were extrapolated backwards to 1960 and forwards to 1968 for all the relevant major agricultural expenditure categories. First, the CBE made the adjustment for the higher retail and wholesale prices by subtracting the transport and handling cost and the wholesale margin from the gross value of production to arrive at the farm-gate value of production. Next, it subtracted the total intermediate consumption, including seeds, fertilizer, pesticides, fodder, vaccine and medicine, fuel and lubricants, packing and binding, maintenance, and interest and commissions. Since the detailed components were estimated only for 1963, the CBE extrapolated their combined value for all the other years of the 1960's. Subtracting it from the value of production at farm-gate, the CBE obtained the GDP at market prices (or gross value added) in agriculture. h0. For the 1950's and most of the 1960's, the CBE estimated separately for agriculture net factor payments to abroad, indirect taxes, and subsidies in order to arrive at the GNP and GDP at factor cost originating in agriculture (see Table A-21). These estimates were then integrated in the national accounts as a GDP at current factor cost originating in agriculture, hunting, forestry, and fishing (cf. 4,731 million sucres for 1960 in Tables A-21 and A-22). Starting in 1969, however, the above procedure of subtracting the extrapolated 1963 values of nonfactor inputs and other components was replaced with an alternative procedure based on the regression of the value added or GDP on the gross value of production in agriculture, hunting, forestry, and fishing (see Table A-22). The latter method is not necessarily inferior to the former inasmuch as hardly any basic data are available for estimating the various components of nonfactor inputs. 1/ 1/ Although the regression method may be considered superior on analytical grounds, it is probably inferior to the ratio method in practical work because it is cumbersome and mechanical. It obscures the crude nature of basic estimates and makes it more difficult to introduce fragmentary information into the estimates. In the absence of computer programs, arithmetic errors tend to creep into the calculations of regressions. Notwithstanding the erudition of the CBE professional staff, the CBE has obtained different estimating equations than the results produced by an electronic computer for the same data and regression method. Thus, the CBE estimated agriculture GDP at 10,271 million current sucres in 1970 although the regression equation which it used yields only 9,652 million sucres for the same year. Moreover, the same basic data fed into an electronic computer produced a somewhat different estimating equation which yielded a still different estimate of 9,693 million sucres for 1970. The correction of these apparent arithmetic errors reduces the 1970 agricultural GDP by 6 percent. A ratio method would probably have avoided the aritbhoetic errors. Moreover, requiring less laborious computations, the ratio method can be used with greater ease and speed for revisions at more frequent intervals, as new basic data become available.  ANNEX A Page :3 -aole A-21 : GDP AT CURRENT FACTOR COST ORIGINATING IN AGRICULTURE, HUNTING, FORESTRY, AND FISHING, ECUADOR, ESTIMATING PROCEDURE FOR 1960 - 1965 (Millions of Current Sucres) Major Output and Input 1960 1961 1962 1963 196L 1965 Expenditure Categories Gross value of production 7,660 8,381 9,256 9,794 10,00? 10,702 Less: transport and handling cost 980 1,098 1,222 1,273 1,251 1,370 Less: income of wholesalers 844 965 1,055 1,117 1,10(0 1,220 Value of production at farm gate 5,836 6,318 6,979 7,404 7,652 8,112 Less: intermediate consumption 828 913 1,009 1,076 1,061 1,178 (a) Seeds 254 h. (b) Fertilizer 72 .. (c) Pesticides . 72 (d) Fodder 281 (e) Vaccine and medicine 10 . .. (f) Fuel and lubricants .. .. 33 .. .. (g) Packing and binding .. 115 (h) Maintenance .. .. 132 .. .. (i) Interest and I commissions 107 GDP at m.p.(value added) 5,008 5,405 5,970 6,328 6,59]. 6,934 Less: net factor payments to abroad -30 -28 -19 -4 -3& -27 GNP at market prices 4,978 5,377 5,951 6,324 6,553 6,907 Less: indirect taxes -334 -338 -353 -411 -L91 -487 Plus: subsidies 57 60 72 43 9, 35 GNP at factor cost 4,701 5,099 5,670 5,956 6,153: 6,h55 Plus: net factor payments to abroad 30 28 19 4 38 27 GDP at factor cost 4,731 5,127 5,689 5,960 6,lsl 6,482 Source: Central Bank of Ecuador, National Income Division. The data are b:ase] on a study by Jan B. Van As and Manuel Arias B., Program of Agricultural Development, Goals and Projections, 1963-1973 (Quito: NEPCB, March 30, 196h).  ANNEX A Page 35 Table A-22: GDP AT CURRENT FACTOR COST ORIGINATING TN AGRICULTURE, FORESTRY, HUNTING, AND FISHING, ECUADOR ESTII4!ATII'TG PROCEDURE FOR CAIETNDAR YEARS 1969-1972 (Millions of Sucres at Current Factor Cost) Agriculture, Hunting, Forestry, and Fishing Year GDP Gross Value of Production 1960 4,731 7,77h.7 1961 5,127 8,359.8 1962 5,689 9,509.4 1963 5,960 10,037.9 1964 6,191 10,366.9 1965 6,482 11,60h.9 1966 7,227 12,h40.5 1967 7,556 13,092.6 1968 7,771 13,030.5 1969 8,562 14,925.7 1970 9,693 17,0h5.9 1971 10,804 19,128.0 1972 11,930 21,239.5 The above variables produced the following estimating equation for GD? at current factor cost originating in agriculture, forestry, hunting, and fishing: Yi = 602.02734 + 0.53333Xi This equation produced the following estimates: Y1969 = 8,461.8 Y1971 10,803.6 Y1970 = 9,693.1 Y1972 = 11,929.7 The CBE computed a different estimatirgequation (Yi 671.01111 + 0.52688Xi) which led to different estimates: Y1969 = 8,871 Yl970 = 9,652 (CBE estimate is 10,271) Source: Central Bank of Ecuador, National Income Division; Mission estimate.  ANNEX A Page 36 41. Reviewing the agricultural value added estimates of Ecuador for the last two decades, it must be borne in mind that there is a limit to the improvements which can be brought about by various estimating procedures. This limit may still not have been approached in Ecuador with respect to the value added estimates in the agricultural sector, although further improvements depend largely on the availability of basic data. The use of average Latin American nonfactor input ratios for the 1950's and the lower Ecuadorian ratios for the 1960's resulted in an understatement in the level of the agricultural value added for the 1950's, a break in this series, and an overstatement in the rate of growth in 1960. The extent of these inaccuracies is not; known because the CBE has never derived the two series so as to overlap in the same year. The accuracy of the Ecuadorian nonfactor input ratios for 1963 and their relevance to the 1950's and the 1970's is also open to question. However, the importance of these considerations is overshadowed by the apparent deficiencies of the basic statistics on output and prices. It is possible that the national accounts estimates could be improved by resorting to the FAO estimates, even though the latter are based on yields and harvested areas quite different from those reported by the Government of Ecuador (see the discussion of basic statistics in Annex B). It would be far better, of course, for the Government of Ecuador to improve its own statistical system which would generate meaningful agricultural data by modern sampling techniques. 