SWP166 ' Thip paper is prepared for staff use and is not for publication. The views expressed are those of the author and not necessarily those of the Bank. INTERNATIONAL BANK FOR RE;CONS'LIRUCTIONAND DEVELOYHEh'T INTmATIQNALDEWPMENT ASSOCIATION Bank Staff Workind Paper No. 166 A Ouide to the Ouidelfnes: The TINID0 Mathod - of B c o n d c Project E . . November ,. x m '.: ... (Rev. L/1/7L) This paper condenses into an operational hmdbcok the Wdelines f o r Project Evaluation, which were writ.+&n by Professors Partha Dasgupta, Amartya Sen 2nd Stephen Marglin and published by UNIDO. The evaluation process has been subdivided into four stages which focus i n in- creasing d d x d l on the social benefits of ~ r o j a c t e . This llOuidewuses an Integrated Document.aticn ,Systm. which is applicable t o all methods of econa2c project eosluaticn. The author is particularly indebted 50 Ste3hen ' Harglin, Glenn Jenkins, Max Corden, Ptcphen (hisinger and Richard Berney for their detailed review and helpful sug- gestions on an earlier d r d t of this paper. The zuthor has also benefitted from discussicns w i t h colleagues with- ir. the World Bank including Colin Rr-ze, Paul Duane, Deep& Ross Parish and Lyn Squire. The ddthor of co-cse asfanes sole responsibility for any e=ror3 which r m a h - an for the views expressed here*. Asia Projects Department 63 Prepared by: John R. Hansen Tzble of Contents Page Stages of Analysis .................................... 2 A Senae of Priorities ................................. 3 --.. ........-..., . DIRECT SOCIAL BENEKTS ..................... ', 8 Recording the Foreign Exchange Impact ................. General Principles of Shadow Pricing ..................... ..................-a Shado~rPricing Specific Resources .......... Tradable Inputs and Oukpput~- = ' ~ * ~ H J : : * . . A P * . ; . ~ . - ~ ~ ~ ~ , . . , ......~ Non-Tradable Inputa and Outdtrz --re Non-Marketed Resources C-L *a **e~vr*ar-na%)F~ --3--- Labor -U~S .................................... A 8 ~ ~ ~ ~ ~ ~ ~ ~ ~ 0 ~ ~ ~ ~ ~ 8 8 8 ~ 0 8 0 0 0 0 0 0 0 8 ~ 0 ................................. Capital Inputs ............................ Foreign Exchange Shadow Price Manipulations ....................... IN=-TEMP- DISTRIWTION BENEFITS 37 INTER-PERSONAL DISTRIBUrION BENEFITS ....................... && A. The UNDO and Little/Mirrlees Methods Compared B. Standardized Analytical Tables . e C. Shadow Pricing . D Distribution&l Analysis -. Swmary Abstract Ths UNIDO Guidelines present one of the most thorough discus- a i m s of the basic issncs of econanic project evaluation available today ye However, the rigor of the analysis and the length of the book have created a demand for a more mndensed and somewhat less demanding introducti m t o the method. T h i s paper follows the under- lying philosophy of tha Ouidelinea very closely, although many of the outward appearances have been modifled s ~ 4 q 8 ~ ~ c i a f i ~cr.ra& t ? b ' an operational guide t o economic project e v a l u u i t P a a ~ * ~can' f.;e wed-* A technicians i n the developing countries . The method suggested i n the Ouidelines may be broken down into four stages: (1) calculation of commercial profitability a t market prices, (2) calculation of the project's net aggregste con- sumption benefit a t shadow prices, (3) adjustment for the social value of investment, and (b) addition of benefits deriving fran the accm- p l i s h e n t of other national goals such as income redistribution. These stages are handled i n tfle present paper through step-? y-step adjust- ' ' menta i n a s e t of Standardized Analytical Tables. he UNIDO ~ d e ~ are isn many ways s ~ l ajro the well- e 1 - known Llttlefiirrlees Hanual published by the O E D 2/, d most of - * ., P. Dasgupta, S.A. Narglin and A.K. Sen. Ouidelines for Project Ehmluation. Vieumax UNIDO (ID/SlB.H/2), 1972. 2/ Im He DS L i t t l e and James A. Hirrleee . Manual of Industrial m5. Project Analyais i n Developing Ccuntries, Pol. 11. Paris r O E D , the differences have been reduced or eliminated i n the course of the e x t a s i v e debates which have followed the publication of these two methods. The changea i n t h ~ Uttle/Xirrlees method are expected , t o be reflected in a revised edition of the Manua1,due for publica- tion in the near future. Similarly, the original Widelines have been modified in this "(hide t o the 0~5ds1inaa.~~,Nevertheless, L t is instructive to note the following Mf&e&;.fn the origin& versions. The major outward diffetmence-k the use i n the Guidelines - f of average consumption expressed <;ar?lestilr currency ."ad. tb-re instead of free foreign exchange i n t h a kmbs- &:r=fie -;r'overnmmt. Thia, however, is primarily a superfj :id difference. The m a i n dif- ferences are i n emphasis rather than content, but as these differences can lead t o the choice of d f f e r e n t shadow prices, they are important i n practice. For example, the (hidelines place more emphasis than did the original Uttle/Mirrlees k m u l an (a) the pursuit of social objectives such as incane distribution in addition t o the madulzation of social CeJP, (b) the externalities of industrial developmt, (c) the use of L an iterative dialogue between the technicians Itat the bottomw and the policy ma!-ers 'lat the topt1 (the nbottom-upft method) to determine the apl .opriate values for subjective p mametera such as the social discount rate and redistribution weights, and on (d) b e possible relevance of w damestic instead of border prices as the correct shadow prices in countries where non-optimal trade policies are llkely t o continue. The United Naaons Industrial Developnent Organization (UNIDQ) recently published a book entitled Chidelines for Project Evaluation v. This book, written by Professors Dasgupta and Sen . , of the Lcndon School of ikon~nics Professor Marglin of ' i f . Hsvard University, started as a s e t of 'backgroundpapers in 1966/67. On tine basis of field testing i n a sertes of rratfmzl'tra3nirrgliork- shops, these papers were modified w 1 ~ $ 3e i n an effort fn gad*v. I , "a set of guidelines which the developing,countries c o d 8 ase fa incorporating the evaluatbn and approval of new industrial projects into their overall industrial planning mechanism. It2/ This l1&dde t o trle Wdelines haa been written t o present, within the context of a relatively few pages, the basic steps of the evaluation process in a non-mathmatical form. Sane camnenta are mcde regarding the theoretical justificatians for various steps, but the reader should refer t o the (hidelines themselves for detailed wathematical explanations .., Since the primary purpose of this paper is to provide a .--.----- - " . * I - 4,I working gutdezto the hidelines, the method has been br &en d m into 9 a series of an-alytical s t e p . Each of these steps is based on * - I y U N D O Ouickhes for Project Evaluation (Project Formulatim and Evaluatim Series, Nb. 2). (New Yorkr Unitied Natiohs (ID/sER. - . H/2, Sales No. E.?~.II.B.II), 1972). Op. c i t . page v. modifications of a s e t of Standard Analytical T ~ b l e s ,which are explabed i n greater detail i n Annex B. T h e ~ etables lead the analyst in logical ateps from standard financial analysis t o the cmpletion of an economic evaluation of a proSect by the UNIM) method. . A s thu reader may already be faniliar with t h e o r i g i n z l Little/ P;irrlees E2nua.l for eccnomic project evaluation 1/, P.nnex ,1 has been included t o ccmpsre the ztpproaches take:. there and i n the U N m O book, which was published four years later. Annexes C and D provide additi>maldiscus- ..' # ' sions of shadcx pricing and the measurement of Fntertemporal and interpersonal becafits. It is zssumed that the ruab ag sa~~cxiaf fluent i n economic terminology. Stages of Analysis In the absence of any explicit stepby-step ~ u t l i n eof the method i n the Guidelines suck as one finds, for example, in the OECD Manual, an applied methodology was derived from the case studies - given a t the end of the book. Considerable liberty has been taken in defining the stzges of analysis, for the case studies themselves use somewhat different approaches. Although the methodology described . . L below is not the same as t h a t used i n any partj-culer case study, it is consistent with the underlying philosophy of the Quidclines and - i w i l l , I hope, provide a logical and unifom appro='d which can be reidily applied by project analysts in the developis countries to 4 m a k d e variety of projects. -1/ Ian M.D. Little and James A. Mirrlees. Manuai of Industrial Project Analysis, Vol. 11. (Paris : OECD, 1968). The Wdelines method of project evaluation has been broken down here into the following four stages: (1) Calculation of commer'cial profitability a t market prices. (2) Shadow pricing of resources t o obtain net aggregate consumption benefit a t economic (" e f f i ~ i e n c p ~prices. ~ ) (3) Adjustment for the projsctBa inveahnent/consunptio~t impact . (L) Adjustment for the project's impact on other social objectives such as incane distsibak&om- A Sense of Priorities Those who evaluate projects by benefit-cost criteria must evaluate their own work by the same criteria. It is easy to became involved i n the theoretical details of economic project evaluation and carry it to the point uhere it produces cnly superfluous infonna- tion instead of better investment decisions. A simple, unifon cutzff point for analytical detail cannot be established for a l l projects in all count.ries because the needs and uses for the information vary too widely. Therefore, sane of the details of the buidelinee analysis which are explained below will not be worth incorporating in sane project evaluations. These details must be- mentioned t o do even partial juetice t o the richnese of ai~alyticaldetail built into the Ouidellnes and to make tllen available i f necjded, but the a n a l p t must decide which parts will be useful i n evaluatir~ga given project. Tn making these decisions, the following guides might be consideredr (1) A good technical and financial analysis must be done before a meaningful economic evaluation can be made. For this reason, Stage all cases JJ. (2) A commercial profl+,ebility analysis would be sufficient only if the market within which the project is to operate were reasonably ." "perfsct If the market is uimperfect,N (as it often is in developiw coun- I tries where market prlcem do not always refled the social velne,of re- sources z/), the llSocial Ranefitft analgzfs of Stage 'Itro issraqafred. Here inputs and outputs are assigned their *efficiency!? sbndw,,rv-5c,es . (3) Stage Three of the U N D O analyaib C~d;IuaQmat,$01: the Divest- ' , ment/Canaumption Impact? w i l l be especially important for projecta which generate benefits going to groups with a higher or lower than average marginal propemsity t o aave, and for countries where there is a significant gap between actual and optimal savings. (4) The importance of Stage Four of the analysia (Other Social Objectives) depends on the existence of goala other than maximizing the gross national product. Of these, regional and interpersonal in- come redistribution are likely t o be the moat important. . These empirical I questions m u s t be anewered on a case-by-caae baeis. l.JSeezAgencyforInturnation&.Developnent Mdelines for Capital - Roject Appraisal (Wpshingtaa, AID, 1971j for a very useful guide t o @he technical and cammercial aspects of project evaluation. 8 0 1/ The tonu ttre80urceaflis used i?lthis paper to refer to inputs and outputA as a group. It may be questioned whether or not it is reasonable to expect a government t o be I1rctionall1regarding these goals in se- lecting project8 when it is llirrationallli n its macro policies. While this is a valid concern, we w i l l assume here that the govern- ment fin& it pcssible t o take actions at the micro level which are not possible a t the macro level because of various political and administrative constraints. .,> This stage produces an estimate of the project's c o ~ s r c i a l profit, or, in the language of the hidelines, the pras&,y+lue of -.., - ._z 4 the net consumption macie possible by the proj8ct--. when a l l i ~ ~ p uand t s - -. --.. outputs are measured a t market prices. -FinancialStatements : The f i r s t step in this stage is to ccmplete an ir.cuma statement, a balance sheet and a cash flow table for the project a t maket prices IJ'. Some r o w of the standardized forn of the tables -All rexah blank a t this stage since they are dasipred to cover economic as well aa financial analysis (eg. Table B1, lines a.1th>ougkL-a.!~).incane The staterrrent ie the central table in this analysis, for it is used to record all of the inputs and . R b outputs of the project. The values entered in the table a t this point are a t market, not shaaow prices. - 9 - In preparing these tables, a cholcs m a t be made between current - C and constant prices. Forecasts in c u r r e n ~ r l c e a ,which take into account m the effects of inflation, a m necerssazy t o give a correct pictare of the y See Annex B for eanplee of tables referred t.o throughout the r e s t of thie paper. actual financial posiuon of a project in any given year with respect to the costs of inputs and outputs. Haever, such forecasts do not give a realistic picture of the ir,trinstc profitability of a project. Inflaticm can imprwa its apparent profitability in a discountd cash flow analysis, provided prices of the output go up a t least as fast as those of cummt inputs, by increasing the net money tenefits i n the future 2s compared w i t h the capital costs a t today's lower money prices. %day 1scapital costs are repaid with ncheapern future dollars. However, h the longer tern, a project relying on current price accounting'vlll find that its deprecia- tion allowances are inadequate to replace capital equipment.a t the prices prevailing a t the tine of replacement. To amfd-TATs;' E%,a.u=k r e ~ a l u eassets in l i n e with inflation I./. - For these remonsj-wsi'mom --- - ,. * - Fmportantly, because the economic analysis based on this table must be done i n c o n s t a t prices, the fi~comeand other financial sta.tements shouid all be done in constant prices. If differential rates of inflation are forecast, adjustments for changes i n real relative prices w i l l be necessary. This is done most easi1.y by a-abtracting an appropriately weighted average rate of inflation from the individual rates, thus netting out the inflationary trend, leadng only , relative upward and downward price movemen-. Assume for example that two inputs and one output are involved and that their prices are forecast t o - r i s e a t u,6%and 8% respectively per year, while the eznual average rate * - of inflation of the goods ,is forecast a t 6%. Subtracting 69 from each of - - * the above gives relative p k c e movements of -8, @, and 2%. rr mese factors should them be applied to the present annual values to obtsin the future -1/ The only real gain fmm the point of view of the project (but not sociew) which inflation may produce can arise if capital is borrowed today and repaid with cheaper dollars in the future. The effects of differential rates of inflation s h d d of course be considered, for they w i l l change the real relative values of resources over time. - ~fcomtant"prizes appropriately adjusted for changes in relative real values. Because both the financial and the e c o n d c analysis in the UNIDO method is based on a discounted cash flow analysis, the income statemat, which r shows annual profits, is not the most convenient basis for doing the actual discounting calculations. 'Ihe Standard Cash Flow table ne able B3 in Annex B) r is much more convenient. Ihe net cash flm i s derived from the standard in- come statement bg standard accounting procedures and is equal to the grqss cash d' flaw (operating profit before taxes and In-st ' plus allowances for w e- c i a t i o n ) . , 4' : minus capital expenses (line 2.1. in Table 33) For the purposes of sh&w pricing capital expenees, it is convenient to have one more tr.bIe, the Standard Balance - Sheet (TaSle B2), which breaks capital co%tr dm Into v~&.mflcdlb~&.&. I t AddLtional tables which feed into this table can be used. &o.%~%g&%d ~ M s a costs into imported and domestic materials, labor, and other costs which can i n turn be assigned shadow prices. Disccuntlng: Once the financial tables have been prepared, the next step is to discount the net cash flow shown in Table B3, Une 4. for each year to the pmsent using selected discount rates -1/. The project is commercially . acceptable if: - the project ha8 a non-negative net present value a t the investor's opportunity cost of capitafl, or, i n the same veln, - the internal financial rate of return is acceptable. - - If the project is not commercially viable but does hare a ! -positive net econodr: value as calculated in the foll~wing steps, ** 8 -subsidies or market price adjustments w i l l be rtjquired. Conversely - - - -- -- y In a l l stages of the Guidelines analysis it is suggested that the analyst use three discoirnt rates to test the sensitivity of the project to high, middle and low assumptions about the discount rate, for this is generally not houri with certainty. In this stage of the analysis, appropriate commercial intersst rates should be used. In later stages, the social discount rate should 39 wed. If x3ak i u c~nsidered,the robabili w of obtaining an adequate rate of return or a positive net preaenr va ue muat also be considered. if it is cmercially viable but is revealed to be econmicaily unacceptable, it will be necessary t o reject it, o r redesign i - L i f ~08sible DIRECT SOCIAL BENEFITS This stage produces an estimate of the net present value of the project measured a t shadow pFfces instead of market prices. Before dealing with tt-a shadow prices per se, a couple of topics should first be considered: How far should shadow pricing be carried? And how can tho foreign exchange .