Mexico - An economic evaluation of irrigation rehabilitation projects in Mexico

 R.p.rt N•• EC ... 180 FILE COpy This report may not ,be .published nor may it be quoted as representing the views of the Bank Group. The Bank Group dbes.'.!\ot accept responsibility for the a~~uracy' or completeness of tb.e report. i.- INTERNATIONAL BANK . FOR .RECONSTRUCTI.ON . AND DEVELOPMENT INTE~ATIONAL DEVELOPMENT ·ASSOCIATION i\ .' AN ECONOMIC EVALUATION OF IRRIGATION REHABILI'r.ATION PROJECTS IN MEXICO .~ . '. . September 1971 Economics Department Prepared by: Richard C. Manning Ofelia M. Ela (Research Assistant) /1  EconoJllica D,ep,artment StUdies on Pm b1ema ot ,. sector and pmject Anaiisla. * IC-117, J~ 1$., 196.3, llu.n G. TE dttr Talc, The, ~onomic. CollJ'arison of CFleCtric P:mJectl!l nth Altemat,ive Dcveloemerits .ot Tberlllal Electric - over Eo.128, ~ 7J 1~641 Hema,n G. va der Talc, The EvalllatioD ot Agdcul turJl Projects: A ~tud,Yot:' So.Eco~om1C and Financial' Aspects. EC-130, October 26, 1964, Lee Charles Neb.rt, A Pnl-IDveatlllent Stuctr ot the Flat Glasa, Industxz ', ~-1)2, Januar;y 22, 1965,. He~ G. van der Tele aQ.cl "ochen K. Schmedtje, Econom:lc~A.spects of Wate,!\ utilization in Ir.rilation ~ro~e~ta. EC-136, October 21, 196,. Jochen K. SchMdtje, On Est.imat;ng the Econamic o,st of Capital '" " * EC-140a, Dacaber 16. '196" Jan de Weille, Quantification of Road User Savipgs * EC-141, Sep.teJDber 26, 1966, Herman Q. van der Talc 8J)d Jan de Weille, An Economic BeifraiSal ot a Road Project: The First Iranian Road Loan or 1959 {IRN-227 ' EC-1SI, December 20-, 1967, Mark Bl:iug, A Cost-Benefit Approach to Educational Planning in Developing Countries * EC-1,8, Jan~ ll, 1968, Alan, A. Wa+~r$, The E,conomics of Road User , Charges i\ March l~ 1968, Hennan G. v~ dar Talc and Anandarup Ii EC...16O, ~, !!!! . Economic Benetits of Read Tran~ort Projects * ~-161~ Mq 29, 1968, Ayhan Cilingirogl.u, Manufacture of Heavy Electrical E9,uipnent in ~veloping ~UDtries. . * EC-162, Hq .31, 1968, Jack BaranSQn, Automotive Industries in Il!ve1oPing Countries ' EO..164, August 21, 1~68, Shlomo Rel1tJ.inger, Techniques tor Project Appraisal under Uncertainjg'. ' " . £c..169, ,MaT 1969, Helen,llqhes, Problems ot Food P~ce:ssing Industries ill Developing o,untries ' '" , ,~ '.. EO-173, November 1969, Hans Heinrich Thias and l.fartin Carnoy, (}.)st-Benefit Analysis in Education 'EC-177, September 1970, Q)nrad J. oort, The Economic Regulation of the HO£! Transpgrt Industry ~~z.ia$ued. in World Bank Staff Occasional. Papers, distributed by John$ Hopkins Preas. / / PREFACE 1. '!'bis study' is the second completed b.r the Sector and Projeots Studies Divisioo of the Economies Department in the area of development project re-evaluation. 1he first such stud7 was prepared by-H. G. van der Tak and J. de Weille, entitled Reagfjaisal of a Road Project in Iran (World Bank Occuion.l Paper No" 1, 19 9. This was preceded by a com- prehensive but .are general review of several WOrld Bank projects b.Y A. o. H1rsclBan of the Brookings Institution, published \Dlder the title Develomant Projects {l)served, (Johns Hopkins Press: Baltimore, 1968). 2. These studies emerged from a growing awareness that the last two decades bave provided an accumulated experience of the development process that bas lain relatively unexplored and poorly' documented. lhe Hirschman study served to distU some general insights from an in depth iJlauersion in several project case histories. The aim was to explore those factors atfecting the successes and failures of projects within the de- velopnent experience of different countries. The Iranian road study was by comparison .ore specific to the appraisal procedures used and their surficienc.y in relation to the eventual realities of the project. 3. Similar:q, the present study of Mexican :L~rigation projects is concerned with a re-evaluation of the appraisal prooedure followed, as well as with identifying the exigencies that might have caused the benefits to deviate from the expected. The report is introduced with a short history of the irrigation districts and the nature of the works involved in the re- habilitation project. Tbis is followed b.r critique of the original appraisal and a comparison with an ex-poet evaluation employing the same methods. 4· The reported 8tudy was undertaken in clos8 cooperation with the water resources ministr,y of the Government of Mexico, S.R.H. (Secretaria de Recursos Hidraulicos) and with the indispensable assistance of the irrigation districts division (Direccion de Distritos de Riego). The authors wish to express their gratitude to the director of the division and his officers at all levale, including those at both S.R.H. head office and the individual irrigation districts, whose willing assistance has made this work poasiQle. Except where noted to the contrary, all the data pre- sented in this r~t were supplied by S.R. H. from i ttl records. The authors hope that this stu~ will provide some guidance to the appropriate basis for selecting future development projects undertaken b.r the ministry and that ita conclUSions will be the starting point for more specifiC stUdies of the pro- blema touched upon. ,. Thanka are owing to S. Reutl,inger, A. C. Egbert, G. Ablasser and P. Duane tar constructive canment on earlier drafts. Mrs. Carole Carr assisted with the, early stages of the study_  T.lBLE OF ann'mtrS It -. PREFACE SUMMARY AND CONCLmICfiS I. INTRODOOTICI: AND BACKGROOND . . . . . . . . . . . . . . ... 1 1. The Purpose or the StuQy • • • • • • • • • • • • • • • • • 1 2. The Role of Irrigation in Mexican Agriculture ••• 2 3. The RortbweaternProjeet _ • . . • • • • • . • • • • 3 4. The Nartheaatern Project f • • • • • • • • • • • • • • • • 5 \1 \ \ S• Evaluation MathodolOl1 and Ou.tlin, • • • • • • • • • • • • 7 '·r II. THE NORTHWlSTBRN PROJECT. ·.• • ·... · .. ·... 8 1. 2. An Ex-ante Review Anaq81a ot Cost ·· · ··•· • · · ·· • 01 • · · · · • · · • ·· · " • · • • • 8 16 .. . fI • 3. 4. Ana],y.ia or B.,erita Seccmc1u7 and Ind1:r~t· Ertects · • •· · • • • · ·• ·· • • · • • • ·• •· • · • • e 0 21 25 5. ··· · An Ex-poet .A.ual7e1s • · • .. • • • • • • • 6. .lltematiYe InYestllent Patterns • • · · 26 7. Cone lu iona •. . ·· · • • • • • , •· • • · • • · • • · ··• • • ·• • • • · 29 36 III. THE NORTRBlSl'ERN PROJECT • • • • • • • • • • • • · . . • 0 38 1. An Ex-ante ReView • • • • • • • • • .. • • • • • • • • • • 38 2• Ex-P08t B...luation • • • • • • • • • • • • • • • • • • • • 42 3. COI1c1_101111 • • • • • • • • • • • • • • .. .. • • • • • • • 47 .... . "  TABLES Page 1 Eet1Jlated Distr1'bution and Irrigated Land in Mexico, 1968 ... 2 2 Land Tenure in the Northwest . . . . . .. . . . . . . 4 Land Tenure in the Northeast ... ....• • • • ·.... 6 4 Original Appraisal. Costs, Benetit8 and Rate or Ret1rn ·.... 9 S Reductian Factor. Applied to Total Coat. ot Production to Acc01lIlt tor Shadow Iricing or Input. • • • • • • • • • •• 10 6 Mexico: West Coaat Suppor I; and &rtiJAated Trade Prices by' Yeua • • • • • • • • • • • • • • (I • • • • • • • • • • • • 11 7 Estimated Production as a Percentage ot the Value of Production per Hectare tor Supportad Crops by Districts, 1963-1968 ATeragea • • Ii • • • • • • • • • • • • • • • • • • • 12 B Original and Rerlsed Ex-ante Evaluations of the Project by' District and Total • • • • .. . • • • • • . • • • • " • • • •• 1$ 9 Areas b7 Salinity C1asse~ in the Hort_at lTroject ~ . • • • • • 16 10 Public InY•• taent in Rehabilitation & Loan ·..·...· • 27S 18 11 Annual Operating and Maintenance Costs • • • • · . . · . · • 18 12 Financial Return to SJi.H. trOll Rehabilitation • • • · . . • • • 19 13 Cbanges in Prodllction Costs per Hectare • • • • • · . . ..'. 20 14 Projected and Actual I.c1d RebabUitation in the Northwest Project • . . • • • • • • • • • . • • • • . • • • • • • 22 IS Cotton as a Percentage ot Total H'arYeeted Area • . .. • • • • • 23 16 S1ln1t1clU'lt, POliti.... Yield Changes due to nTechnolol1~' • • • • 24 . 17 Internal Bate. of Return to Rehabilitation Inv••tJlent Measured lx-ante and Ex-po.t • • • . . • • • • • • • • • • • • 27 18 ReclaJation ~diture" Relative to Potential Outlays by 1>.la1;r1qt • • • • • • • • • • • • • «I • • • • • • • • • 28 19 Over~al1 Sal1n1t7 Incidence b.1 Districts • • • • • • • • • • • • 30  II .h.blea (continued) lK! ;~o Projected SalinitY' Condition in 1967 without RehabilJttaion ••• 30 '21 A881D118d Pattern of Deterioration and RehabUitaticm tn the Eventl that Rehabilitation be Postponed until 19457 • • • • • 32 ,22 Opt1Jlal Timing ot Public Works and the Cost otharing Undertaken the Project in 1962 • • • • . • • • • • • • 33 ;23 f. Ex-post Cost of Undertald.ng the Project :in 1962 far CUliaean J I -_I harte and MaY'o • • • • • • • • • • . • • • • • • • • • • • • • 35 . . '!'be Changing Role of Cotten in the Northea.t • . • • • • • • " • 36 '2$ < • Original Appraisal Coste, Benafita and Rates of Return • It • • • 39 :26 'Original and Ren.ed II:-ate E.,.1:..t10l'l8 or the Project 'b7 Dtstrict and Total • •• • • • • • • • .. • • • • • • . . •• 40 ~'2r,. Revised Ex-ante Mod.lot 'B~etita to the -ortbeutern &oject " • 41 ld ', Progres8 of Sallllit7 with and Yithout Rehabilitation in Ri;o Bravo Dist~ict 1964-1910 ••• • • • •• • • • • • • •• • • • 45 :'29 Rio Bravo: Net Econdld.cValue 'of FaN Procmctlon and Public Expenditure on Behabilitat:l:cr1 • • • • • • • • .' tJ • • " 45 JQ ot EcODOllic Retu:rn, E8t1llateil.tof Rio ara.o OolrlpC'laOll RebabUitation Modele • • • • • • • • • • • • • • • • it, • • • • ,46  FIGURE ,'r PalD'8 l .:.::a: 1 Derivatian of the C08t of a WrOlll Dec1llian trca Present Values • 34 _. I II .p or Mexico - Irrigation District LocatiCll8 Public Vorb Expenditwe III Areas Attected,b,y Salinit7 IV Financial Returns to S.R.H. V Yield Chan,.. Fre. of Area Eftects VI lx-poet Oost., Benefits, Retums VII ".aw.d Oa.tca. or Delayed BehabilitatiOll VnI The Cost of Wrcmg Dec:laions IX District Data I Shadow C08U of Production ..  S~Y A1~ CONCLUSIONS i. This study was developed as part of a program of work designed to assess past success in proJect appraisal as well as evolve improved appraisal procedures for use in the future. The choice of irrigation re- o habilitation in Mexico was deter1Td.ned by !11Q.Y i·actors including relevance to future lending, ease ot communication and existence ot sophisticated data. :i.i.. The i.mportance of i.rrigation in recent develo}ll1ent experience in Mexico is hard to over-estimate. Great reliance has been placed upon it as part of overall agricultural policy which has l.ttd in turn to the deTelopment of serious social and economic inequities that have recently begun to prick the political conscience of the country. . 