42. Mining GYDP. Although expanding petroleum production is expected to make mining one of the most important industries in Ecuador, for the last two decades mining has grown little and its share amounts to only about two percent of the GDP. Mining and quarrying GDP includes the mining of metal and nonmetal minerals as well as petroleum extraction and refining. The CBE has also tried to include a rough estimate of value added for stone, clay, sand and saLt, although relatively few data are available for these activities. Petroleum extraction and refining could not have been separated in the 1950's, and to preserve the consistency with the 1950's, both these activities were also included with mining in the 1960's. For a benchmark year, the CBE estimated the value added data from the 1955 industrial census and for the 1950's and the early 1960's, from its own special surveys and inquiries. For more recent years, the General Bureau of Geology and Mining (Ministry of Natural Resources and Tourism) has provided annual data for licensed mining establishments which had reported their data. The CBE has used these data for the national accounts rather than the results of the NIS industrial surveys, especially for the period since 1965. 43. The basic data on mining and quarrying had deteriorated to such an extent in the mid-sixties that the CBE decided in 1967 to use regressions rather than to continue deriving the mining value added directly from the basic data. After some experimentation with several available production series, the CBE projected the index of  ANNEX A Page 37 Table A-23: GDP AT CONSTANT 1960 FACTOR COST ORIGINATING, TN MINING, ECUADOR, ESTIMATING PROCEDURE FOR 1965 AND 1966 Mining and Quarrying GDP Year (Millions of Sucres) Indeendent ariables at Current at Constan Sucres 1960 Sucre Y X1 X9 XX 1956 228.0 235.7 .. 143.7 15.3 52.0 115.8 1957 232.0 238.2 .. 134.0 17.1 61.3 109.7 1958 222.0 226.5 .. 130.6 20.0 84.5 120.1 1959 233.0 237.5 .. 115.9 18.6 162.6 107.2 1960 311.0 311.0 .. 115.8 15.2 126.4 107.9 1961 315.0 300.0 .. 122.9 15.2 101.2 111.1 1962 326.0 303.8 .. 108.1 20.6 128.1 124.0 1963 369.0 330.1 .. 101.4 21.0 121.8 160.8 1964 389.0 324.7 .. 117.4 16.9 117.1 137.8 1965 .. 379.1 .. 119.7 11.5 70.0 113.2 1966 .. 398.5 .. 108.8 10.9 76,7 68.9 Indexes (1956=100) 7956 100.0 100.0 100.0 100.0 100.0 100.0 100.0 1957 101.8 101.1 100.3 93.2 111.8 117.9 94.7 1958 97.4 96.i 95.4 90.2 130.7 162.5 103.7 1959 102.2 100.8 100.0 80.7 121.6 312.7 92.6 1960 136.4 131.9 130.9 80.6 99.3 243.1 93.2 1961 138.2 127.3 126.3 85.5 99.3 194.6 95.9 1962 143.0 128.9 127.9 75.2 134.6 246.3 107.1 1963 161.8 140.1 139.2 70.6 137.2 234.2 138.9 1964 170.6 137.8 136.7 81.7 110.5 225.2 119.0 1965 . 160.8 ., 83.3 75.2 134.6 97.8 1966 *- 169.1 .. 75.7 71.2 147.5 59.5 Independent Variables: Xi = Petroleum products (thousands of gallons). X2 = Gold production (thousands of troy ounces). X3 = Silver production (thousands of troy ounces), X4 = Lead production (thousands of Kilograms). Estimating equation for GD? at constant 1960 factor cost originating in mining: Y = 422.58 - 2.6415X1 - 0.8o88X2 - 0.1264X3 + 0.3532X4 Sources: OBE, National Income Division.  ANNEX A Page 383 mining GDP at constant 1960 prices by regressing it on the indexes of petroleum, gold, silver. and lead production (see Table A-23). 44. As the cornposition of the mining output changed, the CBaC has used other independent variables for these regressions. Thus, for 1969 and 1970, the mining employment, petroleum Production, production of metals and nonmetals, and the imports of mining inputs served as independent variables for estimating the mining. GDP (see Table A-24). Although such chances in the independent variables have contributed to certain inconsistencies in the estimates over time, drastic changes i-n the composition of minine may, nevertheless, have required occasional elimination or substitution of certain variables. Thus, while the total mining ouitpuit increased, petroleum production has tended to decline: n Ecuador in the late 1950's and throughout the 1960's. This negative covariation resulted in a negative coefficient of net regression for petroleum production (see Table A-211). With a substantial increase -n petroleum extraction expected to take place in 1972 and thereafter, the estimating linear equation which includes petroleum would project a drastic reduction in the total mining activity--an obviously absurd result. Therefore, the mining GDP cannot be meaningfully estimated in the future by regression on the past declines of petroleum productior-. The latter variable must now be eliminated from the set of independert variables (see Table A-2L). If. In addition to the production of crude petroleum, the COE uses four other independent variables for estimating the GDP in minirg and quarrying. The mining employment has the highest coefficient of net regression (see Table A-25). Inasmuch as the mininF employment has remained either stable or even declined in the 1960's (see Table A-2L), this coefficient is negative; therefore, it is not suitable for making estimates as the expected sharp increase in the Production of crude petroleum is not likely to be associated witb a pronounced decline iri the mining employment. Moreover, the mining employment data which the CBE has used are actually the NEPCB projections of economically active poapulation. These projections include employment as well as unemploy- ment and they are not responsive to annual changes in output. They show upward trends for the 1970's, are quite detached from the expected annual changes, and are, therefore, not suitable for making value added estimates. 46. The producti-on data on metal and nonmetal minerals which the CBE is using as independent variables accounts for a very small share of the total mining (about 3 percent for metals and 0.1 percent for non- nmetals). Moreover, both of these series fluctuate ouite differently fromn the mining as a whole. Thus, while the value added of mining showed an increase of about one third in 1960, the metals declined 20 percent  ANNEX A Page 39 Table A-24: INDEPENDENT VARIABLES USED IN REGRESSIONS FOR ESTIMATING MINING AND QUARRYING GDP, ECUADOR SELECTED YEARS 1957-1970 Variables and Indexes 1957 1960 1965 1969 1970 X1 = Mining employment (000) 4.4 4.o 4.0 4.3 4.4 Index implicit in data 110.0 100.0 100.0 107_5 110.0 Index used for GDP estimates 110.0 100.0 100.0 107.5 110.0 X2 = Petroleum production (mil. of US gal.) 134.o 115.8 119.7 65.8 6o.7 Index implicit in data 111.9 96.7 100.0 54.9 50.7 Index used for GDP estimates 111.9 96.7 100.0 55.0 50.7 X3 = Mining of metals (mil. of sucres) 10.4 11.0 13.5 19.5 22.8 Index implicit in data 76.9 81.4 100.0 144.7 170.0 Index used for GDP estimates 92.4 98.0 100.0 144.4 151.7 X4 = Mining of nonmetals (thous. of sucres) 62.0 54.o 627.0 223.0 16,981.0 Index implicit in data 9.9 8.7 100.0 35.3 2,709.3 Index used for GDP estimates 14.5 12.8 100.0 52.1 54.7 X5 = Imports of mining inputs (mil. of JS $) 12.7 13.6 16.7 28.8 28.0 Index implicit in data 76.0 81.4 100.0 172.4 167.6 Index used for GDP estimates 76.2 81.L4 100.0 172.2 167.2 Sources: Xi = NEPC0I Division of Human Resources (projections of economically active population). X2 = Bulletin of the Central Bank of Ecuador, XLV, Nos. 531, 532, and 533 (Oct. - Dec&., 1971), 271, and earlier issues. X3 = Ministry of Industry and Commerce (after 1970, Ministry of lIatural Resources and Tourism), General Bureau of Geology and Ihlines, Mining Statistics (annual issues for 1932-66, 1967, 1968, 1969 and 1970). X4 = Ibid. X5= Bulletin of the Central Bank of Ecuador, XLV, Nos. 531, 532, and 533 (Oct.