@pact of a project be ccnsidered i n a method which usos domestic currency i n the n- -= .-.. -. . '- Selective Shadow Pricing Shadcw pricing should be done selectively i n terms of two criteria. First, what resources figure most promiaently i n the benefits and costs of the project? Second, for a l l resources ii~volved 7 in the project, which market prices are furthest out of line with their respective shedow prices? This concentrates scarce analytical resources on the most important divergencies between m i ~ l ~ and e t social profitability. Houever, two caveats are in ordar: . r (a) Both of these c r i t e r i a muat be a p l i e d aimultaneously. If attention were focused solely on those resu.l=ces which ,9 were most prominent at market prices, a socially undesirable croject - - - C which depen-d heavily for its financial success on, say, g r a l y under- * priced electric power could be accepted because the power would appear as a very minor item in the market price analysis. (b) The danger of a non-optimal alloca5iion of resources is always present when market and shadow prices are u s ~ dtogether in - 9 - the same analysis. l'he problen of l?second-bestlloptimization cannot be pursued h-ther hsre 1/, but it can be generally ass'med that the criteria - above will lead t o a better allocation of resources than wmld have resulte* from the use of distorted market prices 2/. - The most obv',ous candidates for shadow pricing are the output (which constitutes the e?tire benefit stream, if there are no exA,ernalities, and which is often sold at protected prlces); the material inputs (which are also oftea protected) ; and, in sane cases, . unskilled labor (whose market wage often arcedssfts. shnrtrru w e ) Recordinp the Forbign Ekchange Impact Since the UNDO method uses domestic currenqr as the nurneraire 3/, - it is necessary to identify separately the foreign exF)rargb bpi&, 36 that the project's net present value maj be adjusted by an appropriate prhliwn dependir~gon the net foreign exchanga impact of the project. 'dhile it is cmcqtually possible to include in the shadow price of each resource a premium according t o its foreign exchago impact (impact on trade), the approach suggested here is to: (a) record, i n addtion to the basic shadw price of each resource its fomign exchange impact on the input or out~uside of t the project's ledger, . t (b) discount these impact values back to the present, (c) l~ultiplythe total net foreign exchange impact (eg. $1,000) bg - a '. -1/ For z.1 excerlent introducticn t o these problems, see R .a. Lipszy and K. ~ a n c B t e r ,'!The Bneral Theory of Second Best," Review of Zconmic Studies 24:l (1956), pp. 11-32. -2/ To quote a refreshing pdsage from Vijay Jcdhi: "There is strict- - l y speaking, no theoretical guarantee that decisions based on ...in lapproldmately correct1 shadow prices will lead the economy in th6 right direction. my opinion, an 'act of f&itht has t o be made if such cosmic doubts ar raised about the whole basis of a ~ p l i e dwelfare econmica.I1 fVhe Rationale and Relevance of the Ut'Ule44lrrlees Criterltn, "Bulletin of the Oxford Ins+,itutn of Bcono~icsand Statistics, 3h:l (February 1972), p. 4. ) Annex A, pp. 3 4 , discusam tbe qualitim of rmmeraires. a foreign exchange p r d u m (eg. 20%) and then (d) add this product (eg. $200) to the total net present value of tho project. Besiaes reducing the number of individual calculations Y involved, th2s approach has the minor additional merit of providing an indication of the foreign exchange:.%pacttt of the project . . Tne foreign exchange impact of each resource 2s recorded separately from its basic Stage 'Iko llefficiencyllshadow price because, until the impact on consumption and saving is exantined in Stap,+ree, the foretgn ex- change impact may not be apparent or directly related to the sh&m price: -- If the opportunitg cost of labor is say 3s 15 per day,but the wags rate is Rs 25 per day, the basic efficiency shadw price is only Rs 15. However, ol the additional Rs 10 of consumption, a certain proport ion -- perhaps 15 percent or Rs 1.5 -- w i l l be spent on inported goods. Since these expenditures would not have vccured in the absence of the project, it is necessary t o include them i n the cost of the project. The difference between the Itdirect" and llindirectltforeign exchange impac,ts of a project is arbitrary; furthermore, i n an "optimal" . I economy most resources could be used for earning or saving foreign exchange instead of producing non-traded goods -- However, a pragmatic 9 - definition is reconmended here: a foreign exchnge impact exists 0 0 when the project uses or prodlices goods which wuld otherwise be - 0 This approach can be used only i f the shadow price of foreign exchange is assumed t o remain constant over time. If the premium on foreign exchange is expected t o f a l l in the fu%uredue t o a move towards freer trade (or, conversely, i f it is expected t o r i s e due, for example, to mcre restrictive trade policies), the year-by-year adjustment process outlined on page 36 w i l l be required. traded (imported or exported). If tradable goocis are identified in the process of breaking d o n the non-traded inputs to a project, these should also be included on the foreign exchange impact side of the ledger. Thio disaggregation proiess w i l l usually be carried backwards only one or two ~ t e p s ,thus limiting t o manageable propor- tions the problem of i d e n t i f w , , t h e direct foreign exchange impact. C .' $ 6 ' a . This approach is roughly the same as thst suggested {bythe Guidelines : "Any input whose supply fs fneneased t r r reapawe t o a project must be valued according t o the re- ---IS., sources used up in its produc-kia~~.~ ?.-, p- sources i n ~ l u d e - - ~ R n ~ ~ 4 x d W f i g a & b w '':I imported inputs, via expHable6, or -via impor$ , subetitutes -- tW3~~ djrtfeilt the rel"ewma5 net input comists of foreign exchange. (p. 58) Qiven the margin of error i n predicted figures used in most project evaluatioxs, it might be reasonable t o say that, unless the inclusion of a given foreign exchange impact o r group of impacts can be expected t o change the t o t a l output cost by a t least one percent, they are not worth tracking d m . Knowing the extent of overvaluatim of danestic currency (see pages 33 - 36 below), this guideline can be translated into more concrete temps. I f ,for axmaple, the shadow price of f o r ~ i g noxchange is 20 percent higher than the market exchange rate, the input would have t o constitute about 5 percent of the cost of output at market prices t o increase - by 1% final value in shadow prices. the Major material inputs would obviously qualify without difficulty, but many minor nontraded goods, especidly those which are only cmponents i n other non-traded inputs, could probably be ignored. Canmon sense is required, of course; 20 items each constituting only 2 percent each of the total value of output obviously become significant when considered together. To f a c i l i t a t e the foreign exchange shadow pricing done i n t h i s stage of the UNIDO analysi .,it would be best t o s e t up the original Standard Incane Statement (Annex By Teble 1 ) with separa%e columns for the baaic price and ths fmgfgn exchange,impact. Doing t h i s during the C m e r c i a l R o f i t a b i l i t y Stage of the analysis w i l l greatly facilitate the analysis in this section. In reality, if it can be assumed that the foreigq sxcbw~f4Wi%%i$mw~;6rp06i- .- ... , U tion w i l l remain constant over the liraa& pH'@iE%,"i'l* &pFs -- - ~6 may be simplified greatly by recording only t h e total values for each year, discclmtbg them back t o the present, splitting out the foreign exchange impact from the net present value by the foreign exchange/dmestic currency canposition ratio and applying the premium t o the foreign exchange campanent. Oeneral Principles of Shadow Pricing In a 'perfectf market the shadow price for any resource is its equilibriun~marketprice. I n this ideal world, t h e marginal social value assigned from the demand side by the users in the form of mar6 e u m e r willingness t o pay would be exactly equal t o the marginal - 6 w i a l cost of supply +o the producers. If tk.e resource moved in * # s t a r n a t i o n a l trade, equilibrium muld b. reached a t a price equal to- the border price. The perice would not move higher, for consuners would import it instead; and it would n r ~ move lower, f c r producers would export t . . it rather than sell for less on the domestic market y. irlF, - *c:,*, - I- *,? ~~ Annex C describes the 3hadow pricing process In graphical tern. In the real world an equilibrim of sorts does exist between deanand and supply, and this equilibrium bears sane relation- ship t o the border price when a tradable goods are involved. However, because market brperf ections create distortions on both the demand and supply side, there is l i t t l e chance that the market equilibrium price w i l l equal the social value and cost of the resmrCeSm , 'J.. For tradable resources, the market price is 'likely )to.be higher than the border p r i c e y Because of the distortions, it is probable that, given the amount of the resource presently avdlable "to the economy, the marginal social cost as seen from the supply--s2 de -nii&'ttiq mal.ginal . . , & t . social value zs seen from the demand side w i l l not be sqzal. It tken becun ---.. necessary t o examine whether the impact of the project w i l l be on demand o r sapply before determining the shadow price of the good. The Chidelines recommend three different sources of shadow prices depending upon the project's impact on the econoiny q. A project, thrargh its use and production of resources, may for any given resource affect (a) the supply available t o the r e s t of the economy, (b) the levsl of its production i n the r e s t of the eccmomy, an4 (c) the level of imports or exports of the resource. In terms of production of an output, the project may: (a) Increase the t o t a l c ~ u m p t i o nin the economy. - 6 (b) Decrease production t? other parts of the economy. (c) Decrease imports or a c r e m e e q o r t s . f y The main exception is the export good subject t n anexport tax. In certain cases, it may be appropriate t o apply shadow price8 on a regional rather than a national basis. The corollary impacts of a project's consumption of irkputs may be to: (a) Decrease consunption i n the r e s t of the economy. (b) Inti-ease production of resource within the economy. (c) Increase imports or decrease exports. A project might havs, all three of these impacts simulta- neously on the output and/or inpatkaIde.t: Assume, for ,example, that a t present 20,000 automobiles are produced annually wit6in a country and 20,000 more are tnported under a f h e d quota. A prdject is proposed t o produce 15,000 unit&-ph,:tai;~As- ~ . . 3 4 ' ~ u e the government w i l l lower the-~~&qgqzt,qgotaby S,QQQi-r;mr'd6-that ' 5,000 cars produced under the project will substitute f o r imports. A further 5,000 will expand available t o the econmy. Since t h i s additional consumption w i l l involve a decrease i n price, the fins 5,000 unite from the new, more efficient ~ r o j e c tw i l l replace high cost production a t the margin by other domestic producers. The project, therefore, nay be said t o have all threa types of impacts. While the Guidelines generally speak of only one type at a time for a given resource, there is no reason whybmultiple impacts cannot be considered. - - The appropriate source of shadow prices for the output of J * cars under each cf these three mnditions is different: *- C Impact Basis for Shadow Pricing D (a) Consumption within the (a) Consumer willingness t o pay economy (b) Production wiL5in ecmomy (b) Cost of production (c) International trade (c) Foreign exchange value To illustrate the valuation appr0ac.h which should be used for each of these impacts, l e t us continue with the example of the motor cars. First, the 5,000 which increase the number consumed in the economy should be value.= a t the price consumers are willing to pay for the additional cars.'y. Hmuer this pdce f s tnot neces- ,%\ sarily the current price i n the market. For t h i s market price t o be the appropriate shadow price, the following fourconditions must hold: --- - (a) The good is freely a v d l t b U #i&-.mma'~~XiPL~'&ing t o pay the market price. In the case a f " . - , ~ & ~ s s . . e LS artificially restricted -- eg. by a quota on the production of cars -- this condition impiiea that the market price must be high enough t o l i m i t demand so that the number of willing customers does not exceed the number of cars available. (b) No customer buys enough of' the total output t o exercise monopoly buying power (monopsony 1. This would artificially lower the market price below the actual consumer willingness t o pay. (c) The change i n the suppiy'of resources available to' the reqt of tihe ecmany doeq ,not change their p r i . - This implies inelasticity of price ui& respect t o quantity available and/or a L project so small that ifSconsumption or production of resources is * negligible with respect t8 the total danestic market for these 3/ It is assmued here that the number of vehicles imported is fixed s o that the additional consumption would not be met by imports. Under more liberal assumptions about trade policy, the increased consumption would otherwise have come from imports and the pro- ject's output should be valued a t border prices aa indicated i n Annex C. resources. If the project w i l l raise the price of inputs a t the margin or lower the market price of the output a t the margin on either domestic or imported resources, the values a t the margin must be used; This in turn requires a knowledge of denand and supply elasticities. -I ' (d) With respect td;hmts only, the market price is not _ . . . artificially inflated by producers who make monopoly profits on their output. If these conditiam &.d- H a h i . the market ; - c f - > 3 price based on assumptions regas+&g*thw W~W~Q, and consumer surplus must be made to obtain the lltrue" consumer willingness t o pay. I I Taxes often pose a problem i n calctlating the appropriate I shadow price. The general rule i s that, when the impact of the pro- ject is t o take inputs away fran other producers or add t o the goods available to ccnsumers, taxes should be included as part of the indication of the consumer willingness t o pay or the marginal social value. I However, i f the impact i s t o generate more production of inputs or 1 . t o reduce the production of the output by other projects,taxes should be excluded, for they do not constitute part of the marginal social .. $4 *$ :$&.* . - y- j&;--.. !*. hS.4 I 14.t 1%- , *i1 ! I - @---- < +t ' cost of p-oduction y. ? Second, the 5,000 cars which would decrease other domestic I production should be valued in tem.8 of the marginal social cost of -1/ This applies only t o taxes which are pure transfer payments and in situations where the government does not have difficulty raisingthe Nol;timalflamount of government revenues. Taxes which are i n effect charges for services rendered by the government should be included a part of the cost. If the government requires product taxes t o raise the necessarp revenues, the net impact of the project on govemmmt revenues should be evaluated as part of . Stage Four where the different val.ras of income i n the hands of different paFties (eg private versus government) are coneidered. production of the inefficient domestic producer. he net benefit *>f t h i s part of the project would be the cost of the inefficiently pro- duced cars less the cost of producing them i n the new project). The analysis of production costs of qompeting or supplying firms should, of course, be done in terms of shadow prices, anqin somc cases this would lead t o further analysis of the same type farther back the input/output chain, This input/output type of snalysis should also be used when the project's use of i.riptxG3tihuJAtas ~ ~ r . ~ ~ ~ r l ~ ~ , -d ~ q l n e s TT * production of inputs be they p d fsXIY , traded or-non-traded. There is a substantial similarity between the UNDO and L i t t l e h i r r l e e s methods in t h i s respect. However, the authors of both methods agree that this type of analysis can quickly becane too difficult to be worthwhile and should only be carried out t o the extent that resulting modifications i n shadow price are significant vis-a-vis the t o t a l project and that they exceed the uncertainties inherent i n the calculations. .. Finally, athe 5,000 cars uhich would substitute for imported -- vehicles should be valued a t their border prices,for their contribu- tion t o the economy is to reduce foreign exchange expenditure by this amount. A s explained later, foreign exchange savings are msaured at the market exchange rate in terms of danestit currency since domestic currency is the numeraire of the UNIDO nsthod. This sum is later adjusted for the shadow value of foreign currency (see pag- 33-36) It may seem reasonable to ask a t this point i f the shadow price of the care uhich substitute for imports should not consist of two elements, the foreign exchange savings and the additional domestic willingness t o pay. I f , for ex=-ple, the market clearing price (glven the import quota) is $3,000 per vehicle and ths c .i.f. price is $Z,000 per vehicle, s h d d the project be ~ r e d i t a ~ o n with l y the $2,000 per unit of foreign exchange savings or should it also receive credit f o r the additional $1,000 of value assigned the goad by cmsumers given exLsting trade pol!ic.ie&? * The wswer t o t M plausible question is developed ki detail i n Annex C w i t h a graphical derivation of valuation of production replacement and consumption expansion. Briefly, houever, since the product is already sold domestically, there is no change i n the t o t a l consumption of the good and therefore in the consumer willingness t o pay. The oni;.. change is in the supply cost, and the net difference between domestic and import supply costs is brought into the calculatim by adding as a benefit the import supply cost no lcnger incurred. The domestic . I production supply cast of these cars has, of course, already been included as a cogt. - The approach is, therefore, identical t o that used for domesti8 production replacement; i n both cases, the cost - e from the replace&source of supply is added as a benefit. - Shadow Pricing Specific Resources I n t h i s section, the many possible variations i n impact which a project may have on resource availabilities will not be discussed i n detail, but reference w i l l be made t o the considerations set out a b o ~ eunder the '!general principles of resource pricing" ( .2 . Likewise, when questions arise regarding internal imperfections i n the d a e s t i c market, reference w i l l be made t o the llfour conditionsn for use of market prices as shadow prices listed cjn pages 15-16. The mechanical adjusbeak '&antarket..to shadow prices in the standard analytical tables should be handled as indicated a t the end of this section. 