1U. The study deals with two projects financed, in part, by loans 275 and 336 in 1961 and 1963 respectively. Loan 275 tor four districts in north- western Mexico continues to ap~ear satis£actory in prospect even after the original appraisal figures are subjected to caretul. revision. The lowest return appears to be about 8 percent, on the investment undertaken in the Ya4.ui va.l.ley, while returns in other districts range up to 18 percent. Loan 336 covering three districts in the northeast has not .t"ared so well. Here returns seem consistently1 but predictably, low. iv. For the northwestern project, examination of what actually occurred shows that while costs were redistributed among items, essentially the same kind of work was undertaken as planned. The IDOst notable ex:ception is that privately l"inanced, on-farm works 1'a.Lled to materialize. Overal.l about haJ.f as much land 'Was reclaimed from i'tis saline condition by virtue of the public works as was anticipated. Other districts were more successful, but in the Yaqui valley the only impuct was to prevent further deterioration; no re- clamation as such was achieved in the period 1962-68. The project admin- istrative costs, while not exceeding 10 percent of total public works were, nevertheless, three times those expected at the time of the original appraisal. v. Taking a purely financial view of the project, it wias found that the added water and reclamation charges levied on far.mers would represent an insig- nifioant or negative retum, on the capital employed, except in the Yaqui Valley. vi. The most feasible alternative form of investment appears to have been an essentially similar project delayed one or more years. Such an altel'- native is postulated for each district and the value ot postponemSlt given as a cost of having deoided. to start ')lork in 1962 although some lat"r year might have been better. It was,\ found that in CUliacan, Fuerte and MayO the cost of starting in 1962 rather than later was only substantial for ver,y low discount rates such as might be appropriate tor so01al time preterence discounting. For interest rates approXimatj~g returns in the private sector after taxes such as are commonly used in, Bank disoounting" the costs are insignifioant. ~or the Yaqui Valley, postponement of an essentially hol~lng action implies permanent reduction in the area cultivated and this is a possibil.ity that  - ii - should be examined very carefully in the marginal al:"eas, especially as other public works under the Northwestern Water Resource Plan (PIHINO) make new areas accessible for the water of the Yaqui river. It is shown that even a program of work superficially as successful as those in some other districts would not have paid off in Yaqui since prior development had moved so fast that there were fewer benefits to be captured from the inter- action of reclamation and improved farming practices. vii. It is concluded that great care needs to be exercised in determining the value of rehabilitation on marginal areas in Culiacan and the continuation of a staying action in Yaqui. For future studies, the importance of having ra.ndom sample surveys of the area to be recuperated is most important. The present study contains many sources of inaccuracy. since none of the districts in the Northwest Was able to supply data in which both costs and benefits could be disaggregated to a level lower than the entire district. While much such data is currently recorded for every farmer, the means for its quick sampling or summary presentation do not exist and a great deal of work on its original recording is wasted. viii. !Dan 336 in the Northel!stern section of Mexico has been given more cursol"Y treatment. 'This has been dictated by data deficiencies ,:as well as the fact that the project is not yet completed. Two of the three districts involved, Rio Bravo and San Juan have passed through a period in which cotton, once dominant in the cropping pattern, has virtually disappeared under pres- sure of disease and plague in combination with a pricing policy biased in favor of grains and a radical change in the availability of mi gratory labor fo:£" cotton picking. It has c.ommonly been held that it was not possible to predict this great change in the cropping pattern which has had a severe impact on the overall value of farm. production from the area. This study shows that it was only necessary to recognize the extent to which cotton had already been displaced to have concluded that the project had a low social return. Using data available prior to start of construction, the returns are eSitimated at 12 and 7 per cent per annum respecti"'!~ly for Bravo and San Juan. Such returns are low for a country such as Mexico and other factors would have been necessax~ to justify the project as undertaken. ix. Recasting the appraisal in a truly "w1th-and-without-the project" framew,ork and using historic data that reflects cotton t s disappearance does not alter these conclusions. Information limitations have dictated that we do this type of analysis only in the case of Rio Bravo. x. The most important defect in the two appraisals reviewed here is their complete silence on the economic motives and institutional changes that were supposed to lead to private, on-rarm, investments in levelling and drainage. That no motive has so far presented itself and/or that the insti- tutional pr.)blems related to credit have never bem resolved is testified by the fact that virtually none of the expected private investment took place. Thus the scope of the, project has been less than was anticipated at appraisal. xi. Some lessons have been learned for the conduot o.r future stUdies of this kind which ~ve implications for the data collection system established ! ,  - iii - at the start ot projects to be subjected to later evaluation. It seems im- probable that it would ever be worthwhile to undertake exhaustive ex-post analyses of all Bank projects, but the value ot such exercises tor practicing appraisers is sutticient to warrant a careful selection, in advance or i - 1" construction, ot projects 'GO be given this treatment. In the case of the , present study much ot the most desirable, disaggregated information was known to exist but its retrieval cost was prohibitive. This arose tram the tact that too much information ~s being recorded in certain districts and I • ! in a way that did not lend itself to later dissection_ In all cases well designed samples or the farm population would have provided ample information and required tar less busy papen.-ork than is currently undertaken in SQlne districts. The Water Resource Ministry, Secretar1.a de Recurso,s Hidrauli/;~os (S.B.H.) should have a small evaluation team headed by art economist ab:t.f.i to advise on the most ,efficient torm in which to arrange data collect:tCiis-~ xii. Our own choice would be to arrange tor collection or private investment, production, cost and sale price data to be obtained trom a random sample of farms to be rehabilitated and those unatfected - stratified by major rehabilitation differences ot soil quality and proposed timing of work, etc. Accounting records of project outlays should be kept in such a way that 'the regional application ot funds by type ot work would be readily available on, at least an annual basis. wath such material available, statistical analysis could be used to distribute the value ot production gains between those that would have occurred had salinity persisted and those truly attributable to the rehabilitation project. xiii. Fori those Bank projects selected for detailed evaluation at a later date, the most important detj.ciency in present Bank appraisals is a statement of what the anticipated development would be without the project. The impor- tance of knowing what the next llOSt likely alternatiTe was at the time of decision cannot be overstated. This is the only way in which the success of the investment actually undertaken can be measured and clarity on this point is vital to the design or an appropriate data collection system. xiv. For Bank projects not selected r~~ evaluation in depth, it would be desirable to extend supervision missions well beyond the construction period and loan disbursement. In many instances this happens through tollow- on loans, but the principle should apply in all cases. Supervision missions should get involved in undertaking the studies necessary to rectify deficiencies in the original appraisals. For instance, in the projects studied here, the outstandin~~ problem which could have been evaluated was the assertion that farmers cQ~ld and would undertake private drainage and levelling works. xv. "Since the time at which loans 275 and 336 were authorized, Bank appra:isal2) have improved enormously and, apart from the lack of specification of a real alternative, this impr'oved basis will make supervision reporting of more recent loans very much easier. The main task of project supervision is not so much to ascertain that the funds have been spent as intended - the Bank1s audit system.is most efficient - but to detel--mine whether the envisaged benefits have occurred and it not why not. This is always the more difficult question to answer but also the more meaningtul.  I. INTRODUCTION AND BAOKGROUND 1. The Purpose of the study 1. In twenty years of lending for development projects, the IBRD has had the opportunity to' observe a wide variety of public and private investments reach substantial maturity. Continuous contaot has been main- . tained during the early developmental. phases of each project to ensure that the engineering aspects were carried out as planned. At the same time, some general indices of profitabi:t.i ty, repayment capacity of Bub-borrowers and physical prOductiveness of the industries or regions affected have routinely been noted. Where troublesome situations have arisen, the Bank has provided considerable teChnical assistance and financial support to 'save the ~oject. But where positive results have been evident, there has been a tendency to regard the project as a success and close the tiles. In some cases this might Occur within five years of a loan becoming effective. 2. Active concern withiirl the' Bank: about project failure is natural and an indi cation of the reality or the Bank 1 a function as a development agency. The failure of. ~"~ project, after all, is onl7 tenuously linked to debt servicing capacity of the recipient government. Neglect on the part of the Bank to learn from successful projects, on the other hand, is most unfortunate. Much might be learned of the development project and its effective pla~~ing from a careful study of project successes. There have been two previous forays into the .area of Bank project re-evaluation. The fi":'st was,la study of a road pro,jiect in Ira.n by H. G. van der Tak and, J. de Weilie J/ and the other waS Albert o. Hirschman I s review of various Bank i projects using a rather larger canvas foj;" his approach.2/ The present study has been undertaken in the belief that further opportunIties for comparing the relation petween predicted and actual costs and benefits even for the ear~ yaars of projects will h~p in the specification of future projects. It may also lead to insights that will improve the general methodology of project appraisal. 3. Since about sixty percent of recent Bank loans in the field of agri- culture have been :for irrigation works and there are prospects for considerably more lending in this area it seemed appropriate that we should select a past loan or loans in this area for. an evaluation study. Considerations leading to the ultimate selection of two loans to the government of Mexico included: ease of aCcess for multiple visits and rapid communication t.romWa~h1ngton, D.C.; the existence of a relatively sophisticated data oollection service within the Water Resources Ministry (Slt-R.H.); superficial differences in the success of ~he .first two loans made for reh~bilitation works and a degree of familiarity with the Mexican agricultural sector on the part ot the researcher. 