-Dec., 1971), 209-213, and earlier issues (line 5c--raw materials and intermediate products for manufacturing- derived from foreign mineral products).  ANNEX A Page 40 Table A-25: GDP AT CURRENT FACTOR COST ORIGINATING IN MINING AND QUARRYING, ECUADOR, ESTIMATING PROCEDURE FOR 1969 ANID 1970 Mining and Quarrying GDP (Millions of Sucres) Independent Variables Year at at (Indexes, 1965 = 100) Current 1965 Index Prices Prices 196"=1O0 _ . ~~~~~Y Xi X2 - _Xn 4-- 1957 232 285 72.7 110.0 111.9 92.)4 1).5 76.2 1958 222 271 69.1 107.5 109.1 110.1 61.6 77.7 1959 233 284 72.5 105.0 96.8 120.b 115.5 63.7 1960 311 372 94.9 100.0 96.7 98.0 12.8 81.4 1961 315 360 91.8 95.0 102,7 94.5 49.4 84.8 1962 326 364 92.9 90.0 90.3 135.4 8.6 76.0 1963 369 396 101.1 92.5 84.7 157.3 102.6 97.4 1964 389 404 103.1 95.0 98.1 140.8 148.2 117.8 1965 392 392 100.0 100.0 100.0 100.0 100.0 100.0 1966 453 431 110.0 102.5 90.9 107.2 91.3 101.)4 1967 500 459 117.1 105.0 77.1 109.2 73.0 139.3 1968 552 488 124.5 105.0 61.8 148.4 104.6 176.8 1969 588 495 126.3792 107.5 55.0 144.4 52.1 172.2 1970 604 477 121.7095 110.0 50.7 151.7 54.7 167.2 1971 678 489 124.7757 112.5 45.3 155.0 45.0 180.0 1972 238 156 39.7962 117.5 300.0 160.0 50.0 200.0 Independent Variables: Xl = 1ining employment X2 = Petroleum production X3 = Production of metal minerals X4 = Production of nonmetal minerals X5 = Imports of mining inputs The above variables produced the following estimating equation for GDP a-, constant 1965 factor cost originating in mining and ouarrying: Yi = 286.0759 - 1.2531 X1 - 0.6425 X2 - 0.3567 X3 + 0.0504 X4 + 0.3L434 x5 This equation produced the following estimates at constant 1965 prices: Y1969 = 495 Y1970 = 477 Y1971 = 489 Y1972 = 156 The CBE computed a slightly different estimating equation (Yi = 285.2)460 - -1.2503 X1 - 0.6393 X2 - 0.3548 X3 + 0.0482 X4 + 0.3450 X5) Source: Central Bank of Ecuador, National Income Division; Mission estimates.  ANNEX A Page 41 and the nonmetals went down 90 percent. The coefficient of net regression for metals is negative and that for nonmetals approaches zero (see Table A-25). Moreover, the CBE purports to use the data reported by the G3neral Bureau of GeologD and Mines (GBGM). However, the latter agency publishes data which yield different indexes than those used by the CBE (see Table A-2h). The latter show less growth. Although the CRB has not given any explanation for these differences, the data may have been deflated or adjusted to a wider coverage. Thus, while the GBRGM reported an increase of 7,600 percent in the gross value of output for nonmetals in 1970 (larqely due to the new mining of sulphur), the CBE index for nonmetals increased from 52.1 to 51.7 or only 5 percent (see Table A-2)4). 47. The only variable suitable for the estimates would have been the time series on the imports of mining inputs. The CBE has tried to use such a series with virtually no adjustments (see Table A-24). This time series refers, however, not to mining inputs but to raw materials and intermedie products for manufacturing which are derived from imported minerals.-- These import data are not relevant to mining activity in Ecuador. The data on the imports of mining inputs are not readily available. 48. An additional distortion is introduced by the regression of the dependent variable (Y) at constant prices on the independent variables which are at current prices (X3, Xh, and X5). The price changes of the latter are fed as a part of real growth into the dependent variable. These dependent variable estimates are not deflated. On the contrary, the CBE inflates them to current prices with the adjusted cost of living index (see Table A-25). This repeated inflation results in an upward bias. Where there is no growth, upward price movements in the independent variables generate upward changes which appear as real growth in the dependent variable shown at constant prices.2/ It would have been more meaningful to deflate the independent variables before correlating them with the value added at constant prices (or alternatively, inflating all variables to current prices--an economically less meaningful procedure because e.g. employment at current prices lacks a meaningful interpretation although mathematically the results may not differ significantly). 1/ Cf. Bulletin of the Central Bank of Ecuador, XLV, Nos. 531, 532, and 533 (Oct.-Dec., 1971), 209-213, line 5c. 2/ An earlier IBRD report tried to correct for these deficiencies, assuming that the CBE series for manufacturing and construction value added at current prices related in fact to constant prices. Therefore, it inflated the series at current prices, thus superimposing a third inflationary bias on the others which the CBE had built in its original estimates. Cf. Current Economic Position and Prospects of Ecuador, Report No. WH-208a (Washington, D.C.: IBRD, September 22, 1971), Table 2.h.  AIiNEX A Page 2 49. A further distortion is introduced by expressing all variables as indexes for the purpose of regression. This procedure attempts to get around the problem of having different and arbitrary units of measuremaent for the various indePendent variables in a multiple regression, but it introduces a new problem of assigning all variable3 the same weights. This procedure does not solve the original problem and is almost just as arbitrary. The estimating equation measures the unweighted average relationship. 50. The use of multiple correlation with five independent variables gives a misleading impression of a sophisticated analysis which supposedly underlies the CBE estimates. The multiple regression technique obscurpq the various deficiencies. The elaborate mathematical computations of these regressions which the CBF carries out on a desk calculator, contain occasional errors which make the results even less meaningful. 51. Having computed the index values with the estimating equation and having converted them to value added in constant and current sucras, the CBE reviews the result in terms of level and annual changes. At this final stage, the CBS may adjust the regression results according to alternative data and various judgements, thus in effect ignorinq the regression and substituting other estimates which appear to be more plausible. Although the regression based on the CBE data and inflated with the reciprocal of the GDP deflator yielded 588 and 604 million current sucres for 1969 and 1970, respectively, the CBE calculated with its estimating equation 596 and 621 million current sucres for the sane years. Having reviewed the latter estimate in the light of partial reports on the increase of mining output in 1970, the CBE raised its estimate to 641 million sucres. It has made such last-minute adjustments also for other years and industries (see Table A-36). In view of these adjustments, the regression apPears to play an auxiliary role which yields to estimates based on additional information and on subjective judgements. 