7 < fa) Tradable Inputs ;o.&M&f@u'%s~; &qnapeate smmm TI!&' of the shadow prices of t h e s e . r ~ ~ a n r ~ f & i "lie readily apparent d from the general principles outlined above. Unlike the # Manual which recanmends that llgoods uhich would be exported or imported i f the country had followed policies which resuited in an optimal industrial d e ~ e l o p n e n tbe ~ ~treated as traded goods (OED, p. 92), the Guidelines tend t o predict that tradable goods which are not freely traded today will not be freely traded in the future and should be treated as non-tradable goods, measuring their value as discussed . I below. Although thxs philosophy is implicit i n many places in the Guidelines, it is stated m o ~ tconcisely by orre of the authors, Partha - -. Dasgupta, i n another publication: ll&he Office of tl$ Central Project %iven EvtSluat3g is t o accept the state of the economy as - and it is 8 t o *accept as given the policies of the goverrment (includlnq any influence that the OCPE may have)"l/. - Existing trade policy i n most 1/ Partha Dasgupta. "A Comparative Analysis of the UN3DO Ouidelin? s and the CBCD Hanual," I 3 0 ~ 3 4 1 1(Feb. 19?2), p. 36. countries means that s m e goods are fully traded, other are part- ially, and the remainder are non-traded. If the impact of any increased consumption will result in more imports or fewer exports, or i f arry increased production w i l l result i n more exports or fewer imports, a l l other things being equel, the good is fully traded. For such g ~ & there is no change, for exam- ple, in the level of domestic production i f domestic demand increases -- all of the additional demand w i l l be supplied from abroad a t con'stant costs. Conversely, any additional domestic ~q)uoti,on, sud< assihat fzba tbe .. project, will be Psrported and -dl,b.not effect deiestT'cf*c&'$lmption. For such fully traded goods, the shadow price is the border price, converted t o domestic curracy a t the market exchange rate. cu he adjustment for the additional benefit of saving foreign exchange by import substitution or export promotion -- or the converse for using imports or Psrpor+vablesas inputs -- is baeed on the procedure fo,r %cording the Foreign Exchange Impact,I1 as described on pages 9 t o 1'2 above, and the shadow price is then applied as described below on A good that is llfreelyl' traded is not necessarily unextcumbered by import or export taxes. It is only necessary that danestic changes in demand and supply affect only the level of imports or exports. However, rn the Impact of a project's demand for Gputa w i l l not f a l l solely on trade ------ (because the market price will thee &d domestic production w i l l - - expand if: (a) the goods are presently subject t o an import or export ., quota which is fully used uld will not be expanded, (b) the import or export supply is l e s s than psrfectly elastic, implying a rising cost and decreased demand, (c) there is excess capacity i n the domestic industry wuch is not used only s e ~ a u s eof insufficient demand / but which could produce additional output a t the same marginal cost as the present production with additional demand, (d) the additional d d ? - c c w s .inland &ere. tile '*ansport- ation costs from $he-port of entry raise 'the cost of imported yaods above tk.a marginal cost of local produc- tion, (conversely for the production of exportables), or (e) the import price of the good exceeds, or the export price is less than, the damestic marginal cost of production Y. In the absence of these conditions, th.e goxl is ''fully traded" and additional demand, even for a good subject t o say a 40 percent tariff, would fall entirely on ex-ternal trade because, despite the t a r i f f , the prodac* can Le obtained more cheaply by importing it than bjr producing it domestically a t higher marginal cost.8. The converse caa - t+a: ,.r,.+"- example presented below discusses the additional consrrmption of imports, .s:I but the same arguments apply to the additional qonsumption of exportables. ..i - 3 1 Let us assume for example that: (a) the unskilled workers in the project were previously employed i n agriculture where the marg'nal social product is Rs 15 per day and consumption per worker is Rs 20 per day (the additional Rs 5 per day being shared family incane which would remain with the family once the worker, waa employed by the proSect), 1 ' and that , ': , (b) the market wage for the project job is Rs 25 per day (the basic shadow wage will still be Re 15 per day of marginal social product fni thB 'agr'$ci\liM 18- assuming there are no QeoWb- o~%&rati&' etc. f .' The result of the project will therefore be t o increase consumption by Rs 10 per day. Of this incremental incame, a certain percentage will be spent on directly or indirectly imported goods. Lacking better information by i n c a e class, we can assume that, this percentage a t the margin will equal the average percentage of imported content of the t o t a l domestic supply of consumer goods. If, for example, the imported content were 20 percent, employment of the agricultural worker by the project would'rdsult i n a foreign exchange impact of Re 2 (RS 10 x 2%). - Therefore, i n addition t o the shadow wege rate a t this stage of (I0alysis of rL. 15, Ra 2 should be placed 6 under f o rsign exchange impad. * If the shadow price of foreign exchglge is 15 percent higher than the market. price, the Rs 2 will be multiplied by 15 percent a t the end of this stage and Rs- 0.30 per man day will be added t o the cost of labor in the project. While the adjustment for the i'oreign exchange impact 1/ of addi- - tional worker consumption is theoretically desirable, it nay be a waste of time. Its importance requires that this content be substantial and that there be r major difference between the market and shadow exchange rates and between the old marginal product and the new market wage rates. This difference w i l l be significant only in the case of unskilled labor from the agricultural sector in most cases. The i m g af.6Mej.z~. , ~ . adjcstment is further reduced by the fact th'at t h i s tm+af,liqlpq-,.. except perhaps for sane agricultural and civil wrks projects, generally constitutes a very small proportion of the value of the output, and generally relat.ively l i t t l e of the value added. Unless there is a very large' difference between the shadow and market prices of foreign exchange, this adjustment is unlikely t o constitute more than one-half of 1% of the value of output. --I The eocial cost of the additional consumption generated by the project!s employment of workers is sometimes included as one . I element i n the shadow wage rate (eg. i n the Little/Mirrlees Manual). . . However, since employment is only one of the ways Inwhich pr5jects ,* affect the savings/consumption balance in a society, this aspect - 1 of saployme8-t cart i e l e f t t o Stage Three In the U N D O methogwhere this and other impacts on the inter-temporal distribution of income is analyzed. (e) Capital bputs: The c a - i t a l costs of a project can usefully be divided into two parts for the purposes of shado~rpricing. lJ Pis impact may come as increased consqtion either of imported goads o r of axportable goods (eg. foodgrains). For the lack of better terms, l e t us c a l l them the principal and the interest. When Rs 100,000 is invested i n a project, for example, two things happen. F i r s t , Rs 100,000 of financial resources-is con- verted i n t o r e a l physical resources (principal). Second, the investcr remcves these resources from alternative pr&ctfuanses fn the mstqof the economy where they might be used either a s consumption goods"'or as c a p i t a l goods in other projects. mce invested in this project, hm- ever, these resources can yield their benefit (interest,) only over t h e . The shadow p r i c h g of capital thus presents two.prabrefns -- - ... the measurement of the value of the physicrir resources -mr se, and the measurement of the benefits which are foregone by freezing the resources as capital in t h i s ~ w t i c u l a rproject instead of using them in some other project or as cxsumption goods. The pricing of the Iprincipal1 component is exactly the same as f o r any other resource. I f it is a fully traded good, the value is its border price. If it is partially traded or non-traded, its shadar price is its mzrginal social cost ( i f theproject increases its . I danestic production), or its marginal social value measured i n terms of consumer willingness t o pay (if the project takes it away from alternative users). The labor involved i n the constructicn of the - . physical capital f a c i l i t i e s is likewise valued a c c o r w g t o the guidelines laid out above. Ths end result of t h i s phase of the exercise is a series of capital investment values at shadow prices which can be entered i n t b -secticn 2 of t h standard*cash flow table in Table Bb of Annex Be These capital cost figures, when substracted frm the gross current cash flow data in line L of this table, ulll *..* ** *-. -=aw,+ :q.M&?i,%rn*h.-- ", ..',.+", e- -- .-*-, give the net cash flow f o r each year of the project's l i f e . The second part of the capital cost -- its 'interest1 cmponent -- is the foregone productivity i n other uses. The analysis here is s t r i c t l y parallel t o that for other resources; the nlarginal social cost of capital is the cost of gener&*,capitzl resources <. 1 through savings. To the exteat that capital.Cor the F ~ f e clts:&rated fram additional savings, the nagi.n?l social cost is i price which savers must be paid t o forego an additional unit of present ccnaumption -- ' the consamption rate of Lnteres To the e- tbi$,*';ha.,i3&xi'Q' % @ taken away frm competing hvestments, its ms&,mr&saci&~.laEimlliJ-ihs social r a t e of roturn i n the l a s t acceptable investment, ie. %ha inveshent rate of interest y When sa-rings a d investnent are suboptin:al, the m a r r i n d productivity or' additional savings and investment is greater t3an the narginal social valne of additional consumption, and the invesbefit r a t e of interest is greater than the consumption rate of interest. They w i l l be equal only whsn saving is optimal 1/. IJ socialreturnoninves&erit, The which determines t h i s inveetment _,,_,r_M.pf interest, i-a fuarr$ian first of the anntirsl xarginal prc- -- - duct of the margin 2investment dollar, and second of the adcii- tional value which IZ~.Uaccrue *om the annual reinvestment of part of this margins product. Unless investmeqt is optimal, this~reinvestmentd l b e d a premium over consuniption. 2/ This discussion, which is an extension of the presentation in the original Guidelines, was suggested t o me by Prof. Stephen A. MargUn. Responsibility for any errors in its presentation here is of course my own. weighted by the share of each good in the actual marginal import b i l l . Assuming a freely competitive danestic market f o r the good, this approach to valuing foreign exchange implicitly uses the domestic wiIlingness t o pay a premiun above the "officialu price f o r foreign exchange which is needed t o ~urchasethese goods. Other adjustments are neces- sary i n the case of physically rationed goo& (whose price understates .)',.. , the consumer wiliingness t o pay). The approach proposed i n the Quidelines is appropriate i f it is known that the use of foreign exchange by the project w i l l only take foreign exchange from other users. But th'is' wfff happen only iP the effective foreign exchange price is fixed. Such is quite unlikely, for fixing t h i s price involves more than maintaining a stable exchange rate; it requires t h a t there be no change i n t a r i f f s , quvltitative restrictions, capital movement regulations o r any other protective device. This would be almost impossible,except i n the short term, for it would require depleting the nation's foreign exchange reserves. Few govements will allow this for long. Instead, given an increased demand f o r foreign exchange, they w i l l take other actions such as t o a . , devalue the domestic currency, increase trade restrictions, or to increase export subsidies. This in turn win increase the incentives - .# which e d s t f o r the danestic production of ilnport substitutes and -- -- 8 exportables t o earn foreign exchange &d dl1 reduce the incentives 0 t o produce non-tradables. Given these predictable impacts of a demand for foreign ex- change, Ft becanes important to look at the marginal social cost of .'*. -YM?&&r .. * -t -t!t- producing foreign exchange as v e l l as the marginal social value of the use of foreign exchange. The relative importanca of these two aspects u i l l o f course depend 09 the relative e l a s t i c i t i e s of supply and demand for foreign exchange. Probably the best known p o p l a r approach t o determining the cost of producing foreiga exchvlge is the Domestic Resource Cost I method proposed by Michael Bruno y. WtdW thfs,is oriented t o the cost of producing foreign exchange by a single projcct, the same principles apply when taken t o the macro level. However, t h i s becanes more complex i n that the elasticities of.aktei?natiire s & h of ftrefgn exchange suppl, become inpartant.--33- anGflfs'6i'this Gpect of calculating the shadow price of foreign exchange, which is presented i n a recent paper by Bela Balassa -2/,would be a logical sxtention of the UNIDO method. Once the social premium on foreign exchange has been determined, the n e t foreign exchange impact of the project should be multiplied by t h i s factor and the r e s u l t (which may be either positive or ~legaifve depending on whether the project is a net user or earner of fareign exchange) should then be added t o the net benefit resulting fran the foregoing calculations m the second stage of the UNIDO analysis 2/. The resu1.t is the net present aggregate consumption benefit of the y Michael Emno. "Danestic ~edDurceCc3ts and Effective Protectioni .1972). Clarification and Synthesis," Journal of Political Economy, (Jan/ Feb 2J Bela Balaasa. IZsthating the Shadow Price of Foreign Exchange i n Project A p p r a i ~ a l ,Oxford Economic Papers, forthcoming. ~ ~ 1/ It ia possible to obtain a negative result i f the doineatic currency is undervalued. However, given the structures of protecticn i n most developing countries, t h i s is most unlikely. project a t efficiency prices, that is, prices which do not take into account the distribution of income over tins or among contem- poraries. These twr, aspects are considered in the third and fourth stages of the analysis respectively. Shadow Price Manipulations Bef~regoing onto these stages, it may be helpful *".to* add , I 'I, j l a word about the manipulatim of the shadov pri'ces whfch have been derivad on the basis of the atmve analysis. The conventional appro z h is t o < r e ' ~ a l ~ ~ yeare ' by year the resources reccrded in the financ-ial statement, then discs..iat-t?ii4se.' back t o the present. There is nothing &ng with t h i s appmachr - However, when the percentage difference between market and shadow prices caa be assumed t o remain constant over timeythere is a much easier approach which avoids the need t o adjust the numbers for each year individually. The f i r s t step is tc calculate the net present value of each resource for which there is a different constant percentage distortion between the market and shadow prices. For example, if there are three classes of labor earning Rs 100, Rs 150, and Rs LC0 per day, but it is k n m that these market wages are each 208 above the = respective shadow wages, they can be treated as a group a t this stage. -! This discovmting should have already been done i n the . f financial analysis, s o provided it was done item by itm rather than on net annual values and that the financial and social discount rates Ll are t h e same, there is no need t o repeat the discomting process for This applies as well t o the premia on irivestment and income distribu- tion M c h are introduced i n Stages Three and Four below. Since the precise consumption r a t e of tnterest is not known, the OuideUnes suggest that a range of say three discount rates be used. the second stage of the analysis. The f i n d step is to multiply the net present value of each resource or group of resources by the accounting r a t i o of the social to %he market price. This Kill produce a :cries of net present v ~ l u e s h t social prices. Once the net present values of resources whose market/social price ratios change over tims have been adjusted on an annual basis and discounted present. there UFU. bed k s e t of present values which can then be added up t o deterni,na the net present value of the project. -- If the net present value is p o ~ W u e p*,;rrc@d, tan ba acceptable in pure efficiency terms. The rii'sWik* ~ # k t &'59 examined i n t h next two sections. ~ The multiplicative approach out- lined here is slightly different. from that generally used i n the Guidelines. There the marl'st price is multiplied by the percentage differmce between the shadow and market prices, ther. the result Ln added algebraically t o the market price in a separate operation. While the end results are the sams, the accounting ratio method prb- posed hers is sixnpler 1/. -.mm-=ORAL D L S ~ I W T IBZNEFITS ~ The difference between cmumption and savings is, in the- i f i n a l analysis, the difference between consuming now and consuming '? 