1/ Reappraisal of a Road Project in Iran, World Bank Sta.f!.' Occasional Papers - Number 7, Johns Hopkins Press, BaltimOre, 1909. y Dev8?<~eent Projects Observed, Johns Hopkins Press, Baltimore, 1968 ..  - 2 - 4. It should be noted that both the Bank loans reviewed here were for the rehabilitation of existing irrigation system s. They did nothing of themselves to add to the physical irrig,able area though they have, in con- :1UDct,ion with increas'ed administrative efficiency, in some cases led to increases in double cropping. Their principal obj ecti va, however, was ·to recover areas that had "diminished yields or were·entirely out of production because of salting attributable to inadequate drainage and," or permeable canal structures. ~. For easy reference to the two developments examined, they will be referred to as the Northwestern project and the Northeastern project. The Northwestern project was for the rehabilitation of three irrigation districts under S,R.H. and the irrigable +~ds of the Rio Fuerte Commission - a rEJf~i.ona1 Qevelcpment agency headed by S.R"H. nominees. These four areas will be referred to hereafter by the ruWl(3\) Culiacan, ,Fuerte, Mayo and Yaqui. Rehabili tation was financed in part by I.B •.~.D. U)an 275. The Northeastern project covers works in three S.R.H. districts ylhose names will here be abbreviat~d t.o Bravo, San Juan and Delicias. Rehabilitation works in these areas were partly financed under I.B.R.D. Loan 336. The locations of the districts are shown on the map at Annex I. 2• The Role of Irrigation in Mexican Agri culture 6. Because the past deca,de has seen Mexico move from being a substantial net import\9r of food grains t9 being a net exporter of wheat and maize, the country haf:' fJreen widely regarded as a prototype for development elsewhere. All who vi~ilit the country are impre'~sed by the du~lity existing in agriculture. ~Jhile the 11rast majority of the rural population eke a meagre living from small / holq,ings in the Bajio and other upland rainfed areas, about ten per cen:t of farmers arfi3 located in the more prosperous irri gated districts. This duality also showsl up in the oistribption of production among farmers. The 1960 census of ;agriculture showe.d·that ,0 per cent of the fanners'produced 4 per cent of the total; 84 J?er cent pr~uced 21 per cent and, at the other' extreme, 3.3 per Cf;mt produced 54 per cent of output. 7. Since 1926, responsibility for the development of irrigation districts has been With a centralized government agency (more recently a ministry - S.R.R.). Recent et3timates of the area unQer irrigatiqn are given in Table 1. The 2.5 million hectares administered by S.R.H. ars enclosed in large vall~y complexes having j;mpressive local offioes and well trained engineers and administrators in charl~e. These areas represented 15 per cent of all cultivat(~d land in 1967 and yielded 40 per cent of total output. Condi tions for j,rrigated agriculture are virtually ideal on the north-western coast of Mexicoe The local rainf'all is usually well b.elow 600 rom. (2L inches) and heavily con- oentrated in the months of July, August and September. Even for tl1ese three Table 1. Estimated Distr~but.ion of Irri;ated Land in Mexico 1966 .' , (Million has. Within S.R.H. Districts 2.5 other S,~pplement.ary"Systems 1.L Total Irrigated 3.9  - 3 - months evaporation exoeeds preoipitation and for the year as a whole is consistently in excess of 2000 nun. Irrigation water is supplied by the rivers having their sources in the western scarp of the central plateau. For the past twenty years moat o:t these rivers have been under partial or ... .. canplete oontrol with sufficient storage to allCJW for seaond crtopping in certain areas as vater use etticiencies have improved. 8. other irrigation districts are not so favourably endow\,d and Rio Bravo and Rio San Juan of the northeastem region on the Uni ted S,t~ates border are in this class. In the first place, the historical settI-ement pattern was more dense since even the early diversionary systems dependent upon the run ot the river penni tted irregular but widespread irrigation. Wi th the formal coordination ot water control under agreement with the U.S. it has been exoeedingly difficult for the author! ties to continue supplying water efficiently to so widespread a district. The climatic aspeots are also less favourable. With the exception of years in which Carribean cyclones result in devastating floods, the annual rainfall is in the range LOO to 700 mm (16 to 28 inches). Except tor the normally heavy rainfall in Mayor June whioh may exceed evaporation, other months ot the year frequently receive 1.,0 or 50 mm (l.~ or 2 inches). This relatively uniform distribution of rain has combined with the maqy small land holdings to produce a situation in which the labar cost of irrigation is a barrier to the otherwise optimal use of irrigation water. Many small farmers prefer to gamble on the rains than pay the variable costs ot an irrigation. The administrative headache that this situation gives rise to is retl~cted in all sorts of uncertainties for those making regular use ot the water available. 3. The Northwestern Project· 9. This project was envisaged as the .first in a series designed to rehabilitate and, to some extent, canplete, 22 irrigation districts in Mexico. Although the Bank appraised the project most favourably in Januar,y 1961, work did not commence until 1962. From then' on the loan was disbursed over tour years inatead or the three originally expected. Bank finance ot US $ 15.0 milJdon represented 35 per cent of the public works cost estimate and was equivalent tc the foreign exchange component ot these costs. The complementary private, on-farm works envisaged as necessary l'r.~re to be borne by the farmers directly. No special credit facilities were set up tor them. The basic objective was to recover 79,000 has. of land out of production because ot i'ts high salinity and to restore to tull production 158,000 has. having reduced yields. The means to this end were new and improved drains, canals ". and water control structures. Other items included road improvement, tele- phones, dwellings tor the water-masters (caneleros) and construction equipment necessary to the main.tenance of the projected works • .' 10. The land contained by the four districts was farmed by private p~operty holders, ejidatarios and colonos. ~idatarios are benet.iciaries under the Mexican land reform and colonos are small-holders settled in groups through colonization land grants. The most important distinction betw(3en these classes'trom a production pOint or View is that the ejidatarios have had ,in general, les8 well developed credit tacil1 ties at their disposal.  - L- The size ot farm' varies greatly by tenure class and there are some big d"1tferenC8s between districts' u· Table, 2 shows. It is- most noticeable tha t all.. olasses ot, tarmer in the- Yaqui valley are provided wi th relatively large Iota ott average. In interpreting these figures it should be remembered that Mexican ~av does not:' permit an individual to own TIlore than 100 has. ot wi • irrigated land. Family cooperation and other devices nevertheless make it possible for one man to be the administrator of very large farms or Leo to 500 has. and more. Table 2. Land Tenure in the Northwest District Tenure %ot Farmers lia. per Farmer Culiaoan Ejidal 26 $.L Private 7L L2.' Fuerte Ejidal 60 7., Private 32 29.L Colono 8 U.8 Mayo Ejidal L9 6.L Private 51 11.9 Yaqui Ejidal 36 lL.L Private 51 50.) Colono 13 20.6 11. The tradItional crops ot the are. have' been cotton, sugarcane j r-4.ce:j wheat, maize and beans. In recent years., except tor sugarcane and 'Wheat, the others have given 1//&1 to 15011.' new cr~'. SorghUDl tor cattle teed, 8.oybeans in rotation with wheat (double: cropping") and a host ot others. F.t'll1t .U\d' vegetable production baa received a b:Lg il1lpetu$ since suspension 01" tlUl', United States migratory l~or' progr.... Under this pr'ogr. . MexLcan nationals worked on a seuonal buis" ban'esting frai ts and vegetables In the soutnern U. S. Now capital and marketing' knowledge bas ccae esOlltb or the bWder and combined with lQcal inter-elts to da~lop a large .cale and highly ett.1oient export operation. Taaatoes wer, the leader in thia but the i~U"ltry is diverSifying'. u,cal. wh8.t~oduct1v1ty, has be.n grea.tly assisted by, the work done by the international naaile and wheat relearch Cetnter (O".I.M.M.Y.T.) established;, and' tor many yell's maintained by, the Rockteller . Fo.umation. Hybrid maizes vere' also important to the irrigated d1et:rict8 though most ot Mexico's maiz$ is gJ:'CMl under ,rainted conditions and little ,teswar.ch baB been directed towu-ds breeding' tor these harsMr tnv:l.:ronments. i3:be! has' declined in area ,as vat.8r 8tor.~ capacity has expanded to permit doable @opp1ng and a8 new areas ha'At, been opened up adjacent to the traditional rt'ce,arsas. Rice's. high water reqUirements have made it a 188$ interesting crop,;.•8 the possibillties tor more pJ'od~ct:lon ot crops vi tb lower requirements  - 5- have become possible. Although world cotton prices have been declining steadily, Mexican cotton has continued to obtain a good price. Medium to long staple combined with the select!vi ty arising tram hand picking have been important tactors in this. Nevertheless, cotton has declined in area and persists in doing 80. Many factors are involved. Costa are rela ti vely high and yields variable in comparison to the tood and feed grains which have en.joyed stable goverment guaranteed prices for many years. Diseases and pests have been i~creasing1y difficult to eradicate _dding fUrther to costs. One moat iIlportant factor bas been cotton's long grOwing period which, under the given pattern of rainfall, pl-ecludes the double cropping ot cotton land. The local research station now places great emphasis on obtaining more precocious cotton varieties adapted to the local conditions. Researchers there teel that something 'fIractioal could be expected ot this work in another tear years or so. The cropping pattern, yields and tarmgate prices betore and atter the rehabilitation are given in Annex IX. l2. A visitor to the Northwestern area has the impression or visiting a very afnuent fanning district. Almost all operations have been mechanized for a decade on all but the smallest farms. Machinery dealers regard this as a "saturated" nsarket. Apart tram Cullacan which 'Was an old staging t6wn for west coast coache.fb the other towns are new and shout their novelty in neon signs. The well-to-do have their clubs which are often sumptuous, benefitting from the nair "With which Mex:tcan architects design am show themselves adaptive to the local climatic and ute-style condi tiona. Lite here is good tor nearly everybody and there seems little chance of that good lite being threatened. But this is not Mexico. The four districts covered by this project give a Ii velihood to one-halt of one per cent ot the Mexican population. It is no accident that the people are healthier, better built, more relaxed and able to indulge in cheerful social banter - they are among the elite ot the country tor even the ejidatarios ot Cullacan wi:th the smallest plots ot all can 11ve well contracting with a local sugarm111 tor the production ot