52. Manufacturing GDP. The share of manufacturing in the total GDP increased from about 15 percent in 1960 to about 7 percent in 197:. Manufacturing includes large, medium, and small-scale establishments as well as cottage (household) manufacturing industry. Depending on the definition, the share of the cottage and small-scale industry has been estimated from a high of 61.5 percent in 1960 (for establishments witn less than 300,000 sucres annual production)!/ to a low of about one third of the total value added by manufacturing (for establishments with less 1/ Central Bank of Ecuador, Sources of Tnformation and Procedures Employed in the Estimation of National Income of Ecuador (Quito: CBE 196h) p. 12.  ANNEX A Page 43 than 5 employees and 120,000 sucres annual production). For the 1950's, the CBE estimated that the share of small-scale and household industry increased from 57.7 percent in the middle 1950's to 61.5 percent in 1960,21 and for subsequent years, the CBE has generally maintained the value added by the small-scale industry at about the same level, allowing its relative importance to decline below 40 percent, as the medium and large-scale industry continued to grow at an apparently accelerating pace in the 1960's. These general working hypotheses underlie the GDP estimates for manufacturing. 53. Similarly as with the mining estimates, the CBE used the 1955 Industrial Census and its subsequent annual surveys for estimating the value added by large-scale manufacturing, making an adjustment for small-scale and household manufacturing industry. With a gradual deterioration of the basic data and a reduction in the staff of the NID, the CBE decided in 1967 to use regressions of manufacturing value added on readily available indicators. Having experimented with several related indicators, the CRE selected the most suitable and has used them for making the estimates since 1965. 94. Preparing the GDP estimates of manufacturing for 1969 and 1970 (the latest available), the CBE computed an index of the deflated value added with a 1965 base (Y) and regressed it on four independent variables (see Table A-27). Most of the limitations discussed above with respect to mining apply also to manufacturing. Moreover, the independent variables on which the CBE regresses the manufacturing GDP relate to inputs and not to manufacturing output. Two of these independent variables relate to nonfactor inputs. Nevertheless, the sales of electric energy to industry (Xl) has a fairly meaningful positive coefficient of net regression. Although the electrification of manufacturing is not directly related to the growth in its value added, their co-variation is not necessarily spurious. The greater use of electric power may go hand in hand with a greater capital intensity and higher capital consumption allowances, higher paid labor, multiple shifts, and a general expansion of manufacturing. Imports of industrial raw materials represent another nonfactor input which is, however, less closely associated with the growth in manufacturing factor inputs. According to our calculations based on the CBE data, this independent variable shows a slight negative coefficient of net regression which may be partly explained by the deliberate Ecuadorian policy aimed at import substitution. 2/ Although the relationship is l/ Ibid., p. 12. 2/ The OBE estimating equation shows a higher and a positive coefficient of net regression for the imports of industrial raw materials. However, the CGER estimating equation for manufacturing appears to suffer from a deficient arithmetic.  ANNEX A Page 44 Table A-26: INDEPEND'T)NT VARIABLES USED IN REGRESSIONS FOR ESTIMATING MANUFACTURING GDP, ECUADOR, SELECTED YEARS 1957-1970 Variables and Indexes 1957 1960 1965 1969 1970 Xl = Sales of electric energy to industry (mil. of kwhs.) *. 126.7 188.9 283.5 320.8 Index implicit in data .. 67.1 100.0 150.0 159.8 Index used for GD? estimates 51.7 6)4.6 100.0 144jL.5 1560.7 X2 = Imports of industrial raw materials (mil. of US 4$) 40.03 )41.92 57.78 87.06 32.25 Index implicit in data 69.3 72.5 100.0 150.6 1 42.3 Index used for GDP estimates 69.3 72.6 100.0 150.7 1 42.3 X3 = Inports of industrial capital goods (mil. of US $) 13.59 17.43 29.67 44.19 '49.03 Index implicit in lagged data 47.8 54.3 100.0 152.8 Lagged index used in GDP estimates 44.2 46.4 80.3 142.3 149.0 Xt, =Manufacturing employment (000) Employment, large-scale 37.9 43.9 59.9 79.1 31.0 Employment, total 177.3 193.3 227.0 241.8 250.0 Economically active population 185.5 201.8 229.9 253.8 260.1 Indexes iaplicit in data: Employment, large-scale 63.2 73.2 100.0 132.0 135.2 Employment, total 78.1 85.2 100.0 106.5 110.1 Economically active population 80.7 87.8 100.0 110.)4 113.1 Index used in GDP estimate 82.8 89.9 100.0 11)4.1 116.9 Sources: Xl = Unpublished data provided by the Statistical Section of the Euadorian Insti-tute of Electrification (INECEL) for 1964-1970. The source of the- 1960 data could not be ascertained. X2 = Bulletin of the Central Bank of Ecuador, XLV, Nos. 531, 532, and 53'i, ect.-Dec., 1971), 209-213. X3 = Ibid., 209-213. Xl; = National Economic Planning and Coordination Board, Division of Humar. Resources (unpublished estimates dated September, 1970).  ANNEX A Page 45 Table A-?7: GTP AT CURRENT FACTOR COST ORIGINATING TN MVANIUFACTURTNG, ECUADOR, ESTIMATING PROCEDURE FOR 1969-1972 Manufacturing GDP (Millions of Sucres) Independent Variables at at Index (Indexes, 1965-100) Year Current 1965 1965=100_ Prices Prices X " Y Xi X Lagged Lagged X4 _________ ________ ________ Data Index Index 1957 1,625 1,999 60.6 51.7 69.3 47.8 44.2 82.8 1958 1,739 2,126 64.3 54.4 66.8 42.9 45.8 85.1 1959 1,830 2,234 67.7 55.9 63.5 54.5 51.7 87.5 1960 2,011 2,408 73.0 64.6 72.6 54.3 46.4 89.9 1961 2,112 2,411 73.1 74.6 73.9 44.1 58.8 92.4 1962 2,283 2,548 77.2 78.4 74.0 52.9 58.7 96.0 1963 2,523 2,704 82.0 82.3 86.1 74.3 47.6 97.0 1964 3,039 3,152 95.5 91.0 110.0 92.5 57.3 99.0 1965 3,299 3,299 100.0 100.0 100.0 100.0 80.3 100.0 1966 3,501 3,331 101.0 103.8 102.2 118.7 100.0 101.3 1967 3,884 3,567 103.1 113.5 131.0 131.6 108.1 102.6 1968 4,209 3,721 112.8 130.0 143.3 137.8 128.4 111.3 1969 4,917 4,142 125.5531 144.5 150.7 152.8 142.3 114.1 1970 5,565 4,396 133.2608 160.7 142.3 160.0 149.0 116.9 1971 6,450 4,650 140.9516 175.0 149.0 166.4 155.0 122.4 1972 7,579 4,963 150.4)451 190.0 158.0 179.3 167.0 128.3 Independent Variables X1 = Electric energy consumption by manufacturing in kwhs. X2 = Raw material imports for manufacturing. X = Imports of capital goods for manufacturing. X4 = Manufacturing employment. The use of an index implicit in the lagged data produced the following estimating equation for GDP at constant 1965 factor cost originating in manufacturing: Yi = 6.6975 + 0.2959 X1 - 0.0199 X2 + 0.2287 X3 + 0.3871 X4. This equation produced the following estimates: Y1969 = 4,142 Y1 970 = 4,396 'f1971 = 4,650 Y1972 4,963 The CBE, using lagged index, computed a different estimating equation (Yi = 69.6422 + 0.9422 X1 + 0.0977 X2 + 0.1091 X3 - 0.6807 X4) which produced higher estimates. Sources: Central Bank of Ecuador, National Income Division; Mission estimates.  