0 1/ It has a furthe; advantege: if the r e s o u r c ~values at market prices are recorded i n a rectangular matrix such as the pro foma tables given i n Annex B, the entire matrix of values may be converted by canputer t o shadow price values by premultiplying the matrix by a diagonal matrix whose non zero elemen+& consist of the correspond- ing accountJ,trg ratios. in the a t m e , future consumption being possible when the investment yields its returns. For this reason, the consumption/savings impact of a project is often spoken of as inter-temporal income distribution. To the extent that a dollar invested today w i l l yield more-than enough a year frcan now t o repay the cost of waiting u n t i l then t o consume, savings are more valuable than ccnshptfon. Themfore, this Third Stage of the UMDO analysis has been designed b:, evaluatelthe impzct of t h e project on savings within the economy, This is done by: (a) determining which groups lose and which groups gain income because of the project, ._&- . 't I \ (b) evaluating the net impact on savings baseB on the marginal propensity to save of each of these groups, (c) placing a premium on t h e additional investment which the project induces by its impact on income distribu- tion. A 1 1 of these manipulations can be done on the basis of net present values rather than on annual values to simplify the calcul~tions. In this analysis, for every person or group which loses . - I I income, r..~otherw i l l gain; there cannot be a re-distribution of incame without losers as well as gainers. Thjs principle is the basis of the double entry structure of Table BS i n Annex B. - The Standard Income Flov table, which has been developesto assist tha measure- * merit of the distributional impacts of the projact, makes it possible t o identify clearly the campensating gains and losses fron each transaction that result from a project IJ. The approach taken in this section is very different i n form frun that taken i?lthe hidelines, but it is entirely consistert i n substance (except for llmited disagreements with the UNIDO method). It l a much more clear in its execution, and its double entry fonnat lead to more sccuracy. The only case i n which there is not a compensating loss is for the net profit of the project, which is equal t o the net social profit at shadow prices plus the additional gain or loss due t o market price distortions. This net market profit is assmed t o remain with the project as incane for the purposes of this analysis. The net prof it figures sre markad with asterisks i n Table BS. Because of the complaxity of the analysis f o r some of2tke impacts, Anoex D presents a numerical example. Therefore, only the highlights of the analysis are d e s c r h d here. In that Annex, Table E6 is reproduced as Table D l and is used to record the e~raaph*. The heading of the ~ t & d ~ d * ' h c o mPI-. ~ ~ . . s S -is.. 4 s ~ . divided into as many savings/investment groups as seems worthwtdle given the differences i n their marginal propensities t o save -l/and the amount of income which will be redistributed t o each group. The stub of this table (consisting of the t i t l e s d m the edge) has been.divided into three sections: (1) Incane flows a t shadow prices (2) Additional f l o w due t o market prices (3) Net gain ' ' The sum of the first two items is equal t o the actual flows a t market - prices. While it may look as though it would have been easier t o 1 use market prices directly out of the standard incane statement a t e market prices, #any of the flows are difficult i f not impossi'ole Throughout t h i s paper it is assmed that a l l incame which is saved w i l l be reinvested. 1 t o analyze c o r r e c t l ~unless the underlying flows a t shadow prices are f i r s t examined, and then the additi~nalflows due t o market price distort ions zre added on. The f l w s of income a t shadw prices csn be filled i n quite easily irom the Standard Income Statement expressed in shadow prices. In general, purchase by the project of an input, such as labor, re- presents a loss of incone t o the!pm*ct and a gain t o the s e t o r wfiich no longer has t o pay for t h i s input u.Likewise,, the sale of output represents a gross gain in incane by the project and a loss (expenditure) to the consumers. To the ez&mt t h e t)ra ept.p&.8prCI-(. d t". stitutes for imports, there TS'%O loss by consumers, bua d & m l s a net loss of income by the lfexternalllsector. (Since the external sector is external, ita dcmestic propensity t o save is zero and all gains and/or losses by this sector drop out of the analysis.) The net gain of income by the project a t shadow prices is the shadow value of outputs minus inputs, or the "social operating profitw of the project. In addition t o the gain or loss by various sectors at. . I The UNIDO Guidelines implicitly assume that unskilled labor is paid its shadow wage in the sector from which itZleaves and that the margin between the shadow arid market wage ncdrresponds exact- ."(p. l y t o the net extra incane received by unskilled-labor on account t 'of the project 303). Since the shadow wag for unskilled ?labor often a'voaches zero and labor cannot l i v with this wage, it is empirically obvious that the income of unskilled labor must - be higher than the shadow wage In previous employment. This aa- sumpti~nis usable, however, i f one further assumes that any difference bstween the former marginal product and actual incane was due t o transfer payments which, once the laborer is employed '.I by the project, w i l l be distributed t o others In-the same income group- shadow prices, they will also gain or lose depending on the market price distortions for inputs and outputs and on the transfer payments suchas import duties and lncome taxes which must bs paid. Since the values in this section are incremental values, they are determined by subtracting the shadow prices for the respective resources in the Standard Incane Statement expressed i n shadow prices from the same resources i n the Stzndard Income Statemnt expressed in market prices L/. I . The result may of course be either positive or negative. Since transfer payments have zero direct social cost, they are recorded entirely as additional flows a t market prices." *- In the Chidelines the 2's 'considered arr ac m a t . . the sector (private or government) t o which it belongs. It has been' s p l i t out here so that the net profits a t shadow and markst prices for the project itself can be clearly identified and used to double check the figures. So long as the same marginal propensity to save is assigned to the project and to its subsector, this modification does not alter the results, for the internal sectoral transfers of resources cancel out. However, if the project involvss foreign equity and i f some of the prof its will be repatriated, this dcdificaticn is important, for it allows the UNIDO method to take this Fnto account. The same effect could be obtained-by multiplying the shadow prices by the inverse of the accomtlng ratio minus unity and recording the result in sectim of Table B5 for incremental in- cane f l o w due to market .price distortions. he inverse of the accounting r a t i o is rsed because we are going from shadow t o market prices rather than the reverse, and unity is subtracted t o obtain the incremental difference.) P r o f i t s which are repatrieted represent a zero marginal propensity t o save as far as the country is concerned. Therefore, the overall marginal propensity t o save of the project w i l l be s wei~htedaverage of the zero savings on repatriated earfiings plus the actual marginal propensity t o save of the earnings retained dmestically. Once the net gain o r loss of incqme ),as been calculzted f c r each saving/investnent group by adding np the n e t change a t shadow prices and the additional change due t o market price distortions, the net s ~ v i n g a generated 3y these changes i n incane allocation are determined by multiplying the n e t change by the marginal, p m w s i % y t o save or'-the group. Table D2 in Annex D provides an exampla -af t h i s ~ a f c u l e t i o n . To determine the sdditional social v a l w which accrues or is debited to the project, the net present value of the total net s~vings impact of the project must be multiplied D;/the yatio of the social value of an additional unit of investment t o m additional unit of con- sumption minus unity y. This factor, called pinv or social price of investment i n the Guidelines, is calculated by the formula: - MPcaP pinv = (1 s)q = MPC (HPC~P) i eq -- CRI x MPS . I The secmd formulation of the equation is identical t o the f i r s t except t h a t mnemonic abbreviations instead of symbols have been used - ,9 - t o make the equation more intuitively obvious. The variables are 0 Unity is substracted fran this accounting r a t i o to ~ b t a i nt h e premium on investment. This premium shoula be applied t o the net present value of the net savings impact only if it is assuned t o remain constant over time. Otherwise the annual values of the premium must be applied t o the annual net savings impact numbers. MPC = 1 s = naginal propensity to consume - ?fF'S = s = marginal propensity to save lQC ap = q = marginal productivity of capital CRI = i = consumption rate of interest IJ This is a simplified equation for the social price of invest- ment which is valid only if . certain assmptions are made: (a) the time u n t i l saving becomes. optimal and the pre- mium becomes zero is iufbLtafg, far. in tha,futwa, (b) the marginal propensity to save, the marginal produc- t i v i t y of capital, and the consumption rate of inter- e s t are all constant over /.ifme. Oranting these assumptias, thts$czi%@~ I*& 'heffect that t k e ne.t present value of an additional unit of investment expressed in terms of its consumpti~nequivalent is equal t o the annual consumption benefib out of the marginal product of capital (MPC x wcap), dis- counted t o the present by the ccnsumption rate of interest minus the r a t e at which the iaarginal product of capital is reinvested and generates additional flows of consumption benefits (cRI - MPS x wcaqye . I Once the additional social value of net impact of the project on consmption and investment h e been calculated, this in- - c r m e n t is added t o the Sacond Stage ap$oxbation of the net - * - 1/ Since the CRI is treated as an unkncdh (see page 36,footnote l ) , and since the i s a function of the &I, the pUV w i l l vary eccording t o the disco-imt-rate used. It may therefore be appropriate to apply more than one PInv i n the Stage Three analysis. 1/ For a more detailed explanation of t h i s relationship and the formulas which are required i f tha above assumptions do not hold, see Colin Bruce and John Re Hansem, "A Chide for Country Zconomists t o the Derivation of National Parameters for Economic Project Evaluation," Washington: IaRD ( ~ s i aProjects ~ e ~ a r t n e n t 1973, pp. 12-17. ) , his paper is not a stztement of official Bank policy. ) aggregate consumption benefits of tne project a t shadow prices t o obtain the third stage approximation. INTER-PERSONAL DISTRIBUTION B?INSITS If the only objective of the government planners in the selection of projects is t o maximfze!the net present value of the 43 ~ I' .I net aggregate consumption benefits at'lhsdop'prifes, takhg &to consideration the Impact of the project on the national s a w objectives, the above stage of the UNIDO analysis would be the final one. However, i n many countries the redistribution a? m,b~ - . qa@fic regicne or income classes within the country i s of high priority in the selection of projects. It then becomes necessary to evaluate the impact of a l l projects on this objective and t o give the? an additional credit depending on the importance placed on the objective and on the degree t o which they help accamplieh it y. The analyeis i n this stage is made relatively easy by the measurement of the project's impact on income flows which was done i n the Third Stage. bovided that the target group8 for income distribution have . * been identified separately i n the Standard Incane Flow table (~abl-eDl), we already knou the net gain or l o w of incatie by the group from line 3. - of t h i s table. If the analyst wis&a t o evaluate the income distribu- C tion impact on a regional baeie, t$s- table should be brcken down m further by region eo that individuals i n each irrcme class who are 3/ It is necessary to include allyrojects,for if a premium were only added t o selected projects, their net present value could be raised above that of projects which were even better but f o r which the analysis for the accomplishment of other objectives was not done. located in the target regicn caii be identified. The treatment by region of payments end receipts for re- sources including the output, material inputs, labor and capital requires careful thought regarding the source and destination of the resources involvee and the use t o uhfch these resources would have been put i n the absence of the project. For example, when unemployed labor is employed by the project, the wage payments represent a net gain t o both the region and the low income group y;there is a & --. . ' I., ( expansion of employment. - -,. . This is also true even f o r employed skilled labor r f one assumes, as is done i n the hidelines, that the skilled labor will be replaced by immigration from other regions, resulting in a net in- crease i n regional incone. However, i f materials are purchased by the project locally, it may be assumed that, unless there is excess capacity i n the region for the productim of these materials (in which case payments for these materials would represent a net gain for the region), it will be necessary for the region to %nportwreplacement materials from r other regions, resulting i n a balancing decline i n resources available t o the region. For some resource flows such as payments t o expatriata - - workers, it is necessary t o measure the percentage of their salaries 3 which w i l l be respen%within the region. The Guidelines suggest that only the margin between the prior wage and the project wage bc, counted as a net gain, but i f one is t o be consistent with their suggestion that skilled labor wages be counted in full.on the basis that other workers will replace those a' rorbed by the factory, unskilled wage8 should elso be counted in f u l . After the net impact on the region and/or group has been calculated, the treatment of each is slightly different. In the case of a regional incane distribution objective, it is assumed that some percentage of the i n i t i a l distribution of income t o the region will be resyent within the region, fielding further raunds of increased hc me. The initial redistributSo& t o the 'region must be multiplied by l/(l-r), where "rUis the percentage of inccme respent i n the region, t o obtain the net aggregate income gain. In the caee of incane redistribution to an income claee-, tRBlZ 'isT'earfbn asewCcthat +a, 1 relatively l i t t l e of insome w i l l be- respbnt-%Zi*,Wn--thi_s -class,- so no - - - multiplier adjustmat is made. Chce the net aggregate incme gain has been calculated, two approaches t o adjusting for this distribution impact are possible . One demonstrates the logic of the Stage Four adjustment, and the other is much easier t o use. In the Stage Three calcnlation, we have ignored the fact that, from a u t i l i t y point of view, income gohg t o some regions and sane ., income groups is more valuable than that going to others. I* nevelv theless added them all together as though they were equal. The logical step would thua be t o go back to this stage and weight the income of each group separately before adding up the net benefits. If the marginal 8 u t i l i t y of Fncane a t the average level of consumption in the llaveragett region is used as the numeraire or point of reference, benefits going t o regions and income classes below this average would receive a weight greater than one, while those above would be weighted by a factor less than one. While the lo@ back of bhis approach is obvious, this method would involve a l o t of extra work a t this stage. It is far easier to simply add to the results of the Third Stage the differences i n value between the unweighted and weighted net incane gains or losses. This premium w i l l be equal to the ratio of the marginal u t i l i t y of incane going to the target group or region i ' 1 t o the value of the marginal utility;& incame,l( going-to the, atferage group I' , . . I . The marginal u t i l i t y of income is a direct function bf the elasticity of lJ the marginal u t i l i t y of income, tM is,far exrmple;.Wparcantage by which the margiaal u t i l i t y of income declines with a one percent increase - i n income i&el. This value is the link $.cwe.en three important aspects of economic project evaluation @til&&';~"alvl'- &' utyjs!