cane on their land. L• The Northeastern ProJ ect 13. As can be seen tram the map at Annex I, the northeastern project encompassed works on the two adjacent areas ot Rio Bravo and Rio San Juan as well as the more inland area ot Ciudad Delicias. This was the second stage in the proposed program ot rehabilitation works and was designed to recover ~3,OOO or 19,100 hectares out of production in 196) as well as restore to full productivity a further 27,300 hectares seriously atfected by salinity. The works proposed to achieve this end were very similar to those tor the north- western area. Indeed, the second project was formulated so soon atter the first became ettective that there was no time to learn tram eXperience with it. On the other hand, having established lines ot comntunica.tion with S.R.H. under the first loan did result in more expeditious authoriza.tion ot the second. The Bank loan ot US $ 12.5 million was to cover the foreign exchange component ot the public works cost and represented 38 per cent ot these costs. In addition to the public works, it was antiCipated that private, on-farm w9r ks costing 82.3 million pesos (US $ 6.6 million) would be needed to supply al?propriate land leveling and drainage wi thin the farm plot. The public works were expected to take live years in Bravo and tour elsewhere with private investments lagging about two years. The project began in 196~ one year  later then anticipated and the work in Rio Bravo atter six years is still 'incomplete. Work under the loan vas completed in the other two districts ,on :schedule. As in the caaecltthe northwestern project, none ,of the tpri¥atelnvestments took place butPllblic works contingency i tams have \lMrmitted the construction of deeper drains t'han originally provided tor lftiic'b ha\re, to sane extent, been substituted for smaller, private drains. ~1:L. Land tenure patterns in Dell eias and San Juan are rough1y 1Efimilar but in Rio Bravo, as Table 3 shows, relati vely more or the farmers Jare ejidatarios. Here the larger size ot the private farms still ensures that most or the land is privately owned. Table .3. Land Tenure in the Northeaat District Tenure % of Fanners Ha. per Farmer Bravo 'Eji'dal LL 10.30 Private 56 26.10 San Juan Ejitial 2L 8.65 Pri'wte 7L lL.95 Delicias Ejidal 31 6.59 Private 6) lLf'O ;"'."""1 , '15. The single most important crop in .this area has been cotton and the ',success ot the rehabilitation project was largely predioated upon its cOfttinlled importance in the cropping pattern. In tact, ever since the early ;1960 1.8 the area planted to cotton has been declining with maize and sorghum ''1-',&8 increasing to take 1 tsplaoe • While cotton continues to have an ~:lmtSOrtant role in Delicias where, 'due to elevation the climate is drier, in ~Br.vo and San Juan it has vi'rtu"Uy difJappear4!d under the double pressure o.f:ir~8ing coats to fight pesta and ,diseases and decUningrelative profitability :ih:the face ot high support prices tor other crops. A very 'lew hectares l',haVe' been ventured in sane new crops, -especially wgetables,andsoybeans. In ;Delicias,wheat has held its place, but ,otherv:1se the cropping pattern is dominated by maize and sorghum • .16. In contrast to the,(NOrtbweBtern~area, the districts rehabilitated :'i,tlthe second project do not ,give eJtt 'Iany ·aura 0'£ arnu$noe. rarm size is ~n'ller in the private ·seotor, the Olimate leas predictable, ,and service ,:hrtdu$tries have grown slowly in the town$. The 'ejidal oormnunity has been :lunable to secure credit ~tor their opera.tions 1f1thout using the bureaucrat- litdlilly Ir:l;gid Banco Ejidal whieb hasluld tootew technimaans to,ope~ate as "1\aMger to 'so many smalll(tat'lns. The '<repeated 'fa1lure ot cott,~n crops has led ~to~·.n air or despondency :and it seems" 11:kely that it will be ~)everal year's h\ffeore the rehab ili tat:!Jon project will encounter a climate of public policy ,~'titldn which its full 'tru1 ts oan be borne.  - 7 - ?• Evaluation Methodology and Outline 17. This study moves in ea~ stages tram a description ot the original project and the basis on which it was appraised to a new evaluation incor- porating the most recent Bank practices and taking advantage ot ldndsight. The principal methodological issues covered deal with the inclusion ot time in the analysis, the "shadow pricing" ot inputs and outputs and composition ot a representative cropping pattern. 18. Because ot the better data base and considerations ot time, the study of the Northwestern Project is deeper and used to explore ar~9as not open to us in the Northeastern Project. So tar as proved possible, however, the same philosophy has been tollowed in each study. The first steip in the evaluation consi.ts or recasting the description ot the project e:x:p,ected costs and beneti ts in a torm that permits calculation of an internal rate ot return. Doing so results Inreturn estimates that are as 11 ttle as one-halt the nominal returns on which the projects were originally justified. The second step consists ot imprOving data used in the original appraisal in thtl light of what was known but not used prior t,o authorization of the proj ect. At the same time, methodological questions concerning the choice ot a represent- ati ve cropping pattern and pricing ot inputs and outputs are dealt wi t,h. Incorporation ot all these retinements into the analysis leads to a ne'W' ex- ante model ot the projects .trom which a revised internal rate or returtl is deri ved • F.l.nally', abandOning the ex-ante standpoint but adopting the best ot available methodology, a new evaluation is undertaken on the basis of. his- torical outcomes. This expost evaluation still has at its heart the sam,e appraisal model that was used prior to authorization or the project. FOl- lack of a sui table detlating index, all historioal values are lett in current tems. 19. In the case at the Northwestern Project, the analysis is taken one step further. The objective in this case is to decide whether, in the light ot the historical pertonnanc8, price changes and so on, it would have been more appropriate to undertake the investment at a later point in time. It is at this stage that we draw on the 11ttle evidence available to show what kind or development would have taken plaoe without the rehabilitation works. Using this model, values have been derived which represent the present 1fOl'th (in 1961) of the cost ot having undertaken the rehabilitation starting in 1962 rather than in a later year. 20. Chapter II or this study deals with the Northwestern Project and Chapter III with the Northeastern Project. Each chapter is selt-contained,. -. having its own section covering conclusions relevant to the specifio project. At thfl same time, the methodological, issues are explored in greater depth in Chapter II and there is only passing reference to them in Chapter III.  - 8 - II. THE NORTHWESTERN PROJEOT 1. An Ex-Ante Review 21,. The original appraisal model used to evaluate this project WaS qllite typical or those employed by the Bank for irrigation investments at the time it was undertaken (1960). The .f.'1rst feasibility study was done by the Mexican authority S.R.H. Although the documents are now unavailable, :l. t appears that the Bank accepted most of: this study, only adding contingency allowances where it saw tf t. This in no way shirts the onusot responsibility tor the appraisal, but does serve to show that S.R.H. had, even at that time, considerable capacity for conducting feasibility stUdies which met the Bank's criteria, 22. The benentsincluded in the original appraisal model were those ...surable as changes in the net tam incane accruing to areas rehabilitated Iraq their saline state. Quantification ot these bener! ts was done using c~ent farm gate prices. Yields were expected to rise aboutLO per cent a8 a result or rehabilitation on ateas affected by salt but still in produo- tion. Areaa then out ot production nrae aSlNlfted to perform equally well as otl1..~ areas atter rehabilitation. It was aSB_ad that increasing production COlts on the low yielding lands would be offset by reduction in tillage e:xpenaes atter rehabilitation. Cropping patterns vere those operating in the late 1950 1 s and cropping intensities ot 100 per cent were assumed. 23. The costs taken into consideration in the original appraisal Qo"red both public works outlaY's and private, on-tarm, investments as well a. expected changes in operating and maintenance costs once the works were ".plate. Current production costs on the areas recovered trom being out of production were estimated at about 60 per cent ot gross t&rll income. 2L. Theee quantified coata and benefits ....re ~iaed in the form ot a tull-development-rate-ot-return (rna). This 1s simply the anticipated change in net tarm income less the change in operating and maintenance Coslts ......ed as a peroentage ot the total public and private investment under- ",ken_!/ The results are given in Table L. .. .~ lI''J.'he , .J.B.. would be identic.l totbe Internal Rate ot Return in the case ~ot 'an. inve.tment tor whicbJ (a) the benet! t .tream was constant over time .nd ot intinite duration, (b) beneti tsstart at the instant investment i8 Mde ;~ (c) the inTetJtment occl1r. at one point in time.  - 9 - Table b. Ori~nal A~a1Sal Costs, ,Benen.ts and Rit£ot Gum (fiiues ift iiltton ot Pesos) It. District Total Cullacan Fuarte MaY'o Yaqui Loan 27, Costst Public 120.0 2,).0 10la.O 3,.0 512.0 Private 27 •0 119.0 2~·6 7L.L J02· 0 Total 177.0 372.0 158.6 lO9.L 817.0 Net Farm Benetits 35.9 1,6.6 2,.8 20.1 238.L FltD.R. 20~ , L2% 16% 19% 2~ 25. Wi thout departing tram an ex-ante viewpoint ot the project, the original appraisal is open to criticism on both methodolOgical and data grounds. It is very much to the credit ot the Bank's project appraisers that current prectice on similar projects would meet most ot the criticisms that can be made. What tollows then can be interpreted as an fndication ot how tar Bank appr&1sa1 methods have come in the past decade. A critIque ot the methodology applied in 1961 must discuss output pricing and the costIng ot inputs; the rate at return concept usedJ cropping patterns appropriate to rehabill t.ted land J the economic basis tor pri vate investment and how to measure the benet.l ts arising from prevention ot further saliDi.ation. The nrst point concerns the use ot tarm gate prices and input costs in the ori ginal appraia1. In gener'al, it has becane Bank practice to approximate the opportunity cost ot inputs and outputs by their price in the tree world market. On tbe input side, in the case ot Mexico, proximity to the massive Un! ted States I :market makes world prices in fertilizers', tractors and gr-dns ot very real importance. In revising the appraisal, these two major traded inputs and maize, wheat, sorghUJll and soybeans have been valued at c.i.t. import and f.o.b. export prices instead ot the administered prices used in the original appraisal and subsequent project supervisions. Tbe procedure adopted and its impact are described in the tollowing paragraphs. 26. Fertilizer materials represent apprOximately half the costs attributable to tertilizer am ito a'pplication. No such breakdown was available tor machinery services and an arb1trary t.lgure of one-halt the cost ot land preparation has been used to represent this item. The 1969 agricultural sector review conducted by the Bank lIconcluded that fertilizer had, tor some years past, been about 60 per cent more expensive in Mexico than would have been the case with tree trade in this item. 1/ "The Agricultural Economy ot Mexico", IBRD Internal Document, December 20, - 1969.  - 10 - Table ,. tlleduc,tion'Pact .re, . ;,t!u~:tiion :i.to: ,f' ' .- District Cotton Other GrQpS Cullacan 0.885 " Faerle 0.882 ·.Mayo 0.885 ~I.qu1 0.865 .. : ~.L·_ " "'" ..... ' "'ft'"  Table 6. Mexico: West Coast Support and Estimated Trade Prices by Years (pesos per metric ton) CROP Maize 'Wheat Sorghum So.vbeans Harvest atppor-t Export Support Ixport Support Import Support Impart Year Price Price Price Price Price Price!! Price Price 1960 800 601 913 901 525 502 1961 800 751 913 80~ 525 502 1962 800 900 913 829J 550 482 ~Q J',upport l"rice 1963 940 453 913 857 SSG S4S -iii these yeB.l"s 1964 940 704 913 776 625 553 1965 940 842 913 760 625 553 ,.., 1966 800 686 913 15021 625 553 1600 IS88 ~ ... 1600 a.. 1967 800 724 913 141bl 625 611 1425 1968 800 649 913 741- 625 509 1600 1388 a/ Based on the ratio of caasuPO £oreign sales price to series for U.S. sorghum o.1.f. Rotterdam, - U.K. in years 1967, 1968. bl Interpolations. - other supported crops for which trade prices were not much different from the support level or for which data was unreliable were Rice, Se same, Safflower and Beans. Source: Anpario Estadistico del Comercio 'Ext,riar de los Estaclos Unidos Mexicanos.  - 12 - Likewise tractors were found to be about 30 per cent more expensive. On ltbflse 'figures and detailed 'a~lysiiJ '.of the importance of thel!!e 1 terns in 14Q:nqp }h.u,tlgets -tor each district, costs have been :adj'Usted downward using ttbe i'.etors shown in Table .5. Details are given in Annex X. ;£7. For product prices, only products with prices supported directly \itlmo:q.gh the price support agency, CONASUFO, have been taken 1n"toaccount. r.mhe ~relevant support prices and comparable export prices are .g1 ven in Table 6. EXpbrt prices for maize and wheat averaged 8L and 88 per cent respectively of 1iheir domestic support levels over the 1960-68 period. Import prices for .sor,ghum and soybeans were 91 ·and 92 per cent of domestic 'sl.Jpport prices m:e'EJpecti vely • ;·aR. The relati ve impact of shadow pricing both inputs :aneJ ,ou·tputs in ;>tbe case of the price supported crops can be seen in table 7 where costs are $~essed as a percentage of the value .of production per hectu-'e under both l.o.omestic and shadow prices. T)le constancy of these percentag~s in the case ~f 'sorghum, soybeans and wheat 'implies that domestic price arrangements, '.b1l1e making these' cr.ops absolut,ely more profitable than otberwise, did not \m8k~ any oneo! ~them relatively more 'profitable than anoth~. On the other \hand;, maize would have had substantially higber relative costs under the }sna(low pricing regimen. Thus 'it would have ibeen., not only abso111tely less nr0tr! table, but less profitable ,relati'veto other price sUpPOrted crops. I.mm 'ltbis extent it can be said I,that or all the supported crops.,.tze was ~e'IgM!:c1ally favorably treated '. ;29l. The second point '()tmethodology concerns the rate ot return concept. imhe '-ene used in the appraisal i'8 der! ved without any specific r'eterence to 'ttihetime that lapses between commenoementand completion or the project. It :i~ !UWadays common practice to 'utilize·the internal rate of re,turn (IRR) in auk ,appraisals. This return is the discount rate which equalizes the time :.t~eam ·:or costs and benefits ,pt the project. While it still su·t.ters all the ~efl'.ciencie8 or other rate ·CQno.pts wnenabsolllte values are also important, '!i:1t f.,does have the advantag~o'r ,s:\'mIIIfari'zing ,quite diverse investment possibil1 ties "bl one comparable dimension. ;,O\lr _vi'8iono~f the original ,appraisal starts !~1 'casti'~g it in a .~1me path ,mold. Table 7.\ •. w .Yagul Crop Shadow Maiz,e ~8 62 61 '13 ~O ~9 ~9 ,9 Sorgh'Qm ,5 56 ,7 61 60 64 6~ 66 Soya 60 60 60 61 1,6 46 $6 ,3 Wheat 1.t6 L8 15 78 66 68 10 71 \\  - 13 - JO. '!he third point concerns choice of a cropping pattern to be used in evaluation ot reclaimed areas. It is important to understand that one is working with an area in which mree broad land classes exist. There is land which is, and always will remain, unaffected by salinity. There is land lIhich is as yet 1lllsaline but could become so and there is land which has already become saline to some degree. HBhabilitation puts more land into the first category. ;bat is most important for the argument which follows is that about one-third of the irrigable land has always been sufficiently well drained to fall in this top grade of land. The original. appraisal appears to have anticipated cropping patterns on rehabili tated 'lar..d that were similar to those observed, on average, in the districts before rehabilitation. While this may in fact have been the outcome, one needs to ask whether the production of high-valued. fruits and vegetables, especially tomatoes, would not have occurred wi thout the project. It is maintained here that the inCEllti ves were such that these crops wuJ.d have been grown to meet the market demand with or without the project and that, regardless of actual location of the various crops, the effect of the project was to expand areas available for the more marginal of them. The same argument applies, in differing degree a , for all crops but we have chosen to detil only with the easily identified frots and vegetables, ex- cllXiing them completely from the calculations used. to derive production values on either saline or reclaimed areas. That is, we deal with these crops as it they were alv~s grown on areas outside those affected by the project. 31. Fourth is the question of the basis for the expected level of private inves~t. An economic view of the matter suggests that the return on such an investment would not only need to be conpeti ti ve wi tb. the cost of capital to farmers, but that these returns should be capable of appropria- tion by the individuals. The appraisal simp~ aBserted that tarmers would make the investments indicated without presenting any evidence for the private profi tability of doing so. In this case , it is extreme13 hard to posit a more reasonable JIDClel since the basis for the original expectation is not known. However, in view of the signal. failure ot farmers to undertake the j nvestmen ts anticipated of thell, there is an important lesson to be learned; lIhich is applicable to all Bank loans whose success is held to be contingent upon a counterpart private investment, nam.a1y, that an examination of the private profitability ot such investments should be a minimal requirement for their inclusion in the total investment package" Our revised appraisal assumes both that these investments will take place and that the)" are necessary for achieving the full extent of expected land recuperatJ..on over a ten-year period. 32. The last of the methodological issues revolves around the question of how fast othor areas would have become saline it the rehabilitation had not been undertaken. This must be viewed as a methodological issue because it eff'ec:tivel.y asks 'twhat would have happened without the project" and this is the very essence of good project appraisal. The original appraisal of this project accepted that the investment would oc cur and only addresses i tsel.f' to a test of the general acceptability of the project, neither asking whether, nor demonstr~ting that it is the best _ong the alternatives open. In essence, it is a nbefore-and-aftertf model not a "witb.-and-without-the project" model. Sa as not to move too rapidly away from the model us9d in the appraisal" we reserve consideration of these DIOst important issues until Sectilon 6 of this part of the study. ",  - 14 - 33. Apart from these methodological aspects of the original appraisal, IlDY. revision sho-gld also cona1dQr the extent to which a,vai1able data was left U~V.~.~9 il1 the c~lculation of f3XPfl9.t.~q. :returns. In th$ case of this proj ect, q~rr.~I.\~ B~ prac;:tice ot frequent uppating of data was not followed with t.~~ r~sult that, by the time the project was initiated, additional inform- ~~tpn·w~s available on yields, the cropping pattern and the cropping intensity. The or.lly obvious error concerning yields involved cotton in Ys,qui and Fuerte. I~ tll~se two districts, the "before" rehabilitation yields used in. the app~~ sal report were hi gher than even the average yields tor those ~eas. In our revision, they have beep. lowered so as to bear about the same relation to aVerage yields as was the ca$e for other crops. Projec.ted yield increases with rehabilitation were lowered proportionally. Likewise cotton was given ~n ~nduly l~ge share in the cropping pattern and this has been adjusted to refiect the 1961-62 situation in our r4avision. In Yaqui and Cu.liacan, their cfPpping intensities of 130 and 120 per cent r3spectively in 1961-62 were $'Ub~tantially different !"rom the 100 per cent assumed in the app;raisal. B.§p~\!se of the difficulty of kr\9.'W1ng whether it was more reasont\ble to expect tb4~· i~tensity to rem~in the S~e or decline with increased. irrigab1e ax'eas, t~~, revision of the apprais$l is done :first with all intensities at 100 per c~~t~ 34 ~ Table 8 p~(3sents th~. C()$,t,Q J b~pe.ti ts, net be.nefit n~8 and inter- ,n~l ~ates 9f return tor the, orig1~l ap~:r.~$al and tor our present revision. 'l'bft. cl~.fferenoes between the ·two s.~ts or c.O$ts tm..d. returns may be summarized C\~ refiecting an adjustment 0.£ th~, e~ot(td chan~ in annual 11e,;t farm income a~ f\tl1 development. The adjustment renects both updating the data used i~ ~~e way discussed in the previous paragraph and incorporation of the many m$tho9010gical aspects dealt with earl!er. Thus, shadow costs and prices n,,'V:~ been used, internal rate..s or ret\l:rn ~e ~sec;1 (ba.sed on. a. 10· year t:r~p.~i.tion and a La year Ufe) J a,nd. h'uits. and vegetables h.va. be.en excluded f'r91Jl: the cropping pattern for recJ.-imed land~ The revis.ion results in an ~.lJnq~i;. i~perceptible change In the retu~n for the project· as a whole and <)~y Fuerte shows nota.bly, l~~ r.,:t\W~$. t~n lIO~ld have ari$en it an internal r~.twrn had been calculated tor th, ori~~l app~i.$.l. On tbe other hand, t\\~ r~~s1o~ oonfirms the gene,1.?a.Uy D)~.~h 19W,r re;t\\~'ns expe.cted in Mayo and Ya,q~iJ both ot which look lik~ dOl)bttul P~QpQ8.:lt1Qn8 under present Bank p.r,i.p;t~ce in a country $~.ch as. }tfPP;Q.Q.. It shou,ld· b. n,oted that it the 1961-1962 c~p.~ng 1ntensi ties a~ei u=S.ed (120· pt1tr Ctnt, in. Cullacap and· 130 p.-' cent in Y~g,"j.) returns in Culiacan and Y$qu:l. rise to, 16 9 2 and 12.3 per cent respectively. 35,~ It is concluded t~t htd tn~ imp~Ql7,ed. &11d d1saggre~ted n,lodel been ~~l.ab.l. at the time of ~ppr~~..lJ. the: reu.tive11 lower returns to be e~.tlg.ted in Y$qui and Mayo., w'olUg" ~:te. b..n JltQr. cl.~ly higbll gbted than was tb,~ c.se wi tb the FDR' s of T~~l~~ 14, 4M t~e,~e, t..,,' ~eas might Mve reoei ved cAn~i,derab11more atten~Qn·.. Aa 'Q., sn.lla.'I'. f~h:I(s. was esp.e.ciall$ ~rarranted i~ th~ case of Yaqui.  '0 .; '~ Table' 8 : ORIGINAL AND F..EVISED EX-ANTE EVALUATIONS OF THE PROJECT BY - DISTRICT AND TOTAL. {Values in H1ll.1ons of ~eS()8) District Total I T E M Culiacan Fuerte Mayo Yaqui Loan 215 Orig. Rev. Orig. Rev. Orig. Rev. Orig. Rev. Orig" Rev. ""'~¥--' Cost of Public Works 120.0 253.0 104.0: 35.0 512.0 ~ i~:i • On-farm investlOOnts 2Z·0 119.0 J9.2.