ANNEX A Page 46 weak statistically, it is quite significant economically inasmuch as the value of imported manufacturing inputs amounts to about 40 percent of the manufacturing value added. Therefore, this independent variable is usually considered to be suitable for estimating the manufacturing G.77. 55. The value of imported capital goods for manufacturing amounts to about 25 percent of the GDP originating in manufacturing. The cof3fficient of net regression is positive (see Table A-27). Its absolute value is 11 times higher than that for imported raw materials. Moreover, the imported capital goods contribute directly to factor inputs. Therefore, this series appears to be suitable for the regression. 56. The use of the value of imported capital goods for the regression reauires the lagging of the data by one year because the monthly data show that these inputs are usually more or less evenly distributed throughout the year. Taking the mid-year as the average arrival date for the year and allowing six months for the installation of capital goods, training of manpower, and starting of production at reasonable capacity levels, the data on imported capital goods for any given year tend to be related to the output of the subsequent year. Instead of lagging the absolute data on these imports, the CBE has lagged the index, shifting its base from 1965 to 1966 (see Table A-27). This procedure resulted in a certain inconsistency because the weight of the lagged index became relatively smaller. However, in view of the general arbitrariness introduced by the use of equal weights for all independent variables, the implicit assignment of a smaller weight for X3 is not necessarily a step in the wrong direction. 57. Manufacturing employment is most closely related to the total factor inputs in manufacturing. Its coefficient of net regression is positive and higher than that of any other independent variable (see Table A-27).1/ However, the index used in GDP estimates differs signi- ficantly from that implicit in the manufacturing employment data whizh the CBE made available to the Mission (see Table A-26). It appears that the CBE derived its index from the total economically active population in manufacturing, including small-scale and household manufacturing. The latter has grown less rapidly, as indicated by the CBE index. 58. Having estimated the index of the dependent variable, the CBE derived the manufacturing GDP for 1969 and 1970 at 4,201 and 4,590 million constant 1965 sucres. Multiplying the latter by the general 1/ The CBE equation shows a negative coefficient of net regression for employment. This seems to be the result of a faulty arithmetic because both series show steady growth in all years and there is no apparent reason for a negative covariation.  ANNEX A Page 47 deflator with a 1965 base, the CHE obtained its first estimates of 5,062 and 5,921 million current sucres for 1969 and 1970, respectively. Before publishing them, however, the OBF, revised them downward (see Table A-36). The downward revision appeared to be necessary in view of the lower large-scale manufacturing GDP indicated by the 1969 Survey of Manufacturing (3,03h.7 million current sucres in 1969 as compared to 2,659.0 million sucres in 1968). Inasmuch as manufacturing grew (at current prices) only l4.l percent in 1969 and the CBE has assumed hardly any real growth in handicrafts and small-scale industry, the growth in manufacturing GDP which includes handicraft had to be reduced below 14.1 percent. The CBE reduced it to 13.2 percent in 1969 and made a corresponding downward adjustment for 1970. 59. Using the same independent variables and indexes as the CBE, the ivassion has obtained different regression results by means of an electronic computer. These differences are quite significant, especially for 1970 (see Table A-27), and they may partly explain the need for the CB1 revisions of the 1969 and 1970 regression results (see Table A-26). A further revision of manufacturing GDP may be necessary in 1970 to bring its growth rate in line with the results of the 1970 Manufacturing Survey and the assumed zero growth of handicrafts ard small-scale industry. 60. Construction GDEP. -- Construction accounts for about 5 percent of total GDP. It includes value added of public construction, heavy private construction, and residential construction in urban and to some extent in rural areas. The CBE1 Fiscal Studies Division provides most of the needed data for the public sector from the accounts of government agencies and public enterprises. The CBE National Income Division used to collect similar data for the private sector from the private construction companies. In addition, the CBE has used cons- truction permits, municipal cadastral surveys, and other sources for various ad hoc estimates of factor inputs in construction. Having focused itsT efforts on construction enterprises, the NID appears to have underestimated the construction by nonconstruction enterprises and private residential construction, especially in rural areas. The latter data were quite incomplete and as much as 5 years out of data even in the 1950's. There has been apparently no systematic effort to survey construction sites in Ecuador. The NID has also made little use of the 1962 Census of Housing for developing a meaningful frame for its surveys of residential construction activity. 61. With a general deterioration of basic statistics since 1965, the CBE has replaced the direct estimation of value added with multiple linear regressions of construction GDP on cement production, imports of construction materials, and fixed capital formation in the public sector (see Table A-28). Of these three independent variables, cement showed the highest positive coefficient of net regression. This  ANNEX A P age 4d Table A-28: INDEPENDENT VARIABLES USED TN REGRESSIONS FOR ESTIMATING r,ON-TR,TCTTON GDP, ECUADOR, SELECTED YEARS 1957-1970 Variables and Indexes 1957 1960 1965 1969 1970 Xl = Cement production (thous. of m. tons) 141 183 297 456 458 Index implicit in data 47.5 61.6 100.0 153.5 1T54.2 Index used in GDP estimates 52.2 61.6 100.0 153.5 1 -!.2 X2 = zLnDorts of construction materials (mil. of US $) 5.04 4.80 8.35 14.98 13.70 Index implicit in data 60.14 57.5 100.0 179.4 1i' 4.1 Index used in GDP estimates 60.4 57.5 100.0 179.4 1';4.1 X3 = Capital formation in buildings, construction, and public works (mil. of sucres) .. 799 698 1,282 Index implicit in data .. 114.4 100.0 183.6 Index used in GDP estimates 80.4 114.5 100.0 184.5 189.5 Sources: Xl = Bulletin of the Central Bank of Ecuador, XLV, Nos. 531, 532, and 53 3 (Oct.-Dec., 1971), 274. X2 = Ibid., 209-213. X3 = Central 3ank of Ecuador, Fiscal Studies Division.  