%h@~$c?? inter-personal incane distribution an# fm inter-temporal savfnger rs, consumption) income distributiollr As mentioned above (p. 32), the consumption rate of interest, the discount rate used i n the UNIDO method, is the price which savers must be paid t a forego an additional unit of present consw~ption. This price is i n turn a function of the rate of growth of per capita incane, the marginal elasticity of utfiity of income and the rate of pure time prefer- ence. The assumption here is that, in addition to a pure psychological preference for consuming now instead of later, since incme levels w i l l bo higher i n the future, additional future income will be sanewhat less valuable than the same mount received today because of a negative elas- t i c i t y of marginal utility, CKving a higher weight for income going to the poor than to the rich is based on a similar. assumption. Because of a negative elasticity of marginal incame u t i l i t y (the marginal u t i l i t y of incme declines with increases i n income), the marginal u t i l i t y of an additional unit of income i n the hands of a rich man is less valuable than the same additional incane in the hands of a poor man. ' ' Finally, the premim on investment over consumption discussed above (the intertemporal inccme distribution factor) is a function of the consumption rate of interest, among other things, and this is, of course, in turn a functio# of the :elasticity of marginal inccme utillty (p. 42). While these'relation'ships are quite clear from a theoretical point of view, the entire analysis is made difficult by the fact that it is ' virtually impossi&e t o derive empirically the-elasticity of marginal incoae utility. kt can only be done on a deductive basis. Ole approach is to evaluate the incme tax structum of the country. This is based on the assumption that the progresnive marginal rates of taxation reflect inversely the marginal u t i l i t y aeaigned by the governmen* acting on behalf of the society to increasing levels of incme. However, the other assmptions required by this method are so widely violated that the nbottom-uptl approach suggested in the (hridelines and described below seems much more satisfactory, despite its obvious limitations. in th3 average region -- that is, the accounting ratio for incame of the ithgroup -- minus one. (Unity is subtracted from this ratio so that we are l e f t with only the prenium.) The premium w i l l of course be positive for incme going t o the ltpoorlland negative for income going t o the ''rich." When multiplied by the net gain or loss of i n c m in tine respective incame group or region, this premium w i l l give the net present value which should be ad&d'to:or subtracted *om the net I . present value in the Third Stage of andysis t o give the fourth a d final estimation of the net present value of the project. The deternhation of the w & ~ J B 'fci a&p- ar incnmr t o different groups is obviously not a o*im&H??g eco.?odc prueewz7 v k . b depends very heavily on political judgement. The UNIDO Guidelines suggest that the actual determination cf the weights should be an interactive process between the planners %t the bottom"and the leaders "at the top: Ciiven the preferences revealed by the leaders for projects with different net present values and distributional impacts, the plan- ners can determine the ueighte implicitly assigned by these leaders when they accept sane projects and reject others. Assume for Bxample that the planner8 f i r s t put up a project - with a negative net present value of R s 1,000,000 as calculated i n the Third Stage of the UNmO analysis (ie. before considegtion is given - t o income distribution effects). Rro income groups haPe been identified-- s r people with incomes less than R s 100 per month and theorest of the economy. The a n a l p i s of Stage Four shows that the project over its life would increase the income of the low income group by Rs 5,000,000 in net present - --- --- - - lJ This approach depends on the ass\mption that this prcrmium remains conatant over time. If It does not, its annual valae must be applied t o the income distribution in each year. I;, . 1 -A..:d'. . , . value tern. If the planners accept the project, this implies that they place a premium of a t least 20 percent on income going to the poor, for 20 percent of Rs 5,000,000 is Re 1,000,000 which, when added to the net present value of minus Rs 1,000,000, will bring the net present value of the project t o zero, the marginally acceptable level. If the plan- ners reject this project, but accept athers requiring at least a 15 per- cent premium on income going t o the poor, the pmject analysts know that the premiumlies sanewhere between 15 and 20 percent. Provided the policj makers are ccnsistent over time and that a l l other non- quantified factors are equal, a usaEe -v&@j-Ed '60 fjilacebop bmxe distribution can be derived by this proc;essi - 1 . THE UNIDO AND LIT-ES XETHODS CCCIPARg The l4imual of Industrial Project Analysis in Developing Countries, which was written by Professors Ian M. D. U t t l e and Jams A. firrlees and published in 1968 by the Orgadzation for Economic Cooperation and Development (OXCIS),, l.8 today probably the most widely ) ;* , discussed netSod of econcmic p r o m mahattan. The Hanuai com- monly refemed t o as the nOECDll or LiLttlehSimlees" nethod, has obtained this position, apart from being a highly professional work, bcause it was the Pkst, and oMf-tke ~+?S%d%-&~t~s Qt-h - UhW lines the only comprehensive rn- -9 &eamW&'k g d j i c t evaluation available in print. While most of its carponent dements reflect traditional tools of econondc a a l y s i s , they were brought together i n a way and for a purpose which were unique. Acid0 *om the very waak "partial indicatorMapproaches such aa capital/labor ra+;ios and foreigln exchange savings, the only other published and widely recognized methods of prodect bvaluation desigied specifically for the developing cauntries are the Effective Rats of Protectich and the Damestic ~eaourceCoat (Bruno ~ a t i o ) methods IJ.These, however, have been described 3n the litcrature - inietatic (one-yeax) t-enus, and while they csn easil-j be adapted t o - - 6 tiqa discounting, they cannot, in their traditional fom, take into * I ammt the t u of the costs md b e f i t . over th. life of a project= 1/ For a good though somewhat technical summary of these methods, se3 Hichael b m o , domestic Reso19rce Costa and Bffective Protection, ~ l e iicatim-and Synthesis f , Journal of Polltical Bconany, ~ 4 F e b . 1972 A N N m t A Page 2 of 11 Devolopent of the L/M and UNIDO Methods - .- . . Since the Little/Mirrlees (I&) method has been i n print for about four years longer than the UNDO method, most readers of this Vuida to 'Lhe GuidelinesNare probab1,f more familiar uith the fonnar. Also, they may wocder what the differences are L'9twee.n the two .nethods that would justify bringing out another book on the same MpFc. This annex is desigped, therefore, t o summarizeithe major diffsrences (and similarities) between the UNUX) Ouidelines ana the ZJW Manual as originally published. A l l remarks made below the NLSttle/Hirrlees method" refeF-"&pBzia &a-$be.-o~b&'i&---&xt:A 110 ~ttempt'Is made t o record the -ma+&gage~3...-wfiich have W x a d in - . the method i n acdemic publications and in the N o r atraditianff surrounding the method since the time of it8 ori,*al pblication. These changes, which are the result of two or three years of intensive and often i k u i t i . debates, to which both academics and practitioners (and academic practitioners) have contributed, have brought the UNIDO and %4¶ methods very close together in terms of approach and ermphasis and !lave eliminated any basic differences of substance. Tha ? v i t a of these discussions w i l l be reflectediby Professors Little and -lees i n a revised version of their Hanual which is Ale for publica- - - - tion i n the near future y. .9 L y Tn the meantime, t&e moat useful single refermce source on these = . discussions 18 the,Symposium on the U t t l c ~ l e e Manual of s e Btdurrtzial Pro3ect Anblysie i n Bveloping Countries. The papers frau this Sympoeim have been reproduced in the February 1972 --.- Bullstir (381) cf the Oxford University Institute of Econaulcs and 3tatietice. Another nseful zet of pepers was generated by the DNIDO/Inbr-American Development Eank conference r?n the UNlDO method held in Waskington on Hsreh 28 t o 30. The papers from this conference have not, as of this writing, been compiled into a single volume^, however. Deepah- Lal has demonstrated uith - A in hi^ ~ a l ~ankstaff ~ o z f i g d Paper NO. 15eatentitled technical elegance H*ltamatiro A.oj.at Selection Procodma for Drrvlopln~Comtriua Mticd. -7 T@m*' (Vurbb@a, m,197)) that, if tb 8- r s ~ i a m rde, both metL are w l l l produce the same ramilts. It must be stated at the outset of this annex that mast of the differences listed blow, evm- betwean the original texts of the two methods, are primarily mattars of rr,ec~ianics,and perhaps emphasis, but not of substance. Both works are eased on the same body of econmic knowledge and were writzen by highly canpetent economitlts; it would therefore be surprising if tne mothods actually contra- dicted each other on s u b s t a n 6 p o h t s. However, amphasis is very - important t o the reader's Wjr exampIe, the UNDO Chide- ILrles place greater stress cm items such as income bfstrfbution, 7 externalities, and iterative p 3 ~ b p ; m b i c ~ *.B $.ithem a difzrr-b v 6 ii overall tone. The Hanual c-8 - I I t&per .,but often with onIg.,a f o ~ k - note, or at most a few paragraphs. Conversely, the Manual places more empk~asison the benefits of liberal trade policias than do the Quidellnes. Suamary of Rifferences Host of the mechanical differences between the 3wo methods I%-: .ts to the nuneraire used. Although it is customary t o explain -r: ~~onceptof nuneraire i n simple terms such as apples and oranges, . I a r.mualre i n econanic evaluation i n really a multi-faceted unit of account d t h many specifications. Aiior,a the qualities which must be specified are: (a) the unit of currency used t o measure the p n e f i t (local o r foreign currency), ; (b) the relative price/value system used (domestic ' or border), (c) the use which will be made of the benefit (con- sumption or investment), (d) who- gets the benefit (rich o r poor, government o r private) ,and ( 0 )the point in time when the benef'it is received (now or in the future). The list could be extended t o greater levels of detail, but 5 .. :. these are the primiuy distinctions presently made :among bezmfits. Chce a l l the benefits hm6 h e n denaninated in term of the same nuneraire, they can be added up t o find the net benefit, but not before. The WID0 numeraire i; Itnet -curmm@Pba benefits in the hands of the average person ih. e-p~I&+a%e &t%li"ra t bcrder prices measured in damestic currency.I1 Zhe Manual numeralre, on the other hand, is "net present investment benefits in the handa of the government measured in domestic currency at border prices." The reader will not find either of these definitions in the respective books, but only because a shorthand ~ o l o g has y been used. It should be noted that same of the distinctions listed above may be taken into conaideration, even though they are not specified e l p u c i t l y in th.nunarai;e; though the m e of Itequivalent values.n For example, the distinc+,ionbetween income i n the handa - i of the rich and incame,@ the hands of the poor may be dieguised in the final value by &inverting inccrae i n the hands of each class a irrto "equivalt3nt"nccm~ in the hands of a person uith the average level of incane, then ~onvertingthis private incame valueinto the e q u l r ~valru of incam in the h d r of tha gorerawnt. t ANNEX A '&zrof 11 Currency Unit and Price Relatives It is commonly thought, since the LIttls/Mirrlees Manual used "border prices,fl that the resulta are expressed i n terms of foreign exchange (e.g. US dollars). In fact, of more than 20 Little/ Mirrlees project evaluation case studies which this author has reviewed, only one used a foreign currency as the unit of account. All of the rest used danestic curreucg,mrfts. The UNDO method expficitly uses domestic currency, so there is no difference between the methods on t M s point. When the term Itborder pricvt1lla la&=ItJpspeaker is really saying that the relative prices at the border serve as the basfa fiw -*- value measurements. Thus, the tern "border pricesn is really a short- hand term for "shadow price,It not an indication of the currency unit which is being us&. Border pricas are used in both methods when the opportunity cost of producing a good is the cost of this good at the border. For example, a car might have a dmestic market value of Rs 100,000 per unit, while its domestic border price relative (the import price i n foreign exchange converted t o domestic currency a t the market exchange rate) is only Ra 75,000, reflecting a 33%rate ., Ihe Uttle/Mirrlees Manual appears t o go one step beyond - the Mdellnes when it reconmudo that non-traded good be brbkem - -- I d m into m d e d goods, unskilled labor and residuals. . The daded 0 goods sire t o be valued a t border prices, while thtr residuals are to be kokan dawn and revaiued further by the Bane process. In reality, the same process i s implicit in the UNDO method which recornends that, uhen the w e of a resource creates a dmand for additional supplies, +he social opportunity cost of supply is t o be Cetelmined by looking a t the social costs of the component parts of the good. Although a draft version of the Widelines indicated that consumer willingness t o pay (and therefore domestic market prices ) was the f basis for valuation, the approactriId the'pallshed version is in reality very similar to that of th.OECDamethadLsh t agatn'it should be emphasized that border prices refer not t o the' currency of account, but t o the source of the social opportunity cost, - M ~ ~ Y G S & % ~ ' ~ . O . - --- IR * 2 ~ ' . methods use border prices for tradable#, We l X t t l e & s ~ a , - ,. I W c a t e s that, when quantitatgi restrictions prevcil, the dcanestic wF11Fngness t o pay should be used as the measure of value, especially for scarce inputs. Consumption versus Bvestment Investment i n one of the # numeraire specifications, while canmption is used in the UNIDO numeraire. Both methods recognize that investment is more valuable than cor:umptian in e c o n d o s with suboptimal shvinga rates, and both calculate this pramium with very sindlar formulas. a The major difference is that, in the UNIDO method, a premium i a added t o invested benefits, while in the Iittle/Mirrlees method consuned benefits are discounted by a conversion factor equal t o unity minus h e inverse of the 8 premim on investment plus one. For example, if R s 75 Lf investumnt is worth Ra 100 of consrmaption, UNDO adds a premium of 33%t o investment bsnefits t o express them in tenns of their consumption equivalent, -&la the t/n mthod rrmltipliea the R8 100 of consuption benefits by a con- ' -1.0.33 version factor of 0.75, or i n the Uttle/)lirrlees - Manual, this is only implicit, aild to qy knowledge, is brought out only i n the shadow wage rate. This limited use of the premiurn on investment is due t o the fact that the project is considered to be i n the public sector uith all of the net benefita going to government. In this situation, wages i f . to workers would be the only signfffcaatlsource of consumption benefits. i Rich versus Fbor One of the outstanding features of the UNIDO method is that it brought explicitly into the economic project e v a l u a f ; ~ - p m c e # ~ ~ + : ~ i d e r - ation i n a quantitative form of the income dibtifiut3s~- projecta. Thia is done i n a rather primitive way, but it is probably much m r e practical than other more sophisticated approaches would be,given the shortage of necessary information. This approach becomes particularly meaningful i n the context of the iterative planning approach as described below which is favored by the Guidelines. The Little/birrlees Manual admits the existence of income distribution impacts, but does not give them nearly the same emphasis, nor does it supply a particularly useful means of examining this issue from a quntatative vieupoint Y. Private versus Government Neither of thesemethods goes into much detail about the relative 9 value of government a& private incorm, although both mention the problem. - - The present L i t t l e q l e e a Manual Implicitly places zero value on profit going to the F;-ivate entrepreneurs, while placing a unitary weight on that going to the government. This, however, is more a consequence of the way t b method was worked out than of a considered judgement regarding the i- Ihis situation will be reaedied in the forthcoming &sion of the Manual. - ANNEX A Page 8 of 11 Ukely relative values of government and non-government income. It appears that this Kill be m r e thoroughly workbd out i n the now version of the Manual. The WID0 method mentions the value of govermnt income, but p r i m r l l y i n t e r n of the marginal propensity to save and the social value of investment, rather than i n t e k of relative marginal u t i l i t y of ., income i n the hands of the two sectom. A suggestion for t k : h n s i o n of the UNDO method into this area is given i n Annex D of this paps. ?resent versus Future Benefits s o . , Mscounting is standard farqc