& Total 111. 0 372.0 1 8: 81Z·0 Change in Annual. Operating and ~ Maintenance Costs, 1964 on. 2 .• 2 +0.2 1.6 13.9 27.9 Change in Annual Net Farm Income at full development )8.1 ' 43'.2 166.8i 139.7 27.h;· 27.2 3L~O: 2~~6· 266.3 246.3 " ~: 1 ,-- ~...,_ I . .J- f·· ...... ~ Project Net Benefit ~r, ·.1962 -37.2 -31,.2 -78.1 -78.1 -33.-3 -33.3 . ::2.3.0 -23·0 -171 . 0 -171.0 1963 -SJ.l -53.1 -1ll.6 -111.6 -L1.0 -47.6 -36.3 -32.8 -246.0 -246.0 1964 -S[).l -57.6 -ll3.5 -116.2 -52.8 -52.8 -3L.9 -35.3 -259.3 -261.1 1965 -17.2 -16.1 -18.1. -24.1 -16.7 -16.8 -8.5 -9.4 -61.1 -64.7 - 1966 9.2 10.7 39.8' 31.7 6.6 6.6 . -3.7 -5.0 52.0 46.6 1967 13.0 15.1 56.S 45.7 9.L 9.3 -0.3 -2.1 18.6 71.4 1968 . 16.8 19.4 73.2 59.7 12.11 12.0 3.1 0.9 105.2 96.2 1969 20.7 23.7 89.9 73.6 1L .8 14.7 6.5 3.9 131.9 121.1 1970 1971 1972 .,.;,i 21,,,5 32.1 . 28.0 32.4 36.7 106.6 123.2. 139.9 87.6 101.6 115.5 17.6 20.3 23.1 17.4 20.2 22.9 9.9 13.3 16.7 6.8 9.8 12.7 158.5 185.1 211.8 145.9 170.7 195.6 1973 - 2004 35~9 41.0 156.6 129.5 25.8 25.6 20.1 15.7 238.4 2?o.4 '. Internal Rate of Return (%p.a.) 12.73 14.07 21.28 18.65"' 10.73 10.66 10.13 8.28 16.35 15.44  .. 16 - 2. Analysis of Cost .36. We now begin the process of compsring what the appraisal expected vith what actually took place. '!his section deals with both public and private expenditures undertaken in conjunction With the rehabilitation pro- ject. The cost of the public 'WOrks and the extent to which uni.t costs and .- the composition of the investments changed from 'Wha.t was expected at appraisal are dealt with. first. Next comes an analysis of how operating, maintenance and other current outlqs in the rehabilitat$d .areas have changed since the " . public works were begun. Thris leads into a statenent of the financial re- turns to the project. These retlU'ns treat S.R.R. as an autonomous project operator and show the extent to which direct water charges have contributed to rep~t of the capital works. The last part of this section deals with the wq in 'Which productioIl costS,i the main component of private project costs, have varied over tim.e in the four districts. As indicated in paragraph 18 above, all actual costs and values are :in. current terms. This should be borne in mind in making cOllJ~arisons with appraisal. projections though the problem is not conSidered a major one for the short and economically stable period under review. 37. Of the original 48jr ,,6 million pesos of public lIOrkS expenditure provided for in the apprais,ll report, ~.1 were itemized, the balance being a contingency allowance. .l(=tual expenditures were 466.7 million pesos or 96 per cent of the appraisal. sum. Against this should be set some general meaaure of performa.nce.A tJlrude butbelptul indicator is the extent to vbi.ch reclamation was schiel'ed.. Table 9 s:Uows comp ariB on of the situation as seen in the appraisal reJllOrt and as revealed by the JIOre detailed su:-veys conducted thereafter. Table' 9. Areas b.l_Sallnitg Classes in the Northwest Project (Thousands of hectares) .. Prior to 'l.t'oJec·t ChaDge Salini't,y -At In 1961, lPPri1a81 1962-1968 CLass (a) Appraisil. 1962 'b} !,,£ectatlon Actual Unaffected 343 271 +237 +108 Moderate 158 190 -1.$8 - 59 Severe 79 102 - 79 - 49 ..... (a) Various classificatioDs have been used. in each district. The areas classed. as most severeJ.;yaff'ected 0,.. out ot production have been. classed as "severe" and are equated with "out ot production" as used in the appraisal r~pOrt. (b) Surveys were not all do_ in the "ame ye·ar. Note that total. areas do not quite agree,. 1'hj.~ is because of dit£erences in def'ini tion of the irrigation district~.  -.17 - 38. Not ollly were the appraisel estimates or the severity ot the problem an understatement, but ~h~'Projected recuperations were much greQter than those achieved by 1968. It may be argued that the recuperation process will continue into the future, but, in those districts where some ~ . headway had been made, the rate at recuperation has dropped substantially in recent years. It is, therefore, felt 'that most ot the gains arising trom the proj ect outlays would have been achieved by 1968. Tables 1 to 5 of Annex III detail the saliDi ty class changes by districts and they are discussed in more detail in Section ). 39. Returning now to thft individual cost categories, it can be seen trom Table 10 that" apart from the .cost ot direction and administration, the overall outlays vere indeed in about the same proportions as those anticipated in the appraisal. It is at the level ot tha individual districts that the underlying changes in costs show up more clearly. The most notable aspect is that the unit\,cost ot improvements to the existing drainage system were consistently und~reatimated by more than 20 per cent, whereas the cost ot new drains and canal work was conai stently overestimated. Details are available for Cullacan, PUerte and Mayo in Annex II. Le., It is also notable thalt tar more canal. work was undertaken than was envisaged in the appraisal. :nus renecte the tact that in the course ot identifying the specific saline areas, porosity ot existing canals was frequently tound to be more impol'tant than drainage. In all, 2,170 lans ot canals were either renovated or b\~ilt in contrast to the 1,581 lans proposed tit the time or appraisal. Henova'tion and construction of drains J on the other hand, atfected only six per cent more than the 5,216 kms anticipated in the appraisal report. InSlmmary, drainage works wre overstated and canal work understated in the original appraisal. 41. The other important public outlay anticipated in the appraisal report was for increases in the future operating and maintenance costa in each district. Measuring these in direct relation to the capital works under- taken has not been teasible because data available was or too aggregated a nature and the most that can be presented here is a rough approximation ot what these costs ware in years prior to construction and what they now are, comparing the change with that expected by the appraisers. The reason for approximation of these costs &rises tram the fact that the district accounts of S.R.H. do not distinguish between two categories of expenditure - "conservaoion" and "mejoramiento". The tormer is clearly i~tended to be a maintenance charge while the lattet;"oettermentlf,is ot an iri:.;estment nature. In the absence ot any better guideline J one halt this i tern has been added to current cost items to obtain the operation and maintenance (0 and M) costs gi ven in Tables 1 through 5 ot Annex IV. Table 11 shows that this very broad assumption gives 0 and M costs tor 1961-62 cqu1te si~~lar to those in the appraisal report. On the other hand, applying it in-recent years sugge~ts that the probable increase in these outlays was substantially overestimated in FUerte and I_qui and underestimated elsewhere.  - 18 - ,~J.Jt';:I';wr,}.;,~".. i~;;., ~ ,', .. - · .. Percentage Item ~():eosed % Actual Difference :'- 'c ....-." .. t.'! ~ ( EKel. Conting.) :"..:Z93%::i."'f'< . ~:P.r.ainage 271,660 61 229,122 50 -15 iI 81.,62~ 10~,188 20 ·~nals 19 +27 ,otber Works 39,811, 9 56J'~81 12 +~2 f¥.f;intenance Eqpt. 36~S!'0 8 32,291 7 -12 ~rection and Admin .• 1.3:,!!77 3 L3,L~8 9 +222 !~tingency/Errors* 'AA,1ah~ 1,17S Jqtal. Cost L·87,,623 !,66,70, -L '~-"',*"",,~,,-,,, A "-. * Resulting fromuae or aeveral d..ta sources tor actual nlues. Table 11. At Apprais.l (j'YM"caJ1. 7,000 7,797 2,200 ,,119 Jq.rt~ 11,~OO 9,280 10,200 6)99L ~19 1,000 ~,a.6B ., 600 1 2,60L lHui 9,8,0 7,512 1),'$0 7,,1, '''''.'' ~v·' _"'. l"- ''''' ..." .• "j •• ",),.~ ......  - 19 - L2. While it i~ quite untrue that S.R.H. operates its irrigation districts in an eco~omic vacuum, isolated' .trom other social pressures, 0"'j,9 view ot the project is to observe the extent to which the tarmer beneficiaries ot the project made direct contributions to its capital cost. Such contri- .. . butions were to be collected as rehabilitation quotas charged to all tarmers in the districts on the grounds that even those remote trom the works would beneti t indirectly' through the improved conditions elsewhere and the pre- vention ot increased salim ty poss1 bly threatening their own land. Once more, the breakdowns ot income do not clearly 1d6ntlty this component but Tables 1 through 5 ot Annex IV show how total water charges have behaved over time. Taking the extent to which the charges exceeded operation and maintenance costa as being the nettinancial inflow and attributing any increase over the pre-construction level to the rehabilitation quotas we arri va at, a cash infiow which may be compared with the cost ot the 'public works over tim's. Table 12 summarizes this information as internal rates ot return under two assumed lite terms for the works. iTable 12. Financial Return to S.R.H. trom Rehabilitation (Per Cent per Year) District La-Year Lite 60-Year Lite Culiacan 0.26 1.,1 FUerte Negative ,Negative Mayo Negative 0.41 Yaqui 3.60 4.)0 L3. Ddrect payments such as these rehabilitation quotas are but one ot the ways in which project costs may be recouped by the public authorities. Production and export t~es may also be used and, at least in the early years ot the Northwestern project, were an important torm ot revenue to the central government. On these grounds, it may well be arguable that the tarmers in the Yaqui district have been unduly heavily charged for works which, as is shown in Section ), have maintained the status1Uo but achieved little in terms ot posit1v~ recuperation. L4 • Private investments complementar.y to the public works were antic- ipated in the appraisal report where it was held that the "public works alone would greatly improve conditions but, to take run advantage, farmers would also have to undertake farm drainage, leveling and leaching of heavily salted lands." !I It was claimed that virtually all the salt attected land would 11 IBRD Report TO-258b Page 9  - 20 - ~equire ditch or tile drains to keep the water table lOW' enough tor "optimum Fod~ction'· • It has. only been veTY recently that any such works have been und'~t.ken and the very small amount completed to date has been undertaken on a p;rQlllotiona1 basis by S.R.H. at a nominal cost to the farmer. The ben~tj. ts that remain to be captured .trom the rehabili tatton almost certainly .. depend upon such investments. For the purposes of projeotions in this study, both tbe costs and output relations of 1968 are assumed to hold in the future", No attempt is made to conjecture on the probable ~Jt~re oourse of private investment and the retur-ns 'thereto as there is no empirical basis upon which to build such forecasts. L5. There has been one way in whioh some minor land leveling has been undertaken, namely through more thorough land preparation including the use ot land planes. It is believed that these outlays will be reflected in the general upward trend of farm production costs to which we now turn. ~6. An absence of' farm survey data linking farm expenditures and pro- duction has been a real limitation to what 'has been possible on the cost side. '1'11., only consistent sets of time series data on productiQn costs are those prepared by S.R.R. and the crop insurance agency. These are not representative oX' average costs, but apply to .fanners seeking production credits from the pr1 vate and official banks. They OUght to refiect how costs have changed over time but will always contain some bias towards the mix or inputs used by more advanced or more capital intensive production methods. Table 13 shows hw cost ind~x~~ based on these data have changed in the past five years. '!'be crops used to construc'l;, the indexfiJswere sesame, cotton, maize, wheat and sorghum in all four districts together with rice in Culiacan and Fuerte and safflower in Mp,yo and Yaqui. The costs included cfter all direct cash in.p"t$ as well as cultivation charges at custom rates f'or mechanised operations. No charge is made ro~ land or taxes directly related to the rehabilitation wQrks such as water charge~, but other crop specific taxes to finance research have be~ included. Therelntively low oost :Lncreases in Ouliaoan reflect the ,.era.lly small sC$.te of fa:mu.ng and lower ~vel of teclihology. that persists, wbtle the sJJiall incre~es" in 18.q14 are a 81l1Ptoa ot the great technological IdVWlcOS achieved in tha~ Cli~trlct p:rior to 1964. 