ANNEX A Page 49 Table A-29: GDP AT CURRENT FACTOR COST ORIGINATDING TN CONSTRU]CTION, ECUADOR, ESTTMATING PROCEDURE FOR 1969 AND 1970 Construction-GDP (Millions of Sucres) Independent Variables Year at at (Indexes, 1965 = 100) Current 1965 Index Prices Prices 1 965 =100o . 'X X2 1 - _ 1957 380 467 59.3 >52.2 60.t 80.4 1958 388 474 60.2 53.9 48.2 69.9 1959 462 564 71.6 52.9 46.2 94.4 1960 499 598 75.9 61.6 57.5 114.5 1961 574 655 83.1 67.0 48.6 119.1 1962 560 625 79.3 65.3 43.8 94.3 1963 593 636 80.7 79.1 47.8 98.6 1964 698 724 91.9 89.2 51.6 104.7 1965 788 788 100.0 100.0 100.0 100.0 1966 845 804 102.0 115.8 148.2 96.6 1967 1,040 955 121.2 131.6 109.8 119.2 1968 1,140 1,008 127.9 146.1 162.2 148.4 1969 1,329 1,120 142.0922 153.5 179.4 184.1 1970 1,446 1,142 144.9370 154.2 164.1 189.5 1971 1,663 1,199 152.1540 163.3 174.0 199.0 1972 1,918 1,256 159.3797 173.4 188.0 207.0 Independent Variables: Xl = Cement production X2 = Imports of construction materials X3 = Fixed capital formation in buildings, construction, and public works The above variables produced the following estimating equation for GDP at constant 1965 factor cost originating in construction: Yi = 16.2866 + 0.6587 X1 - 0.0807 X2 + 0.2127 X3 This equation produced the following estimates: Y1970 = 1,142 y1971 = 1,199 Y1972 = 1,256 The C3E computed the same estimatin-g equation (yi = 16.2805 + 0.6587 Xi - _ 0.0807 X2 + 0.2127 X3). Sources: Central Bank of Ecuador, National Income Division; Mission estimates.  ANNEX A Page 50 probably reflects the emphasis on heavy and urban construction where cement is more widely used than in rural areas which appear to be largely excluded from the estimates. The small and negative coefficient of net regression for imported construction materials may reflect to some extent the gradual import substitution, although statistically the result is primarily due to the lack of growth in the imported construction materials prior to 1965 and very large increases concentrated in a few subseauent years. For the fixed capital formation of the public sector, the coefficient of net regression is positive but three times lower than that for cement although public construction accounts for a large part of the total (see Table A-29). It is possible that the accuracy of public capital formation estimates leaves much to be desired although statistically inadequate estimates of construction GDP coulid equally be responsible for the relatively weak covariation of these variables. The inconsistencies between the value added by construction and expenditure on capital formation should be further investigated and explained. 62. Ecuador lacks construction and fixed capital formation estimates by industrial origin. The latter estimates could be developed first for the public and then for the private sector. Partial data for such estimates could be prepared by the Fiscal Studies Division of the CBE for the public sector and by the NIS for the private sector. 63. Electricity, water supply, and sanitary services GDP. -- These public utilities account for only about two percent of total GDP. They exclude gas and some private companies whose relative importance cannot be readily ascertained. For the 1950's and the early 1960's, the CBE estimated the value added by surveying municipal services and compa- nies in a few selected cities. The 1962-63 Electrification Census provided additional benchmark data for the electric power estimates. 64. Since 1965, the CBE has regressed the GDP estimates for these utilities on the production of electric energy in kilowatt-houtrs and on water charges reported by municipal governments in millions of sucres (see Table A-30). The coefficient of net regression is about 12 times higher for electricity than that for water charges. In fact, the data on water charges given at current prices may contribute little more than some distortion to the estimates. If they were properly deflated, the data on water charges at constant prices would reveal more clearly their lack of relevancy as their coefficient of net regressLon would probably tend to approach zero (see Table A-31). Therefore, tshe CBE may ccnsider eliminating the water charges series from the independent variables, basing its regression entirely on the production of electricity and then adjusting the results for the other services.  ANNEX A Page 51 Table A-30: INDEPENDENT VARIABLES USED IN REGRESSIONS FOR ESTIMATING ELECTRIC ENERGY, WATER SUPPLY, AND SANITARY SERVICES GD?. 7C,0TADOR, SMLECTMD YEARS 1957-1970 Variables and Indexes 1957 1960 1965 1969 1k7O X1 = El'ectric energy production (mil, of kilowatt-hours) 298.8 387.0 575.2 843.2 948.8 Index implicit in data 52.0 67.2 100.0 146.5 164.9 Index used in GD? estimates 51.9 67.3 100.0 147.8 164.1 X2 = Water charges (mil. of sucres) 15.5 28.6 34.8 53.7 .. Index implicit in data 44.5 82.1 100.0 154.3 Index used in GDP estimates 44.2 81.5 100.0 153.0 159.0 Sources: Xi - Unpublished estimates provided by the Statistical Section of the Ecuadorian Institute of Electrification (INECEL) for 1965-1970. The estimates for 1957 and 1960 were provided by the CBE. The primary source of the latter estimates could not have been ascertained. X2 = Unpublished estimates provided by the Fisc.al Studies Division of the GBE.  ANNEX A Page 52 Table A-31: GDP AT CURRENT FACTOR COST ORIGIWATING IN ELECTRIC ENERGY WATER SUPPLY, AND SANITARY SERVICES, ECUADOR, ESTmIATING PROCEDURE FOR 1969 AND 1970 Electric Energy, Water, and Sanitation GDP (Millions of Sucres) Independent Variables Year at at (Indexes, 1965-100) Current 1965 Index Prices Prices 1965=100 1957 120 148 54.8 51.9 44.2 1958 12. 154 57.0 5S.3 42.7 1959 138 168 62.2 60.7 59.8 1960 152 182 67.4 67.3 81.5 1961 179 204 75.6 74.7 63.5 1962 197 220 81.5 78.8 78.6 1963 224 240 88.9 82.5 9Q6.0 1964 215 25L 94.1 91.0 108o.0 1965 270 270 100.0 100.0 100.0 1966 311 296 109.6 105.8 117.7 1967 338 310 118.8 115.2 1L2.7 1968 393 347 128.5 132.0 181.3 1959 466 393 145.6291 187.8 153.0 1970 548 433 160.2829 164.1 159. 02 1971 646 866 172.6343 177.0 175.0 1972 758 497 183.9715 189.0 187.6 a/ Data estimated on the basis of the 1958-69 trend. Independent Variables: Xi = Electric energy production X2 = Water charges The above variables produced the following estimating equation for GDP at 1965 constant factor cost ori2inating in electric energy, water supply, and sanitary services: Yi = 'o1390 + 0.8743 X1 + 0.0671 X2 This equation produced the following estimates: 9 169 =39 Y1 97 = 433 Y1 971 466 Y1072 = 497 The CBE computed a similar estimating equation: (Yi = 6.3892 + 0.8783 X1 + 0.C671 X2). Source: Central Bank of Ecuador, National Income Division.  