atborder prices, This, i f the appropriate assumptions are nade in assigning values,is the excess of line 7. from Table B1 i n market prices over border prices divided by 7. from B1 i n border prices. The domestic resource cost calculation could be calculated on a variant of Table a. ANNEX B Page 3 of () In choosing the major categories which are llsted i n the following standard analytical tables, it was decided to keep the tables relatively short so that they could serve as a su~l~nary which could be reviewed quickly to get an overall picture of both the economic and finan- cial structure of the project. However, since a summary does not usually provide sufficient information for those,.,~tuallyvorldng wtth the,project, . . 9 a flexLble codified numbering system has b.fntrqdaced,toallau,&re detailed ancillary tables t o be integrated into the doclmentatian.system. The codified nuhering system is based on a mdified decimal code and allows arly desired number of 'slili&f&&h a t aqy 1eml"through t ? m ~ means of a vary simple punctuation syst~s: l%ge'for example the section VQerating Expensest8i n Tgble B1. It is designated (8.) and the following section "Operating Profitn is designated (9.). The first subcategory of "Operating . , -1lr9ss~~ which is l'Invisiblesll becomes (8.1 ) while the second subcategory is (8.2. ). llInvisibleswis broken down further into lVtilitiesll (8.1 .1 /),tlWansport!' (8.1 .2/) and so forth. If sdditional information was required to docuwnt the data on flTransportU,for example, it could be entered i n a separate table runbered 8.1 .2/ and this table would have lines labeled /I. ,/2. and so forth, Each of these could be furtlrer subdivided into /l.I., /1 .2 and so on. The f_ull - designation of line /I .a would of murbe be 8.1.2/1.2, which makes it 4ry - easg to trace the datiback to its origins from the very swmary f i g u r 6 . for Wperating Expensest1in Table Bl . w Although it is usually unnecessary to continue this process of subcategorization much further, it can be done according t o the following general schema: x,x.x/x.x.x//x.x.x///x.x, etc. The &sr of slashes between each s e t of th-ee digits clearly identifies the level of disaggregation at which one is operating. Since this same codified numerical d a r i n g system :s used ir. a n of the tables, it might be well to attach the following prefixes to the nuxnber for aqr auxiliary tables giving supporting details: SIS Standard Income Statement SBS Standard Balance Sheet Sm Standard Cash F'bb . .. SIF Standard income' Flaw The income statement table on trlnsport costs mentioned above would therefore be labeled SIS 8.1.2/. Also, since the samo tables uc:used far- &+A.=get and social prices, it would be well to in&cate,#rs" fanner with the suffix '13" and the l a t t e r with the suffix "SU. Theecor-onric or social price version of the above table would therefore become SIS 8.1.2/ S. 2E?!Ez The Integrated Ibcumentation System presented here provides a means of placing i n a readily urderstood format a l l of the i;lfonnation which i a necessary f o r the financial and economic evaluation of projects. Because the nurrbering system allows one easily to find further details 'on aqr given b i t of data, 'summary tables can be prepared to give the bFoad pi:ture without sacr:',ficing usable detail. Because the same fonnat is used for both market and social prices, the important market price dis- tortions cars readily be identified. Since the farmat 'can be modified withott losing aqy of its organizational precision to Include special categorizations of data (e.g. domestic and foreign exchange impacts zs i n Table &) the IDS provldes a means of spplyifig various methods of , econonric project evaluation t o the sams s e t of data. IZ-$2-."~ , - * t SHADOW PRICING - This annex demonstrates graphically some of the economic 1/ principles whlch i i e behind the UNTW amroach t o shadow pricing.- The discussion here is develo~edi n t e r m of shadow peeing the production of importables; M a done, &re is a biiaf .*rcussion . , . I . , '. ai' ths shadow pricing of other wads and of the .constmrptfm of a l l resources used as i n ~ u t s . ... . . The best place r.o lxegia -,with t h s list given In the main I , t e x t of the impacts which the production of a pmfeca rag h m cm the I r e s t of.the aconolqy. These M ~ be to: V (a) increase the supply avatlabla t o the r e s t of the economy (domestic market sxmn~ion); (b) decrease the production of other users (domestic productlion replacement); or to (c) decree-se imports or increase exports (import substStution/ export expansion). Each of these impacts may be snown ~ceparatelgi n papiua:c,L . I tens, provlded certain special assxnpticmregarding the ahapa of the supply and denand cum-es are made so that only one of these impactt 9 - w i l l occur a t a t h e . This i e done i n Figures C-l t o ~ f k . - The varlous * effect8 are brou&t together i n figure $5, which r e f l e a s a aore complex - b n t more r e a l i s t i c s k u ,,tion. -. I/ I am indebted to Max Corden for co-~ectingan error of mine in an earlier draft and fcr suggesting +J.e approach used i n W e Annex. .A more or l e s s complete paphical exposition of shadow ricing given by h a s t o Fontaine i n "Social P r i ~ i n g ,Chapter ~ E i $8 n OAS ;. Course i n R o j a c t Evaluatioa and Investment Rograuaning In m i d a d and Tobago, W a q t o n , OAS, 1973 PNNEX C Page 2 of 22 R re Market Expansion: In this situation, the output of a project adds t o the t o t a l supply available and t o t h e t o t a l size of t h e market KLth~utchanging the price. This is possible only with perfectly e l a s t i c demand. I n F i g ~ r eC-1,the project expands the market sup?ly from S1 S1 t o S S by adding an amount AB to t h e , 2 existing supply curve S1 S1. There ts,a perfectly elastic demand for tha output, as represented by OD, so the market expands irom t o Q2, but the price remains fixed a t PI. Since demand is perp ztly elastic, there i r no consumerls surplus md-L\w&.cr@no change i n it. However, .- < there is an Increase i n producer.!,c .surplus equal t o the shsaect area ABC. The consumer willingness t o pay is represented by the t a t a l shaded m e a under t h e demand cume. ANNEX C Page 3 cjf22 Pure Supply Replacement: In t h i s situation, the ou?:;ut of tha project does not increase the size of the market but instead reduces tha prevailing market price and, i n the process, replaces marginal production which takes place a t costs above the new market price. A drop in price with no change i n t h e volume of sales is possible(if there is no simultaneous change i n the demand curve) only with a perfectly inelastic demad'curve such as IlD i n Figure C-2. R i c e Q1 Quantity In this figure, the new project produces AB of new product, shifting the old supply curve S1 S1 outwards t o S2 S2. Since demand is perfectly inelastic, the total market remains a t Q1, thus replacing Q2 -Q1 of ANNEX C Page-Lof 22 the production of efisting producers. The consequent drop i n price from P2 t o P1 remits i n an increase i n conPe er surplus equal t u the shaded area ACB. The coot of the production replaced is equal t o the total shaded area under the supply curve S1 S1. Pure Import Replacement: In t h i s case, new domestic production, represented i n Figure C-3 by SD SD, replaces Q1( P ~ B ) of the t o t a l market supply Q2 uhfchlorfgfnaUy was inparted from a perfectly elaotio import supply source r e p r e s e ~ t e dby. S1 Sf. This reduces imports t o the quantity Q2 Q;, ..Tha.&w&raxe DD and .*:. *-. . i the market clearing p~.ice P1 remain iulchanaed, Thia.iarp,ort, replacement .- process creates a producer's surplus equal t o the area ABP1, b u t leaves consumer surplx:, unchanged. Price '4 Q2 Quantity Fime C-3 Before bringing together all of these impacts i n e single diagram, it may ? a helpful t o extend the analysis of this section t o examine the graphical representation of trade restrictions, which are included i n the motor vehicle example developed i n the main tzxt. First, the free trade situation i s depicted i n Figure C-h by t h e import supply curve ql % and the demand curve ID. In t h i s .d; ' , I situation, Q of t h e product is sold at prfce PI. Next, dtmrestic production begins, as represented by SD SD. J u s t as i n Figure C-3, domestic supply replaces some imports (Q;),, Fr-, . *A,ha?a.J.a-~a. ' -,*-- - * A,; change i n market price, t o t a l volume o f sblbs, .or c o e a ~ ~ , q q l u s . ._.I -. ,. Quantity Finally, a specific import tax equal t o P2 - P1 (=AC) i j imposed. This shlfts the supply curve upwards t o % The representation would have been the same i f the tax had been an - ad valoren duty equal t o (p2 - pI)/pl percent. The quota equivalens of the duty shown i n Figure C-4 Is AD; if imports are restricted.,t;o t h i s amount (q-92)) the domestic market cZearing price w i l l risesLC P2 as i n the case of an import tax equal t o P2-PI. rS +h*qqmta*it;s- equal t o or Larger than BE (-94 - Q ~ ) the free Z rade 3aqw-t .e~~our.t+ , - . the price w f f l remain a t the c.i.f. level. Now that the individual canponents of the impact of a project have been examined, they can be brought together i n a sin* graph based on the example of a project for the production of 15,- motor cars which wau given cn page 4 of this piiper. Of these v-2.. 5,000 would replace prodllction from less efficient danestic p r o d u c e . 5,000 would replace iaports, and 5,000 would expand consumption. . m c - the project, the *port quota was 20,000 vehicles; afterwards, it-a- to be reduced t o 15,000 vehiclee. -2/ -1/ %Is device of shifting the supply (or demand) curve can be e-J: *o - any situation i n which the market price and the social o p m & cost (shadow price) are d l f f e r 8 n t . b ~assuming that the d i f f e r e ; i r equivalent to a positive o r negative tax. . Import duties could have been used instead of quotas. The 0nlgaG-z change would be the allocation of producer surplus t o the gove-1 in the fonn of tax revenues instead of t o the ir~~porterai n t h e of quota pmflta. ANNEX C -of 22 This situation is depicted i n Figure C-5. The domestic supply prior to the project w a s SS. Given the market demand and imports, domestic production was 20,000 vehicle3 a t a price of $3,000. An additional 20,000 vehicles were i q o r t e d under quota, as indicated by the Ilne segment'AB,,bringfqg, t a ~ . c o n ~ too n b0,000 units. The import supply is infinite* elastic t o the country, as indicated by a %. The introduction of the project w i l l shift - . > the domestic supply cwve t o the righ*, u~.~~+.,7~4;'~',''&&&ck time the quota dl1be reducd"to iFitW4ihAi&P 5&cktedd%~the line segment EG. This quantity, added horizontally to SS to produce the new supply curve St St, b r i n b the market into equilibrium a t . point G w i t h a volume of bS,000 motor cars selling f o r $2,500 per u n i t . y It is helpful to note that the line segnent CG is equal to AB, the original quota of 20,000 units. Strictly speaking we cannot allocate specific portions of our 15,000 W.ts of new danestic production t o various f'unctiors such as domestic supply replacemer,t, import . L . I rephmment and market expansion We cen however assume f o r the analysis her without distorting the final conclusions,that of the 15,000 units i repreeented by the line segplent LB, whfch shifted t h e domestic supply - curve from SS t o st Sf ,the f i r s t 5 , q (LK) replace an equal amount oC supply tlhich uaa produced domestically by other producers a t higher marginal cost, t h e second 5,000 (KC) replace - an eqcal amount of b p o r t s bezause of tke reduction of the quota AB t o Ea, while the third 5,000 units I! expand t o t a l daaestic market supply. -I./It is obvioua that, given the domestic supply crrve, the market clearing price and quantity are established by the size of the quota, provided that the quota plus the danestfc supply do not exceed the lev91 which would make the market clearing prCce equal t o the import -C ANN= Page T o f 22 - Motor Cars (thousands) Figure C-5 ANNEX C -of 22 Definitions Before looldng a t each of these impacts i n d e t a i l and explaining the UNIM) approach graphically, a few terms should be defined. Cost refers t o the r e a l resource c o s t (including a m o n a b l e - . 3 . 1/ return on investment) of supplying the product.- Total coat fs equal to the area below the relevant supply curve, and marginal cost is ? , equal t o the distance from the horizontal d s to a given point along *. the supply curve. Consumer willingness to4pay qfers-t&tlie -- nmdmum price consumers would be willing t o pay for e-speckf$!.pl%B&ity or unit of the product. Total CWP is equal t o t h e area under the demand curve, and marginal CWP is equal t o the distance from the horizontal axis to a given point on the demand curve. Surplus comes i n two basic forms, consumer and producer. C o n m r surplus is the difference between -1/ A l l of the supply curves i n t h i s diagram represent costs measured at social rather than market prices. Thus, excess market costs of production due, for example, t o factors such as taxes on inputs have been netted out. This corresponds t o the instntctions i n the c tert (pages 12to18) to look at the real social opportunity costs of input+. The present graph could be expanded along the lines indicated - -- at the end of this annex t o reflect market as w e l l as social costs, b u t only at the expense of making the graph unreadable with too many - .* lines. *- C -2/ * Unlike cost curves which must be corrected f o r market diGortions , r) the demand curve by defin!,tion reflects willingness to pay. The main axcoption to this,in wMch the social CWP i s not reflected i n the demend cnr~e,wouldbe conditions of a r t i f i c i a l l y constrained demand (physical rationing). ANNEX C Page 10 of 22 the market price t o the consumer and the price he would be willing t o pay, while producer surplus is the excess of the market price which the producer receives over the price a t which he would be wiiljng to s e l l - (Total and marginal concepts are as above). On the supply side, .I: surplus may accrue as producer. ~ ~ r p ~tloudomestic produce.rs 7 are to be Iinport--.$! a ~dert!:e rim quota, the graph covering Lhe l ~ s 5,000 of t.hsse hes :r\i:a t wed to value the addj.tLci;~aldmestlc production which 'mcreaur-s%he zize of the domestic market becauso this, not. impcrta, is ?.ha rnatgi..ialchr+nj.e. ANNEX C Page ~ c 22f Impact on Surplus The overall inpact of the project and trade policy changes can now b e summed up i n terms of the changes brought about i n costs and the allocation of consumer surpluses. Consumer c u r p l ~ sh m increased because of the.drop i n 1'' '., market prices. M a is represented by the area %+Ib+C4+lh+M3+B3. Of this, all but Q and B3 represent a change from importer t o consumer surplus. % is a net gain of consumer surplus because of a reduction - .-- - F, 1 I" i n domestic production costs. Bg is -a mt,gaia@ iilcddsume'r suG1ns- 0-: ..- . .., I,..' Prcciuuzer surplus incr%es by-?* -(&ma tic-'supply replacement) ,12+13 (import replacement), and C3 (supply expansion 1. h e t o the reduction i n imports, the importer quota rents are decreased by 14+C4+Mj,+f3+13+C3+C2 - B2, 1eavir.g only M2+B2. -1/ Domestic prod~lctioncosts decrease t o the extent that high marginal cost production is replaced (P2+P3+P4), leaving only F1 for t h i s part of dagnestic producticn. On 'the other hand, because of impart replacenrent and market expansion, domestic production costs will It is iqteresting to note throughout this annex, and especially - - i n t h i s last section on expansicn~!of consulaption, t h a t marginal - instead of everage costs and aver* market prices have been used. I f the demand and supply e l a s t i c i t i e s are l e s s than infinite, use of average value? c mld lead t o wrong results. If, for example, the benefit -l/If protection had been i n the form of import taxes instead of quotas, the decrease w x l d hope been i n t a r i f f revenues to tke Gioernment. PNNEX C Page of 22 of the project were measured i n terms of t h e $3,000 market prize f o r motor cars prevailing before the implehentation of the project, the gross benefits of the expanded consumption made possible by the project would be overestimated by the area B I G. On the other hand, i f only the marginal price of $2,500 had been u=ect,:the benefits)would have been underestimated hy the area B F G. The theoretically correct way of h ~ z d l i n gtohisprcblcm would be t o take the integral of the demand curve DD betwem the points B and G. Howevmr y 5 & ~ b ~ , t ~ ? & r , + k zurve is roughly linear, it is sufficient to %aim-- ;hySap of-tha previous market price and t'le mardnal price t o estimate the M area. Other Situations The conclusions drawn on the basis of Figure C-5 as t o the ~ : e tbenefits and costs are of course highly dependent on the position of t h e relevant curves i n this figure. 