'l'ab1e 13. Changes in Production Costs per Hectare Cost Indexes Annual Growth District 196L-~5 196~-68 in Costs (%) pz7? u. Culfacan 100 109.6 3.12 Ftierte 100 128.8 8.50 "-yo 100 123.0 7.00 Ya.qu.i 100 117.2 5.S0 b1'~ It is not pOtS sible to ma.ke .ny meaningful comparison of' the cost cl\t.;ng." shown in Table 13 tnth those anticipated in the appraisal report  - 21 - • i' where technology vas treated as 'l;hough it were unchanging. In fact, pro- duction practices' were changing quite rapidly and costs with them. It makes 11 ttle sense to d1 vorce the yield increase arising tram a change in salin! ty class fran that arising trom overall technological advances. Given the large area unaffected by salim ty, the advances in technology would have taken place whether rehabili tatton was undertaken or not. Their etfect, so tar as the project is concerned is to make the reclaimed land even more valuable than would have been the case w.f. thout technologi cal progress. As an exogenous innuence, technology has had a benign etfect generating otherwise unexpected gains in the productive value or the land. In our ex-post evaluation ot the project we have counted as costs those increases in production expenses that were necessary to adopt the improved technology to the extent that such costs were greater than the costs that would have been incurred on saline lands using the appropriate level of technologicallytadvanced methods. 3.~lysis of Benetits L8. This section is 11m! ted to an examination ot the beneti ts defined in the original appraisal and compares their components with observable counterparts over time. As indicated already, only the expected increase in output on land artected directly tiy the desalinization project was regarded as a benefit to the investment. Definition or the size of these benefits d~anded upon expectations about changes in areas, yields and relative'prices. While tor the present we restrict ourselves to this narrow approach, it should be noted that a broader view ot project benefits is given in Section 6. L9. The overall picture of how land was recovered from its saline etate was g:l. ven in Section 2 along wi th an analysis of how total publi c works costs were incurred. As measured by changes in the distribution of land among salinity classes, the overwhelming impression is that while rather less was achieved than expected by the appraisers, the districts or Culiaean, Fuerte and Mayo all made subst£\ntial progress. On the other hand, the Yaqui valley shows alJnost no change flrom 1961 to 1968. Indeed the only change recorded between those years shows that while there was a reduction ot 130 hectares of areas moderately saline, 'there was an increase in the area severely affected of exactly the same amount. Thus on balance, there was a slight net dfJterioration in overall saliDi ty. This instance serves very wen to pOint up the importance ot baneti ts beyond actual recuperation. It cannot be said that the invested .tunds actually had a negative return just because the salinity situation did not improve. Unfortunately, the original appraisal mQdet does not coma to grips With this question which involves specifying what would ~ve happened in the absence ot the rehabilitation program undertaken. ,0. • rehabilitation against the Table 14 sets the areas planned for situation as it appeared in 1961 to 1962 and shows the extent to which land in two broad salin1ty claoses was recovered. This table is based on data given in Annex III where the ditferences in nomenclature tor affected areas is clearly apparent. Broadly, the most severe class or salimty in each district is shown as "severe" and an others having some degree of salinity  - 22 - ,are shown as "mQderate". Since actual classes have names such as "light", '~.,(J.~~Jnn, "heavy,ll, livery he$-VY", ,1'lCW't ,of product:i9JJ." @Id "be:ing rehabilitated" ~, " '801ft. rOQn for di,..greementq~ bow to form cQmparabl~8Ub groups. ~l i,~.lieved that the method used is consistent and a fair reflection of '~ ,,,a-tive order of magnitude otthe problem tram district to district. ~·,allcases, the greatest short-taIls in rehabil1 tation are 01) the ,~a;t,ely affected lands where on-tarm investments in dr.,:iJ)lt.g. ~d leveling _~d -have played an important role. It must be concluded tbat either the re:t;~ns to such investments litre not financially attractive to farmers or ~r,e have been ~nstitutional deficiencies that have pr~vented their taking pl.-a.P.,e. An example or t~ latter is that loans tor dr$1~g~ .nd leveling not been a part or tne ~:iV'ate bank~1 portfolio in t~ 'MIlt 4Jld it has ;WA:1'~ only quite recently that the ofticial banking ,sy~\em ~ "o,Qgr;U.zed l;).ee,n 19¥..-tenn lending or this B,ort as b.eing wi thin its means. 0J\e ou,teame ot tlQ., experience has be~~ that in similar projects tor the I4~~ re,gion in ~:t!.r81MexicoJ the Bank has m.d,e specific provision for c~ed1t to meet .~ cost or such on-farm work~. '-WH;i!f,(.j"Q ... . ';'lppr,ls,ii;t"Un ~ct~al Salinity ~.t~~~t,e S:q~ey aeb$bilit.tion Class .ll to Qe 1961~1962 1961/62 - 1968 Ibhab1li ~at'.ed I~.~,acan Hoderat, 2~8".QOO 23,000 ~,350 Severe 7,000 1.5,000 9,650 Moge;rate l~,OOO 86.,~j9 $6,019 Severe ,~~:"QOO L,),12.0 20,878 Moderate 8;,000 lb,~31 1,363 ,Severe ~ . ,QQO 19,9~3 ).;8,803 Mod~t\te 22,,000 '66,,176 130 Sf;Jvere 8,,000 2'1,9[2 - 130 '~l.~ Return,s to t:g, WC)j.!Qt", ,-u,~ by ~ ~a;l8-.l lJao.d.el are ;~,.~i,llBd in Hrt by t~crQ_q p',~pn" ~d ,s;ltlcecrQ,ps do $howsqme ,,.,~..t.1on in their net vllU,88 ,Mrb..~tM'.",. .trJ, ~~8 rep.rd, tbe O,uy serious \~Y~r~'JJ1t1m.te ot a crop"~,~m~.e QOQW'" ',1IJl,~b Co:t:tOlle I.t.,~ that the ~qris..~ CQf!Jt ot !c;Q)~'.~i~C!V;.; ~.M . " ;rut ,~ ~is crop f s i• •iage over 411 bu~ . _ IWl,t,t, ,.~~tt4.~,at.~ 'two""lIfJ.U. rotatiQ~, was IlQt t~ '4ue "p.t. aela~D:t;9 tIJI' '~ ·IO:~~ ~oted tD cottQlJ. in 1961~2, _, .r~Hl eatimate •• l'J,:i.u .~ .~ dl'~~p'y'.. $iDee c~_~t~~ were .,~ t• • • •t "over timf;; ~ t@ ,.,~..,;L .. ~~iJ.44oItort4C1l can • .,. a w.bst. .ntial tttect. ~ ~.u ;_'~rlpts b,lli; ,....~~ '1.bd~. ~,..~...lW,t 1Dcrease4 .ub8t~tia1.~y 1)1"  - 23 - 1968. Table 1, g1 ves the details. This implies errors in other crops to offset, but no one other crop shows such a consistent bias as does cotton. To the extent that promotion ot export crops was seen as a subsidiary benel! t of the projectp such benetits would have been overstated since the land sub- sti tution that took place was in tavor ot danestically consumed crops, albe! t, especially in the case ot oilset\Qs, import wbati tutes • Table 15. Cotton as a Percental$e at-Total Harvested Area '-~ District Appraisal Report 1961 - iP~\2 1967 - 1968 . ,~ Culiacan 17.8 6.2 1.2 , ... ~ Fuerte 33.0 22.3 31.6 Mayo ,0.0 2,.0 15.6 Yaqui 3,.0 23.L 20.3 52. A more aggregative aspect of cropping levels concerns the extent of double cropping. Cropping intensities have been calculated by dividing the total harvested area by the physical land area not classed as severely affected by salimty • As against the appraisal assumption of 100 per cent intensity, the 1961-62 average is 120, 93, 103 and 130 per cent in Culiacan, Fuerte, Mayo and Yaqui respecti ve4r. These intensities had risen (deepi te increases in the irrigable land base) to lL9, 105 and lLO per cent in Cullacan, Mayo and Yaqui by 1967-68, ralling to 87 per cent in Fuerte where the level of recuperation was so very m~ch greater. The factors behind these changes are quite canplex but it is believed that they are a refiection ot the ., . increased efficiency ot district water administration and on-farm application methods. In part, the renovation at canal structures and the small amount , . ot canal lining would have contributed to this higher cropping intensity too. Because ot the relatively small part of the total irrigable area that benefitted in this direct manner , it has been assumed that such administrative improve- ..... ments took place independently or the rehabilitation thus resulting in otherwise unexpected benetits to the project. Year to year fiuctuations in the cropping . 't,' intensity are mainly the result of variations in water supplies determined .. ' by weather variability • ,3. In the absence of crop yields by salinity classes it is not possible to detect more than the grosle$t ot tJrrors in assumed yield levels. The saline areas would be expected to have lower than average yields for most crops. Comparing appraisal estimates of yields before rehabilitation on salt atrected land with average yields in 1961-62 (18 months atter appraisal) the only  - 2L - ~important error or this ldnd is for cotton in Culiacan and Fuerte. While qllq$t ,other crops grown on saline '~nd were assigned initial yields 20 to 30 ;~ le~nt below ·average yields, in'tbese two, cases cotton yields were assumed ~9 'be, ';respectively, 6 and 20 per cent higher than the average. This would .'~.•pd ,to ~understate the beneri ts due to cotton offsetting, in part at least, ~the overstatement that tollows trom the cropping pattern errors referred to ::;i!l)p~~g1·aph 52 above • .5L. Comparison ot assumed yields after rehabilitation ·wi th historical averages is not likely to lead to any strong conclusions. While the estimates for different crops ~lere both ;&bove and below 1967-68 aver,age yields in most cases, the only case of consistent bias is in Mayo where ''Dhe three principal crops of cotton, wheat and maize all were assigned lower yields .on reclaimed land 1ihan those actually achieved, on average, for the district. In ,an attempt ,toidentiry the extent to which there had been yield increases ·over time that :(:lQ~ldnot be accounted tor by changes in the area cropped, ,a series of re gres- 'sion analyses were undertaken, the results ot which are given in Annex V. 'iBJ!;tefiy, the idea was to acknowled ge that as the area devoted ,to a gi van crop i'Jlcreases (decreases) yields will tall (rise) other things 'b4d,ng equal. Any "~~J\\~ tr,end remaining atter accounting for area changes valS treated as if it ',_Qse ·fran technological improvements. Variations due t() weather were ,rM$Dned randomly distri·buted. ,Table' 16 :g1v.s the significant mean annual ~y;I;.