ANNEX A Page 53 65. Transportation, communications, and storage GDP. These activities account for about four percent of the total GDP in Ecuador. They are small as their development has lagged and there are considerable underestimation problems. For the 1950's and the early 1960's, the CBE estimated the transportation value added on an ad hoc basis; it surveyed a few trucking cooperatives and bus companies, and obtained some data from the railroads for 1950-1958, from Panagra and a few domestic airlines for 1950-1959, and from the Great Colombian Merchant Company. For the subsequent years, even the railroads ceased to provide statistical data and the crude estimates for earlier years had to be extrapolated on the basis of trends. River transport has apparently never been included in the estimates; warehousing and storage has been assumed to be negligible in Ecuador and these activities have also been excluded from the estimates. The public communication enterprises provided some data on factor inputs which the CEF. included in the national accounts estimates. The value added of radio stations was excluded from the communications and included implicitly as a residual with services. 66. The above estimates of GDP originating in transportation, storage, and communications have been extrapolated since 1965 by a linear multiple regression on three independent variables: (1) index of registered motor vehicles, (2) consumption index of petroleum production, and (3) in- dex of transportation employment (see Table A-33). The number of registered motor vehicles is rather poorly related to the changes in the value addedl of this sector: the coefficient of net regression is low and negative. Mlany of the registered motor vehicles are passenger cars which divert passenger traffic from carriers whose factor inputs are included in the national accounts. Registered motor vehicles include also many trucks whose transportation services are counted as a part of agriculture, construction, manufacturing, trade, and other economic branches. There- fore, it would have been more appropriate to regress the transportation value added on the number of busses and trucks used by the common and private carrier enterprises. 67. Consumption of petroleum products has a higher and a positive coefficient of net regression and it is somewhat miore suitable than the total number of registered motor vehicles for regressing the transportation value added because the common and private carriers consume relatively more gasoline per vehicle than those operated by other sectors or those used for pleasure. Transportation employment has about the same coefficient of net regression as the consumption index of petroleun products. However, the employment index is actually derived from the economically active population which the NEFCR projected from the 1962 Population Census. These population projections show continuous growth which outpaces the value added growth. In fact, the factor and the  ANNEX A Page 54I Table A-32: TI\DVEPENDENT VARIABLES USED IN REGRESSION FOR ESTIMATING TRANSPORTATION GDP, ECUADOR, SELECTED YEARS 1957-1970 Variables and Indexes 19969 X1 = Registered motor vehicles (000) .. 28.3 37.9 55.8 Index implicit in the data . 74.7 100.0 147.2 Index used in GDP estimates 57.2 7h.7 100.0 1h7.3 165.6 X2 = Consumption of petroleum products (mil. of US gal.) 86.41 173.6 229.h 322.3 362.1 Index implicit in the data 37.6 75.6 100.0 1h0.5 157.8 Index used in GDP estimates 5h.8 75.0 100.0 150.9 151.6 XX = Transportation employment (000) 35.8 h0.0 48.0 59.9 62.9 Index implicit in the data 7h.5 83.h 100.0 124.8 131.0 Index used in GDP estimates 74.7 83.3 100.0 124.8 131.0 Sources: X1 = GBE, National Income Division. X2 = Bulletin of the Central Bank of Ecuador, Vol. XLV, Nos. 531, 532 and 533 (Oct.-Dec., 1971), 271. X3 = National Economic Planning and Coordination Board, Division of Hummn Resources (unpublished estimates dated September, 1970).  ANNEX A Page 55 Table A-33: GDP AT CURRENT FACTOR COST ORIGINATING IN TRANSPORTATION, ECUADOR, ESTIMIATING PROCEDURE FOR 1969 AND 1970 Transportation GDP (Millions of Sucres) Independent Variables Year at at (Indexes, 1965=100) Current 1965 Index Prices Prices 1965=100 ______ ___ - -__-_- X1 _ X9 ' X2 1957 522 642 88.6 57.2 54.8 74.6 1958 530 648 89.4 61.5 60.9 77.3 1959 546 667 92.0 69.8 65.5 80.4 1060 548 656 90.5 74.7 75.0 83.3 1961 595 679 93.7 81.1 77.6 86.5 1962 600 670 92.4 83.1 77.0 91.0 1963 649 696 96.0 91.9 83.9 93.1 1964 701 727 100.3 96.0 92. 1 96.5 1965 725 725 100.0 100.0 100.0 100.0 1966 789 751 103.6 109.3 109.2 104.0 1967 822 755 104.1 127.5 118.8 107.9 1968 902 798 110.1 138.8 136.4 119.0 1969 980 826 113.9410 1L7.3 150.9 124.8 1970 1,050 829 114.3842 165.6 151.6 131 0 1971 1,193 860 118.7045 178.0 168.3 137.7 1972 1,365 894 123.3040 196.0 187.1 145.0 Note: The 1970 figure is revised to 1,149 because it is considered to be underestimated. Independent Variables: X1 = Registered motor vehicles X2 = Consumption of petroleum products X3 = Transportation employment The above variables produced the following estimating equation for GD? at constant 1965 factor cost originating in transportation: Yi = 63.o904 - 0.0530 X1 + 0.21614 X2 + 0.2036 X3 This equation produced the following estimates: Y1969 = 826 Y1970 = 829 Y1971 = 860 Y1972 = 894 The CBE computed the same equation. Source: Central Bank of Ecuador, N.Iational Income Division; Mission estLnates.  ANNEX A Page 56 nonfactor related inputs show faster growth than the value added of this sector (see Table A-33). The implied decline in the average productivity cannot be readily justified on any ground other than the deficiency of the basic data. An inescapable conclusion emerges that the growth, as well as apparently also the level of this sector, have been greatly understated. It should also be noted that the regression excludes completely communication variables. 68. Wholesale and retail trade GDP. The wholesale and retrail trade accounts for about 11 percent of the total GDP. These are crude estimates based on income tax returns of trade establishments and foreign trade data. Foreign trade establishments account for about 60 to 70 percent of domestic trade and the OBE has assumed the same proportioni for the value added, based in part on a study of trade margins carried out in 1950-1054. 69, Starting in the middle 1960's, the CBE began extrapolating the trade GDP estimates with regressions. At first, it based the regression entirely on the combined value of exports and imports. In more recent years, the number of independent variables was increased to four indexes: (1) agriculture GDP, (2) manufacturing GDP, (3) value of exports, and (h) value of imports (see Table A-3h). While all three GDF series are at constant 1965 sucres, the index based on the value of export permits indicates faster and that on import permits shows slowier growth than the indexes which the CBE has used in trade GDP regressions. It is possible that uhe latter two indexes were also deflated with different price deflators for exports and imports. 70. The selection of the independent variables for the regression of trade GDP is more appropriate than for the other regressions: agi~i- culture, manufacturing, and foreign trade reflect the commodity flow3 of the country. Nevertheless, the relationship of exports to domestic trade is very weak -- the coefficient of net regression on exports is almost zero. This independent variable may be eliminated from the regression (see Table A-35). Manufacturing shows also a relatively small coefficient of net regression. This is consistent, however, with the relative importance of domestically manufactured products in the total volume of domestic trade. She coefficient of net regression for agriculture is about twice as high as for manufacturing, and that for the imports is the highest. Although the level of trade activity ma;y be underestimated, its projection over time is generally in line with the overall quality of the national accounts. 