'vhile t h e analytical principles should be obvious from this discussion, it might be worth mentioning a few of the major changes which would be induced by moving the cmves. Shifts to the right or l e f t In the i n i t i a l supply curve w i l l require an increase or decrease respectively i n the amount of the import quota to maintain a given market clearing price, ceteris p d b u 4 The position of the n& supply curve depends on the production costs *Z 8 of the new project. ff the production of a given quantity of motor cars were more expensive than i n Figure C-5, the S' S' curve would s h i f t t o the left. This would i n turn reduce the producer ~ r p l u s(p2+p3) and, depending on changes in the quota, could even reduce the consumer surplus ?4. With respect to import substitution, an increase in cost ANNEX C Page of 22 per u n i t would increase the s r e a 11, thereby reducing or 2erha.s eliminating entirely the producer surplus 12+13.i f the new domestic production were more expensive than the imports, the area below the new supply c u m would exceed the area below the impcrt supply curve and there would be a net loss tcr:the e c o q . Simi$arly,.i;tf the . . 9,. production were more expensive, t9e cost o f expanding consumpcion (indicated b e h the S ' S' curve as Cl+C2) would increase; thereby *- increasing the cost i n excess 00 + t %pi='u W + Z z &&,gaucing or evm eliminating the producer s q J ~ - . .:.. , . I n the rsccmnendations that the output of a project -3 be valued at the cost of replaced domestic production, foreign e x c h ~ g e saved f o r import s u b s t i t ~ t i o n ~ e r lconsumer willingness to pay f o r d consumption expansion are entiraljr consistent with the basic principles of traditional welfare economics provided eldsting trade policies are assumed t o prevail. However, t o the extent t h a t trade ?olicy cmld be predicted t o become more liberal, it m a y be necessary instead t o value domestic production replacement and production f o r market erpansion at world prices. - - Application to Other Resources 4 * - & A s stated a t the outset of this annex, the graphical analysis * . used f o r traded goods can be adapted t o explain the UNIDO apsoach to the shadow pricing of all resources involved i n a project. This analysis, which represents distortions between market and shadow prices as a tax Jlich yields, for example, p a r a l l e l supply curves representing the - ANNEX C of 2 market and shadow costs of supply respectively, owes much t o the work of Professor Arnold Harberger. -I./-2/ Other examples of its application are i n the calculation of the social value of capital, the shadow price of foreign exchange, and the shadow price of labor. The graph f o r the s a d a l . v a l u a of capital gives savings and investment on the horizontal axLs and the interest r a t e 'on the vertical axis. The supply curve indicates the supply of savings, whose social value may differ fran its market value bx fac+mn,~uch as the difference . . - . 7-.2- '4 % - between the private and the social $ure time p-reference factor. Rn ---- - ',. , > . I demand curve indicates the demand for savings i n the form of investments, and this curve may be distorted from its social value, f o r example, by income taxes on earnings from investmmts. The intersection of the L qand and s ~ p p l ycurves indicate the accounting rate of interest which balances demand and supply for capital. The graph f o r foreign exchange is similar except that domestic and foreign currency are on the tv3 axes, and the intersection of demand and supply(in terms of the amount of. domestic currency which is offered an , asked foi. foreign exchange)indicates the exchange rate. Import taxes on goods imported w i t h foreign exchange w i l l , for example, distort the .3 demand curve, yielding different equilibria i n social and market price * -l/ The analyais also applies t o demand curves where the true consumer willingneee t o pay is distorted, f o r example, by taxes on returns f r a capital. Also, the two curves do not have to be parallel; i n fact, this will be so only when the distortion is a fixed rather than an ad valorem amount. The curves may even cross, as was seen in Figure C-5, where the domestic(market value) and import (shadow value: supply curves intersected. ANNEX C Page 20 of 22 The graph for the shadow price of labor is shown i n Figure G6, where SS represents the social cos5 of supplying .Labor and SSlc the private offer curve. The private supply price i s determined by factors such as the dlsutiUty of effort, the cost of migration, the cost of education,union pressure, and minimum wage legislation. I f , i n the view of Oovernment, the cost to society of the disutility of effort,for example,is overstated i n market wages by an amount Pj-PI, the curve SEx&itfy~ep?ese$t' the 3hai9Y kmtm supply curve. I n t h i s case, i f tb.pjCs$&Vt&re'smaflt an&'meitl .. .f . . Measurement of the Savings Impact a T h i s section is related t o the third stage of the U N D process -- accounting f o r the impact of the project on savings and investment. To do this, t h e amount of income gained or lost by specif%e WMI~ the society (e.g. government, private, consumers, workers) because of the project must be determined. Once t h i s has been done, it is relatively easy t o calculate the expected net savings impact by sach sector based on its marginal propensity t o save. The net saving impact can then be multiplied by the social premium on investment over consumption t o determine the additional social value of the savings impact of t h e project. Considerable care must be Laken i n determining the amount of income which each sector gains and loses. This annex gives an example of the reasoning . I which goes into t h i s process. The analysis is greatly facilitated by usjng a table such as Table BS (Standard Income low), duplicatgd here with numerical 9 entries as Table Dl. - This table identifies each of theLitems which may generate -- incom#flows and a sanple of major sectors within the e i e t y which it may be - significant t o consider 1/. By breaking down the income flows in50 those due t o . It may be desirable to add or subtract income groups depending on the type of project -- AfLiiJ 3 Table D l L ~ c e2 or'l5 pages -s t m . w d h e e m now- - - - - I t e m Prlnta RoJect a c t o r Gowr-nt iprkurr Coruumrs Extarnal .I/ atplt A. Wietrd .I. Subrtitution .2. elpursion /2. Non-marketod .l. Housing .2. O t h n r .2/ #torial Inputs P-. Forelm /2. 'hmrtio .2. Other /2. L.bor .l.SLdl1.d .2. UMkil1.d /3. Capital -1.htarlala .2. Labor .3. Currency .b. Orhr .b/ Social Qerating Prefit .4/ net Social %qratFng Profit .2.Additional Flows chu to lbrlnt Prices .I/ (k:>ut Jl. Yuketrd .l. Qoda .2. wrt diaLdorhrmw /2. Won- Uuketrd .l. HJluinC .2. :Foreip tahnge .ld Exass ?hmcial Proflt (bfore -s) 4 AImx D Page 3 of 15 pages payments a t social prices arid those due t o additional payments a t market prices, this ?able makes it easy to obtain the relevant net present values from the standard analytical tables prepared in social and market prices and t o distinguish the source of the income flow 1/. Since a l l relevant groups are identified and since, with the soie exception of net profit (social and- private), there is a gainer for every loser, the table creates a double entry form of bookkeeping uhich makes it easy t o check for errors. -.,. 9 The following section uses a simple example t o shm)rarr a tahb _ .-- 4. like D l should be completed. The entries in the shadow and market price sections w i l l be handled as pairs (eg. foreign material inputs a t shadow and market prices). output The project w i l l produce output worth Rs 100 a t siiadow prices and R s 115 a t market prices, so Rs 100 of income is credited to the project in section .1 and an additional Rs 1s in section .2. It can be seen on line .l.l/that the Rs 100 gain of income by the project is balanced by a combined 1 loss of income by the workers, consumers and the external sector. The reasons for this ellbe q l a i n e d below when the components of the Rs 100 are looked a t 9 individualfy. A t this point, it is only important t o note that, except for - &* operating profit before and after tax i n social and market terms (lines .1.W, rn 0 w .1.6/, and .2.L/, .2.6/) a l l lines a t both the subtotal and total levels w i l l A l l values in the remainder of this paper are net present values. ANNEX D -of 15 pages add t o zero -- the gains and losses of incone must balance out exactly. Of the R s 100 worth of output a t shadow prices, R s 90 le a marketed product (eg. razors) and R s 10 is a non-marketed output (worker housing f o r which no charge is made). Of the R s 90 at shadow prices of marketed razors, R s 60 are substitutes for imports, so the external sector l o s e k ~ e60 of income, and Rs 30 represent an expansion of the domestic supply, so consumers lose (spend) R s 30 over and above the R s 60 they previously spent on the razors. If the project would only produce impart &&&uwq . ~ r,.j&-,.,hese d t were t o be sold t o the consumer a t the sahe'price as. & e 4 ~ ~ t ~ d razors, the project would have no impact on consumer incomes. In addition t o the imome flows generated by payments of the social vahe of ';he marketed output, an additional flow of R s 25 from the consumers t o the project is generated by the fact that the project sells the razors for Rs 25 more than thei.r social value (line .2.1/1 .I). It is interesting to note, by the way, that if there had previously been a tax of R s 25 on the i-lnported razors and if the projectwere t o produce only importLsubstitutes, there would have been a R s 25 gdri a t market prices t o the consumers in import duties no longer paid, which would have just balanced the - - Rs 25 loss of income t o the producers, leavidg the consumer's income position L * unchanged. The net effect would have been tZj transfer Rs 25 from the government w t o the project, which is often the case withwimpor%substitution projects i n developing countries. ANm D -of 15 pages Housing, the mn-marketed output, mst be treated carefully. The workers do not pay the project for the use of the houaing, but i f one is thinking in t e r n of social values, the consumers w i l l sacrifice R s 10 of income for the housing uhich, again in social p ~ i c e s ,is gained..bythe project. ( h e .1.1/2.1.). However, i n t e r n of the additional flows of income generated by distortions between the market and social prices, the worker gains the Rs 10 of hou-skagd I?W*jdW P;~SGW,'Sfit= ( e 2 1 / 2 1 ). The nee;result is-p+caimmxq - khaW&;~,-s&c& gains --c=A ' ' I no income from the o u t p t of worker housing and the workers lose no income by occupying this housing, but it is necessaqy t o record these implicit flows t o make certain that everything has been accounted for. Foreign Exchange It has been implicitly assumed in this example that the social and market foreign exchange rates are identical. However, provision is made in line .2.1/2.2. of Table D l for tho loss and gain of domestic currency through the purchase and sale of foreign exchange a t undervalued prices (due t o overvalued domestic currency). Although foreign exchange - i per se is marketed, there is no market for the net foreign exchange impact ,* which a project w i l l have on the economy -- it is an externality much -a 8 r l i k e housing which is provided free of charge t o the workers. This m externality is recorded in the market price instead of the shadow price section because the net foreign exchange impact becomes important only uhen the market price of foreign exchange is not equal to its social AWEX D P a g e o f 15 pages value. If,for example, the market and social values of foreign exchange had been different and there was a positive forelgn exchange economic benefit of Rs 20, t h i s would have been recorded in item l4.2. of the Standard Income Statement. It would now be recoraed heze as a gain of R s 20 for the government'and 3 loss of Rs 20 fop the consumers, assuming that the gove~nmenthandles a l l foreign exchange i n the cuuntry and that the importezs who had previously parchased foreign exchange from the qorernment to ' ~ D H ~ E& the- M .>,. on t o the customers She "savingsn which-~ ~ & d " f r ' opurchasi,sg~Z~sltsipp r n exchange with an overvalued domestic currency. Material Inputs In the material inputs section of this table, four sectors are involved -- the pmzect, ~ d e producers i n the private sector, the r government, and the external sector. This vrojeci uses R s 68 cf matorials when valued a t market prices. However, their value at social prices is only R s 50 (line .1.2/)j of the other R s 18, Rs 15 go as excess costs t c inefficient (or making) daxestic prodncers (line .2.2/2. ), and Hs 3 go t o the . - - government as duties on the imported materials (line .2.2/1..2 .) Sinca , g these materials would not have beer. imported in the ~baenceof the project L * - * - the duties represent a net ,ain t o the government, O f the Re 9 of material inputs a t s ~ c i a lprices, Rs 20 u e paid t o the extern& sector . and Rs 30 are paid t o ot.her prochcers in the private sector (lines .1.2/1 and .1.2/2. respectively). ANNE D ?age 'I'of 15 pages Since it is assumedthat the other producers i n the private sector have the s x w marginal propensity to .save as the projsct, the income loss of R s 30 by the project is balan~edby a corresponding gain ir, t'-s remainder of the private sector =d there is ~ i onet impact on savings. The same is 3 , s true with eqenditures oil labor and capital inputs. For +.hisreason, the h i d e l i n e s do not specifically identify these resource flows i n the case ~tudies. However, t o avoid any confusion regarding t h s +4wqf#fz&#G'i*~nfdn~ of such flows i n special cases, the methcd presented h e r s is.4e- -to& r i l e the project analyst t o cross check the completeness of his analysis both within the income flow table and against the standard income statement. Operating Expenses T h i s project ases no invisible inputs i n the sense of real resources, hence there are no entries i n section .1.3/1. Nor does the project i n t h i s example pay for any invisible inputs which have a value only at market prices 1/ mch as indirect hsinesc taxes . - Labor: In the unskilled labor category, the project pays out a t o t a l of .Rs2 ( 1 2 p 1 2 . 2 . ) This is composed of a real expense at shadow - pr5aes of R s 30 (which represents an income gain t o the rest of the przvate .c eector because it no longer pays f o r these workers) and a net gain at market - = prices of R s 5 due to- the differences between the market prices for labor which I) the producers pay and labor's shadow value (due t o t.raditional wages, labor legislation, atc.). This loss of lncome by the project is a gain to the unskilled 1/ The U N l M Guidelines k ~ v egenerally assumed away the possible social -~aiue of taxes, so we will cantime to do so here. ANNEX D -of 15 pages workers (line .2.3/2.2.). On the other hand, the project pays skilled labor a t o t a l of R s 10 less that1 i t 3 social value. This is recorded as a gain of income by the project and a loss by the other producers i n the private sector who themselves no longer enjoy the benefit of t h i s "underpricedtt labor (line :. 2 . 2 1 ) I f the project had used foreign labor uhich would otherwise have been used elsewhere in the domestic p-rivate sector, the income l o s t by the project would have been balanced by a gain in this sector. The external sector i.- would gain nothing, provided the rate of inc&'repa- resmfnecI the same. ---. .. . ...>. ' . . I However, if it changed or if the labor ; b u 1 ~ M 1 o t h ~ h r ihave been used by ~ e the country, the project 1s income loss w i l l be balanced by an external sector gain equal t o the repatriated wages, the remainder of the wages constituting a gain t o the worker sector. Capital: It is convenient t o think of the capital expenses and returns (cash flow) of projects as consisting of three elements -- principal, interest and profits. Speaking in terms of social values, the principal comesponds t o the real resources valued a t shadow prices which are . a invested in the project; t h e interest corresponds t o the social discount rate, and the profits correspond to the net present social value of the projecC. - Profits as defined here are excess profits i n a sense. Lf the project just earned the opportunity cost of capital, there would )e no Itprofits". The project would only recover the opportunity cost of capital and yield a zero net present valu I n the standard income flow table, the income flows generated by the nprir,cipallt investment i n real capital resources are handled in lines .1.3/3.1. and .2. at shadcw prices and in lines .2.3/3 .1 and .2. a t market prices. In terma of ahadow prices, tho project invested Re 10 of capital in:.-~terialo PNNM D Page 9 of 15 pages (equipment, building materials, in7~ilt0ries,etc .) and in labor (to build the ~ l a n t ) . On the assumption that these capital goods were acquired from the rest of the private sector, thie generates a loss of incorrie for the p m j e c t and a corresponding gain for that sector. The ftinterestffo r opportunity cost of capital has already been handled implicitly by discounting the a m a l :.valueo,~~associatedwith the project t o the present by the social disccunt rats. The marginal product of capital which w i l l be sacrificed because of future consumption out of the income generated by the project is onBCb2-G h .