71. Financ-,insurance, and real estate. -- These activities account for about three percent of the GDP in Ecuador. The CBE estimates  ANNEX A Page 57 Table A-3h: INDEPENDENT VARIABLES TTSED IN REGRESSION FOR ESTIMATING DOIMESTIC TRADE GDP, ECUADOR, SELECTED YEARS 1957-1970 Varia'Dles and Indexes 1957 1960 1965 1969 1970 Xi = Agriculture G;DP (mil. of constant 1965 sucres) 4,81h1 5,672 6,482 7,211 7,655 Index implicit in data 74.7 87.5 100.0 111.2 118.1 Index used in GlDP estimates 74.7 87.4 100.0 113.6 115.9 X2 = Manufacturin- GDP (mil. of constant 1965 sucres) 1,999 2,408 3,299 4,142 4,396, Index implicit in data 60.6 73.0 100.0 125.6 133.3 Index used in 3DP estimates 60.6 73.0 100.0 127.3 139.1 X3 = Value of exports (mil. of us $) 98.7 102.6 133.8 151.9 201.5 Index implicit in data 73.7 76.6 100.0 113.5 150.5 Index used in GDP estimates 70.5 82.1 100.0 108.8 129.1 Xh = Value of imports (mil. of us $) 97.8 115.2 168.9 261.9 247.6 Index implicit in data 57.9 68.1 100.0 155.0 146.5 Index used in GDP estimates 63.5 70.7 100.0 156.h 167.5 Sources: Xl = CBE, National Income Division. X2 = Ibid. X3 = Bulletin of the Central 3ank of Ecuador, XLV, Nos. 531-533 (Oct. -Dec, 1971), 209-213, and 237-240. Xh = Ibid.  ANNEX A Page 58 Table A-35: GDP AT CURRENT FACTOR COST ORIGINATING IN DONESTIC TRADE, ECUADOR, ESTIMATING PROCEDURE FOR 1969 AND 1970 Domestic Trade GDP (Millions of Sucres) Independent Variables Year at at (Indexes, 1965=100) Current 1965 Index Prices Prices 1965=100 1957 1,306 1,606 77.2 74.7 60.6 70.5 53.5 1958 1,338 1,636 78.7 75.5 64.4 75.9 65.9 1959 1,357 1,657 79.7 80.1 67.7 79.5 61.9 1960 1,482 1,775 85.3 87.4 73.0 82.1 70.? 1961 1,555 1,775 85.3 90.3 73.1 73.2 69.9 1962 1,620 1,808 86.9 98.0 77.2 82.4 72.2 1963 1,772 1,899 91.3 98.6 82.0 83.4 76.5 1964 1,921 1,993 95.8 99.1 95.5 89.5 90.2 1965 2,080 2,080 100.0 100.0 100.0 100.0 100.0 1966 2,185 2,079 100.0 106.1 101.0 103.3 97.5 1967 2,420 2,222 106.8 107.0 108.1 111.5 113.? 1968 2,680 2,370 113.9 106.0 112.8 116.9 135.3 1969 3,039 2,560 123.1 i1ii. 125.6 108.8 156.4 1970 3,397 2.683 129.0 118.1 133.3 129.1 167.5 1971 4,330 3,122 150.1 120.1 141.0 129.9 ?31.7 1972 4,987 3,266 157.0 120.6 150.4 182.1 251.2 Independent Variables: X1 = Agriculture GDP X2 = Manufacturing GDP XR = Value of exports Xt = Value of imports The above variables produced the following estimating equation for C!I)P at constant 1965 factor cost originating in domestic trade: Yi = 32.8597 + 0.2413 X1 + 0.1315 X2 - 0.0082 X3 + 0.3057 X4 This equation produced the following estimates: Y19- =-2,560 Yj970 = 2,683 -11971 = 3,122 Y1972 = 3,266 The CBE estimating equation has a positive sign for the X3 coefficient (Yi = 32.8592 + 0.24127 X1 + 0.13145 X2 + 0.0082 X3 + 0.3057 X4) which produced slightly higher estimates. Sources: Central Bank of Ecuador, National Income Division; Mission est imates.  ANNEX A Page 59 the value added of banking and insurance companies by aggregating compensation of employees, rents, interest, depreciation, and profits of financial intermediaries. The Superintendent of Banks provides these data from the accounts of financial intermediaries which it supervises. Real estate is considered to be quite unimportant in Ecuador and it is largely excluded from the estimates, although the CBG. has made a rough estimate for 8,000 brokers without a permanent business location. The ownership of dwellings is estimated separately (see below). 72. Prior to 1958, the CBS made no imputation for banking services. In subsequent years, the CBE imputed the banking services as a difference between interest and dividends received less interest paid. There is, however, no indication that the imputed service charge has been properly divided between the business and the household sectors. The Superintendant of Banks may not have the bank deposits properly classified by type of depositor to enable the NID to estimate the enterprise share of the imputed service charge and to deduct it as a business expense from the gross profits. 73. Ownership of dwellings. -- This sector accounts for about six percent of total GDPo It includes imputed rents of owner-occupied dwellings of the private sector. The irputed rents of public buildings are included in the government sector. For l9qO0-l96, the CEE derived the estimates from the rent registers of Quito Guayaouil, and from municipal cadastral surveys of a few othe-r cities. For subsequent years, the CBE extrapolated thpse estimates on the basis of trends. 74. Services GDP. -- Private services accounted for 9 percent of GDP in the early 1960's. Since 1965, the CBE has estimated them together wi th public administration and defense is a r,esi.dual, that is, total GDP less all the separately estimated sectors. The CBE estimattes of the 1950's were considerably understated. Services included education, health, religious institutions, welfare and social services, independent professions, entertainment, general services rendered to enterprises, hotels, bars, barber shops, beauty parlors, etc. Many service enter- prises could not be properly classified between the private and public sector. The major stumbling block remained the lack of basic statistics on services. 65. In the middle 1960's, the CBE tried to extrapolate private services by regressing their value added for previous years on five indexes: (1) index of matriculated students, (2) index of incomes by educational level, (3) index of incomes by health level, (h) index of public entertainment, and (5) index of imported books, records, and paper. These extrapolations, however, failed to produce meaningful estimates and the CBE finally decided to combine the services with public administration and defense.  ANNEX A Page 60 Table A-36: INDUSTRIAL ORIGIN OF GDP AT CURRENT FACTOR COST ALTERNATIVE EOTIMYATES OF THE CBE AND NEnEPB, ECUADOR, 1969, 1970, ATD 1971 (Millions of Sucres) CBE Estimates CBE in -Mission Estimates Year NEPCB of NEPCB Based on Final Regressed IPublished Estimates Estimates Regression Es;tiates AgricuIture 1969 bj574 8,574 5,716 101.8 8, 472 T37fl42 1970 10,271 10,271 10,008 102.6 9,693 j 2,693 1971 .. .. 11,343 .. 10,804 1 CG,804 Mining anuL Quarrying 1969 -5-9-6 596 585 <~101.8 5d 1970 621 641 630 101.7 604 641 1971 .. .. 773 .. 678 734 Manuf acturlrng 1969 1 53062 4,764 1 4,650 101.7- 4,7917 4,7h4 1970 5,921 5,657 5,555 101.8 5,565 5,438 1971 .. . 6,735 , 6,450 j6,412 Construction 1969 1,350 1,350 1,324 101.9 1 ,329- 17350 1970 1,473 1,803 1,770 101.8 1,446 1 ,8603 1971 .. .. 1,982 .. 1,663 Electricity Gas, Water SuppI and. SanitarServices 7969 474 474 l 466b l 101.7~ - 466 -L74 1970 559 559 549 101.8 548 559 1971 . .. * 65t 646 } j Transportation and Communication 196 --'-995 - -995 977 10195 092