&&(irs that detem&as=di&%, .' ' prerniwn on investment which b r i l lbe appliekat the end of t h i s stags th analysis, Subsidized interest rate3 offered by governments t o firms a s an investment incentive may create a slight aaalytical problem. Since the "interest* cost of capital has already becn handled i n the above paragraph t h ~ u g hthe social discount rate, any suSsidies t o the "interest" rate actually paid rmst s p i l l over into the profits of the firm a t market prices. Since both the principal -d interest componente of the capital cost of the project have now been measured a&shadow prices, #ashave the other inputs and the output, the !*social operating prvfitw in l i n e .1.4/ does not need t o be adjusted for any subsidies y. market prLces, however, the income flows generated by & interest subsidies do, however, require adjustment a t market prices, for these s effectively transfer income from the subsidizing agent (eg. the governms~it;t o the project. The adjnstment, which is entered i n l i n e .2.3/3.3. (currency) is Line .1.5/ (Income ax) has been included only f o r symmetry between the shadow and market price sections of the table -- t h e social price of such taxes will general* be zero. This of course makes line .1.6/ (net social operating profit) redundant es far a s the preseyt exanple is concerned. ANNEX D PagelO of 15 pages calculated i n three steps. First, multiply the net cash flow i n each year by the fac%or : . r where r is the social discount rate and t is the time factor. The result is the ninterestrdeducted from the net benefit each year by discounting. Without going into a proper mathematical ber-kWq~,,:of.$his lmltiplicative . . d < factor, it is obvious that t h i s factor-plw ~ . , p s c o u n tfactor .' . F is equal t o unity. Therefore multiplying the net cash flow by each of these simply divides it into two parts uhich add up to the total undiscounted net cash flow Second, calculate the actual interest paid i n each year based on the market intf-yest rate. Third, calculate the excess of the social interest fipaymttntll over the market interest payment i n each year and discount these values t o the present. If the excess is negative (if the market interevc rates are subsidized), t h i s becomes a lass of income t o the Government, for example, and a gain of income t o ,the finn. The result is recorded for this project i n l i n e .2.3/3.3. (currency). The "profit" of the project i n shadow and mark& prices represents ? ' a net g o to the economy and to the producer' respectively, so there is no * balancik loss in lines .lob/ and .6 or in lines .2.4/ a& .6/. It is assumed in the Quidelinea that market profits remain with the firm t o be saved and consumed according t o the firm's marginal propensity t o save. However, if fcreign equity is involved i n the project and has the right t o repatriate part or all of its profits, two more lines should be added both ANNEX D Page 11of 15 pages i n the market and the shadow price sections of the income flow table. The first line should be for expatriated income and the second for net social (financial) profit t o the country. The repatriated profits are a loss t.: the firm and a gain t o the external sector. Since repatriated profits indicate a zerc domestic marginal propensity t o save, the repatriated earnings would contribute nothing t o domestic savings 2nd investment. I F t h e presence of for&@ equity gave ihe iirm a marginal propensity t o save which was significantly different from t h e marginal propensity t o save of the private sector, this should be reflected i n line .4. The difference bet'weerl" tlie ,v.a2~a:.&f+qined,by t h i.s . ; ., adjustment is roughly the same a s the d i Q f e m m . c. 4ross ~domestic product ~ ~ ~ and gross national product. The income tax of RY 2 paid by the project is an additional flow of income due t o =ket prices (line .2.3/) and should be recorded a s a loss of RE 2 by the project al-A a gain of the same amount by the government. There is no entry urider "flows at social pricesll because income tax does not, i n t h i s context, have any social cost -- it is a simple transfer payment. The data on the bottom line of Table D l (BS) indicates the net impact on the projectaon the flow of income within the econow. T ~ L Sdata can be used t o calculaSe several other important data including the net impact on savings an? investment, tho net impact on government revenues, and the redistribution of income among s ~ c i a classes. l - 8 Calculation of the Inve'stment b p a c t Value The impact of the on savings and investment is an inter-temporal .?om of consumption distribution -- savings generate consumption benefits vhich are received i n the futcre instead of immediately. The value of the savings/ investment impact depends on three f a c t a x : who gets the additional income, ANNEX D Page 12 of 15 page8 how much each recipient class normally saves out of additional income, and how valuable is the additional investment i n t e r n of consumption. The flrst factor hae almady been detemined by Table Dl. The income received by class i r in ,w entered on the f i r s t line of Table D2, the Social Invesment Value Table. Next each value is multiplied by the second factor, the marginal propensity t o save, t o determine the Net SavFngs Impact by i l s ~ s . , These results, which add t o 28.0 i n thin example, are th&.multiplied by the third factor, the social premim on investmect, which is the percentage by which the value of a unit of investment exceeds the value of a unit of _.I- .", consumption. The derivation of the social prim W f ~~- 2 3 5 1 . - 1 . - tenas of its' consumption equivalent, which is equal to the pre&bwpZasa'xty, is described in the text on pages 42 to &3. Table D2 Social Value of Investment - Private Sector i t e m Project (excl.~roj.) Government Workers Consumers External - 55 - 40 Net Oaln (lone) of Incacs Marginal + 25 75 - 1 0 . + 5 1 + Propensity t o Save 0.5 005 0.6 0.1 0.3 0.O - - 16.5 Net SarLngs ,@pact (+ 12.5 +37.5 6.0 +0.5 0.0, Y - L Total Savipgs- Impact 28.O r 8 * (. Social Reuium on fnvestmmt lJ 6.O 2.8 1.06 Invs '-merit Impact V&W . See the following text for the asrtumption Mch generate the alternative values ehoun here. ANNEX D Page 13of 15 pages Ln the present example, it is assumed that the marginal productivity of capital is 25 percent, the margins1 propensity to save and reinvest out of profits arising from future reinvestment is 30 percent, and that the social discount rate may be 10, 13 or 16 percent. The sensitivity analysis implied i n these l a s t three figures generates three different social premiums on investment. mltiplying the t o t a l savings impact by these alternative social premiums on investment i n turn generates the t h e e possible investment impact values as l i s t e d on the next line. These values may then be added ~ ~ to the net social benefit of the projec$~&f#m 4.66tainadtn&t%gp~~Z..@a! the UNIDO andysis thus taking intc acce=??&-h;&~%ertempor& 2bpas6.d~f,tbe project . Governme!.t/?rivate Income Impact Although the UNIDO Guidelines do not go into the subject in any detail, it has been noted by other analyets that, due t o the difficulties which governn~entaoften face i n establishing an lloptimalfl tax system, income i n the hand. of the government may be mare valuable than income i n the hands of the private sector. . r ~ & l eD l providee the information on the inpact of the project on - government income, thereby making it easy t o place a premium on this impact. - For example, if the government regards its d d i t i o n a l income a s 20 percent - more valuable than additional income i n theaanda of the private sector, the * D net income impact value f o r the government sector i n l i n e .3. of Table D l should be nultiplied by 209 and added t o the value obtained i n Stage Two of the andlysis. If, as i n the case of our example, the government suffers a net lose, the premium on t h i s loee is appropriately subtracted from the net benefits of Stage Two. ANNEX D of 15 pages Rich/Poor Impact Stage Four of the UNIDO analysis haidles income distribution and other social objectives, and Table D l pro-iride~the information necessary to calculate the benefit of the project i n terms of its h p e c t on the distribu- tion on incone among different income classes, provided ox' course that the table has been broken down into a sufficient rnrmber oi sectors. To the extent that workers are s y n o r p u s with the poor, the breakdown given i n Dlmay be satisfactory. However, if a more detailed income distribution impact analysis is desired, it may be desirable to include ab Zt2gs* tejo'&- 4U&SJU= and perhaps even two classes of consumers-.[~pVfeiSts; 'sutibc~~dwiqp 'by cost domestic consumer goods for cheap imported goods, can adversely affect income distribution). Provided the appropriate breakdowns have been made, Table D l gives the net incremental distributional impact by groups. The only task is then t o apply a premium t o extra income going t o the poor. If, as is true in most conventional cost/benefit analysis, the income of all groups is regarded as equally valuable ,the conversion ratio of the value of income i n the hands of the poor t o the value of income i n the hands of the rich will be equal t o unity, and unity minds unity gives a zero premium. If, however, income i n the hands of the poor is considered to be of 1.5 tines - - the value of income i n the hands of the rich,for exanple,the net gain of the po?r as indicated in Table D l should b8 multiplied by a premium of 5Of, and be added = * t o the net benefit as calculated by Stage Three. . 0 , t It should be noted that the UNDO method does not build in the measurement of income distribution by creating a r t i f i c i a l shadow wage rates i ANNEX D Page of 15pages as is done i n some nlethods. This has the dual a&antage of leaving the St-ge Three calculation in terms of pure efficiency prices, thus making clear the cost of seeking income distribution objectives through the use of project selection, and it allows for the inclusion of income distribution impacts in cases where this would not be measured by wage rates (eg. the production a t high cost of consumption items for the poor). The logical extension of this type of analysis would be to weight the income of each group according t o the ratio of the value which society , * Y*> i . places on the income of that group t o the ~r;~hrb?%?iTkid'itplaces +om - . income going to persons with the average level & ~ . ~ 6 3 h done, i b . m 3 d i s be possible t o add up the consumption benefits going t o each income group in terms of the consumption equivalent of income received by persons a t the average level of income without implicitly ignoring the fas'; that consumption of these different groups has different values. This wczld also solve the gmblem that, technically speaking, it may be necessary to use a different consumption rate of interest t o discount benefits received by differeut inco:ne groups t o the present. Once a l l consumption benefits are expressed in terms of equi~lalentaverage consumption, the CRI relevant t o consumption a t this - level of benefits can be used. -? Regiolial Sncome Distribution * If the government wishes to evaluate projects in term5 of their rn impact on the regional distribution of income, Table D l should b% broken down further by region. It would then be possibla to place a premium on income going to the target region much a s was done o r income going t o the target income group. *r+. Ar - :. Table 3.3 FINANCIAL CASH FLOW (FCF 1.0 - 2.0) - i Net Present Value Years Item 0% 10% 20% 0 1 2 3 3ta - b 1. Net OperaJing Flows [lO,OOO] [3,633] (-7271 -14,000 -2.000 -1,000 7,000 20,000 - - .l. Sources [18,000] [14,358] [11,735) 2,000 7,000 9.000 - - .l/ Operating - p r o f i t BIT' 10,000 7,809 6,250 4,000 6 . W - . 2 /Depreciationd .. 8,000 6,549 5,485 2,000 3,000 3,000 .3/ Externalities ·- · a · 1 · , r e · .2. Uses (~nvestrnent)' [8,OCO] [lO,726] (12,462) 14,000 4,000 8,000 2,000 -20,000 .l/ Current Assets [0] [2,2181 [ 3 , 6 2 0 ~ , ~ ~ ~ , & @ ~ , : 4.&~#7$,am rk-1s;;k3 I h 11. Cash - 0 482 790 l,,, ~,mi-t, r.mt . -3,000 -- 12. Receilables 0 458 T l G . .. -.-. l,tM$" 4,000 1,300 -7,000 - 13. Inventories 0 1,288 2,116 3,000 2,000 3,000 -8,300 - - .2/ Fixed Assets [8,000] [9,498] [8,842] 10,000 -- -2,000 -- 11. Land Bldgs. - - - 6 4,000 _ 4,498 4,842 6,000 -2,000 - - - - 12. Equipment 4,000 4,000 4,000 4,000 - - - - 2. Net Financing Flows 20,000 - . I . Sources 400 700 7,000 .l/ Net Operating Flovs (1.0) -14,000 -2,000 -1,000 7,000 20,000 - - .2/ New Borrowing 7,000 2,400 1,700 - - - - .3/ New Equity 7,uOO ' - - .2. l'ses 4 00 700 7,000 - - .l/ Debt Service 400 300 4,700 - , - I 11. Interest 400 300 700 - - - - 12. Principal 4,000 - - .2/ Taxes 900 -706 - - -iW3 - 400 500 - - - .3/ Dividends 900 3676 - 521 - 900 - - - - .4/ Securities 900 -676 521 900 aColumn @ 3' records liquidation of p r o j e t a t end of third year. m - b ~ h i ssection contains the cash flow components resulting from productive operations--the cash flows required f o r calculating the net present value of the rate of return. (1.0 1.1. - 1.2.) C Operating Profit before interest and taxes from FIS 9.0. d ~ r o mFIS 8.3.11 (page 19. Table 3.7). e This section contains the cash flow 6 ~ p o m e n t srelit. short-falls and/or other expcn!es, or (b) use of funds fo Oxpensee. SALES PROJECTIONS (FlS 3.0 - 3.2) I Year I - - 3 . Sales (factor cost) - 10,125 29,813 38,925 -, .1. Razors--Single Hcnd .l/ Veluc 5,375 15.938 20,675 , &Northern Market 875 1,628 -1,880 12. Scuthern Market 4,500 14,310 18,795 .2/ ~uantit~*" 120 ' 355 46fJ /1. Northern Market 20 35 40 /2. Sou~hern-Xartct 100 292 376 .3/ Price (each)' /l. Northern Market 43.75 46.50 47.00 1 2 . Southern Xarket 45.00 49.00 49.99 - .2. Razors--Double Head ..,, j ' , .l/ Value 4,750 xi"'-"^ i 1&2W "" ' 11. Northern Market 1,500 1,875 I -?+z2,k.! ' /2. Southern Market 3,250 , 12,aQl.h , '" r6.000 .2/ Quantity 75 175 230 /I. Northern Harkct 25 25 30 12. Southern Market 50 150 200 .3/ Price (each) --- 11. Northern Harket 60 75 75 12. Southern Flarket 65 I 8 0 80 *Indicates supporting documentation is attached. Assumptions (sample only) : a Southern Market vill begln at five r h e s the size of Northern Market and grow much more rnpidly. b~ecauseof higher incomee, price in Southern Narket will be higher than in :i~rthern. In both markets, prices will be lover initially as an introductory offer. - L Table 3.5 SALES QUANTITY PROJECTIO%. SINGLE HEAD RAZORS IN NORTHERN MARKET , (FIS 3.1.2/1.) I - I - Year I -Single Head Razors 11. Northern Ma-'--- -// Dnbba 1 . . . // ... . I -1 Toble 3.6 MATERIAL REQUIKEHENTS (FIS 6.0 - 6 . 2 . 1 / ~ ) Item 6. Material Inputs a- .l. Domcstic --- .l/ Basic .--- 11. Steel .2/ Other /1. Paint /2. Other Process .2. lmportrd .I/ B n ~ i c /l. CKD Components Toble 3.7 OPERATJ NG EXPENSE DETAILS (FlS 8.0 - 8.3) Net Prescnt Value Year I tern 8.1. UTILITIES AND SERVICES 6,500 .. 5,227 .. 4,305 .. 1;000. 2,500 -3-,000 .l/ UtlliJ*~ -200 .. .. .. 100 200 .2/ Trannport -400 .. .. .. 100 300 .3/ Selling and Admjn. .. .. .. -500 1,000 1 , 0 0 0 - -- .4 / Technology .. .. .. 100 100 .5/ Indirect Bus. Tnxcs .. .. .-. I , 300 800 9QO .6/ M i ~ c elnneous l - - 100 8.2. LABOR 9 , 6 6 6 ] 7 [ 7,9741 2,OC -- .1/ Unsklllcd 0 ' '4,912 4,050 1,000 2 , 3 0 0 - - 2 @ 0 0 .2/ Skilled 5,900 4.754 3,924 1,000 2,200 2,700 8.3. CAPITAL CONSUMPTION 11,000 9,035 7,591 3,800 .l/ Depreciation 8,000 6,549 5,485 2B00 5-- .2/ Ot- 3,000 2,486 2,106 .. .. .. 1,000 ?,000 1,000 /l. Rcpoirs 6 Hnlnt. .. .. * . /2. Lease&Rent Exp. -. /3. Pre-Operat ional .. .. . ,Expt.nses 800 TOTAL . 29.500 23,928 1 9 , 8 7 0 "Value included in t'otnl (thcrc is no nced to calculate thc net yrcsent vrlluc of . item8 whose oarket value w l l l not be adjusted individually to r ~ f l r c te c o ~ i o ~ ~or~ lsocial c values. The ~suumptlonimplicit i n these linen that thcrc a r c no distortiona w l l l probably not be rensonable in practice. lu 0 >I .-I v a J a J P ) d ) > C U X ' r l U P ) . r ( U N > k L s C 0) P) up: . g P) 0 4 s - * u 0 3 I, COCT nlw a a m U ) d 01 E U X w 2 2 a P E . U V ) C l n b U ) m b d) 0 P) U a m P ) . 4 R Q C b . I , * a , 0 a u 0. rl 0 0 1 4 U C 3 ~ s a m u o u c 1 . 4 X - O 0 : u c n 14 c.4 O: - P ) ulv d a * ?IC P)w a u b l . r ( r l a u U ) m a J u 2.: s 5 4 2 X P) u - ~ a~ ) w b l u a l u a J I,@ v m u P ) w c v I , c r-abacu UI, d -"' 4 X-4 o o w U b U C & W' 3."* I O I , d 9 0 c a ; r u > c 0 3 u-lbl o u r l 0) 0 u P) I, a o w U D arc c J : 0 b a 2 2 0 U ) O U ) : ~ S ) C.0 P) C 0 -rl U 3 C a d - O d I , U OI m m 3 U b V O Q U ) rl4JP)U)Cbl JJ r l a l u c aJbldbl L C r l r l C b UP)*: r l a 0 S U ) r l lu ELL) c m u S , , D P ) O U b n c d f aP) P ) Q . b > a"-,-5 4.4 u o w a u c 0 1 EZW s a c P ) r l F ? , Z V 0 w U P I,- CU o e r l u 3 a u a P) c U U)rlu-lO aba Q l u O r l W O U Q u N.-II,P)P. u b U ) O c ~ a u j 1 s " D r S Z z U P 0 a m U) N V O D ¶ ) a w :sU)::: C1u.c P C ? u-l Qrl X U m O . l I , d r l u o n a 8 3 Q I , I , C o a u a a a C G b s O r l J P) 0 4 U O C 13 Q'CI m u a 3 C I, u Q r O U l u 4 U i5,:Oaa n ~ s u r u c 23: :a 1 a n U) C 0 b a a u "-3i C . 4 a, ooh ail 5 -a! 0 rl LJ I, U U J I , Q ) O U T:Tq 0 C a Q ) P i $ 0 N 9 E a l . 0 4 M d , r-l r'l u m,. -64 u.7 7 4 . 4 e w h r-l C Q bl U - J Q ) O 4 a l o a m C C d W r : I , .d 0. O a J o U o U 4 I rn Q,0 r-l r-l 0 a" u-3 W U U C X b ) - 0 c u m u a , , 2 . a a , ln I f N I S v.4 0 C C N d m a,'d a, e m . : :- U + + 9 N r-lt n 4 ' U U M d 4 W c d l Q . a , Q ) O Q ) O 4 I, I,5 U I,C I, 0 0 UI U U L Q ) w U @ J B M L E 0 4 O C O U C W I, 4 C.d u O W E I: u 4 2 h Q ) a,.,.; -+ Q).a - M a d C'O 4 m 6 U C I, d U U :Sd 54,: iii u c ; w ~ r l dxuQ 0.u ln 0 3 X.% 1 X d Q 0 - Q ) . u lnd N Q) 3 z