~ld growth rates that ";,;were ':detected in thi;8WIq'.· It should be :noted that the '~~h rates given would be representati·ve or the year 1965. Though not '@:Qstrained in this way, none proved to be exponential. Yields of these .~s have been increasing but ·at,a decreasing rate over time. Table 16. Significant, Positive Yield Changes aue/to" "'techrioiogy"" . (Aven$ge Per' 'cent per ' Year ) ""./. t~ri,sated Districts Crop Sesame Cotton ,1.3.2 Sugarcane 6.0 ,.2 ,Sarnower Beans Tomato 6.9 .O" Maize '14 .'7 6.,9 13.2 8.6 Sorghum '19.~ '11.2 12.0 · Soybeans 11.0 Wheat 8.5 L.7 Linseed 6.6  - 25 - 55. The last component of gross benefits is the prices used for valuation of output. The question of shadow pricing has already been dealt with in Section I and what follows is confined to differences between the prices used at appraisal and those observed before and atter the proj ect tor crops that were not subj eat to direct support. The only important error is in the case of sugarcane in Culiacan and Fuerte. Prices of 50 and L~ pesos per ton were used whereas the averages for . " 1961-62 and 1967-68 were 56 and 61 pesos in Cul1acan and L9 and 59 pesos in FUerte. The appraisal report did comment on the tact that the sugar prices used were conservative in view of the recent (1960) improvements in Mexico's export quota prospects. \ 56. In all,it must be concluded that, apart from the failure to predict that on-farm works would not take place and the level ot recuperation thus be substantially less than expected, the appraisal report was a fairly accurate prediotion of the changes that took place. L • Secondarz an9, Indirect Effects 57. In terms or supplying productive opportunities for otherwise un- enlployed resources there were few external effects ot the project beyond utilizing the latent capacity of the S.R.H. field personnel. However, on the production side, the big increases in output ot importables to which the project contributed may have raised conswnption of these items beyond what they would otherwise have been. It is very dou~trul that the project made a positive impact on any of the exportable crops since other factors, exogenous to the project, were of much more significance in deter.ining changes in the output pattern. 58. There have been local multiplier effects trom the general prosperity to which the project added. These are most noticeable in the form or light ind·~strial plants, service facilities and stores. A Bank supervision report in 1968 listed the many agro-industries and services that then existed close to Ciudad Obregon in the Yaqui Valley. Apart from 127 of the more obviously agriculturally oriented operations - fertilizer distribution, cotton gins, seed merchants, tractor sales and repairs, etc., - there were 272 small industrial plants. Since the Yaqui valley was the one in which the project had least impact, it ia instructive to note how wealthy its towns became. Most or this growth would have ocourred in any case and can be attributed, in large part, to the generously sized farm lots especially among the private farmers. 59. This raises the question of how benefits were distributed among farmer classes. Wi thout a doubt, private fanners able to mortgage and sell .~ their land were greatly advantaged as compered to ejidatarios who could only realize the increased value ot their land in the form ot improved yields through time; even this impact was constrained by a lack of production cradi ts. On an even broader scale it might be speculated that the opportunities for government investment in irrigation has diverted attention from the urgent needs of the remaining 90 per cent or Mexico's tarmers. While S.R.H. was  - 26 - pumping many billions of pesos into these public works, research, extension and supplementary irrigation works in the rest of the country was provided out Or.8 budget usually one tenth the operating budget of S.R.R. excluding 'capi til works. So far as the projects funded by the Bank are concerned, at that point in time a lending program in general rainfed agriculture would have been unlikely. It is equally possible to argue that the Bank ioan for irrigation set other funds £.ree for the rainfed areas especially as we argue the project would have been undertaken later if not at that time. This seems unlikely. The government was desirous ot achieving selt " . sufficiency in strategic food grains and saw the opportunity tor quick (though possibly expensive) realization of this goal in the established irrigated areas • The administrative strength ot S.R.H. also argued strongly tor concentrating public investment where the technical personnel were already established and able to administer a large program or public works. 60. In the recent past, programs to invest in rural infrastructure 'outside the favoured northwest have gained in political popularity and a start was made on village improvement projects in 1968. This, t~gether with '. planned expansion of small-scale and supplementary irrigationliorks under S.R.H., will redress some ot the imbalance ot the past. At the same time it ril'ges serious questions about the optimal rate ot development and stafring pattern tor the ~orthwest. The S.R.H. establishment there 1s large, the oppt?rtunities (physically speaking) under the Northwestern Hydraulic Plan (PI:HINO) l:Lre still immense and some serious decisions need to be made about hi,l,ional priori ties if the natural inertia or bureaucracy is not to result in '. ~ge number ot supernumerary engineers and other professionals in the hen few years. 5. An Ex:-Post Analysis ~1. Berore departing entirely from the "before-and.i.arter" appraisal ~ddel used to test the "8conomic feasibillty or this project, it will be worthwhile to spend a little time 'seeing the extent to which patching that ~el in the light of hindsi'ght 'enables us to obtain ratesot return that r8re:comparable wi th the ones g.lfth in ,Section l.That is, we viII look at lfha't we believe to be anhistorloally accurate betore-and-.arter description lnst;e~d or the one given in the 'Appraisal report. For this 'step in the >8nalys1s, three different beneli t streams have been defined. They differ 'only in the basis on which the i·ncreased output due torehabllitation is vuued. The first one utilizes observed termgate ,prices, the second uses 1sbaaow costs and prices as d43f1ned in ,Secti'on 11. and the last eliminates Mll ts 'and vegetables from the croppicng pattern on reclaimed land, thereby 'approXimating cropping patterns 'on. ·mar:ginal are.s e 102.. In all three caees,average yields, cro.ppin~ patterns, cropping -. ilft~nsitles and other physical 'aspectsvereaa they historically appeared ~Yert.he period 1962 tot968. Years ;beyend 1968 were projected using the 1907..:68 caverage of costs iand benetlts. It :h&$ 'been assumed that the yields r<m tbe moderately l1tfectea -arM'S ve:re 15 per cent ot yields on those in full raM lllormal production. 'Ttrf;d" ~onrorms 'closely to the original appraisal  - 21 - projection tor yield changes and seems a reasonable ratio in view of the fact that the severely affected lands are assumed to be out of production entirely which is an exaggeration irl most cases. As already noted, the absence of any hard data on yields for different salinity classes is a major impediment to precision in this analys.is. The detailed cost and benefit streams tor these three situations are given in Annex IV and the resulting rates of return are summarized along with those derived trom the ex-ante analysis otSection 1 in Table 11. Table 11. Internal Rates or Return to Rehabilitation Investment Measured EX-ante and EX-post How Benefits,were ·V&l.ued~ At Shadol!r Costs District At Domestic Costs At Shadow Costs and Prices, Fruits, and Prices and Prices etc. excluded Ex-ante Ex-post EX-post Ex-8,11te Ex-post Culiacan 12.13 16.25 17.oL 16.11 7.13 Fuerte 21.28 LL.36 L'.23 18.6, 21.56 Mayo 10.73 18.39 18.L7 10.66 11.81 Yaqui 10.13 (-L5,026)* (-L2 ,~97)* 8.28 (-121,0,0)* All Districts 16.3, 27.97 28.81., 15.L4 12.43 * Since rates of return were negative, the 1961 present value of the net benefits is shown in thousands of pesos. Discounting was done using the whole project rate ot return. 6). For the districts ot Culiacan, Fuerte and Mayo, a comparison of these returns with those obtained on an ax-ante basis shows a big increase at domestic costs and prices. Retllrns are l:J.'ttle affected by shadow pricing, all being slightly higher since the overall cost disadvantages outweighed price conces- sions under the cropping patterns adopted. On the other hand, when fruits and vegetables are excluded from the analysis the returns in Culiacan fall sub- stantially below those antioipated, ~lerte comes out about as expected and Mayo shows an increasfJ. Fr~om the standpoint of this type of appraisal model, it seems likely that ()f the three districts Cul1acan would be the least favourable for turthe:r 1~nveatment of this sort. ~~ ~ 6L • The ne gat j. '!ri.,ty ot returns in Yaqui should, come as no surpri se since, so tar, we have restr.icted ourselves to considering QS benefits, only the value of the increased pr'oduction from reclaimed land ar..d this district showed a slight net deteriora,tion. The level ot public investment in this valley, whether judged on the basis of what the appraisal report estimated was to be done or on the basis ot what the 1961-62 surveys revealed, was small relative to the sums invested in the other valleys.  - 28 - 65. Us~ng the limited information available, regression analysis was used to distribute the total public costs between moderately and severely affected but r(Jcuperated land. Th:$ derived average costs lft9re 2,065 pesos per hectare for moderate land and 3,181 pesos per hectare ot severely saline land". !! In Table 18 these costs are applied to the potentially recoverable areas in each district and the actual expenditure on public works is then expressed as a proportion of these potential outlays. This proportion for Yaqui is very significantly lower than in any or the other cases. This has led to consideration of wbat the outcome would have been had more money been spent and a result more like those seen in other districts achieved. Using aRXOpriatl8 averages from the experiences observed in Fuerte (a similarly large administrative unit) and Mayo (Yaqui's nearest neighbour among the four valleys) it has been estimated that an outlay ot 111,07, thousand pesos (86% of the potenttal given in Table 18) would have covered the cost ot rehabilitating 37,100 hectares of the 66,116 moderately saline areas and 15,500 hectares of the 23,520 severely affected. Such an outcome would closely parallel the exper nee ot the pther districts. Tabl() 18. Salinity Area Recuperation Actual Ratio Actual Coat Expenditure to Potential - District Class 1961/62 Cu;liacan Moderate 23,000 ' ~ Severe 15,000 95,210 108,6)1 1.J.L Fuerte Moderate 86,239 Severe L3,120 315,231. 228,96L 0.73 Mayo Moderate 1.3,808 Severe ~,O86 13,321 95,6,8 1.30 Jaqui Moderate 66,176 Severe 23,520 206,6)S 33,1.~2 0.16 66. The physical and economic details of this hypot.het1cal situation are g1 ven in Tables 6 and 7 ot Annex VI. EYE 1I1thout excluding b-ui ts a,nd yegetahles trom the cropping pattern on reclaimed land, the aum of costs is ... 74 .321 million pesos more than the beneti ts. Thus, the return remains nf3gati va. b 11 Fot: Y • Total cost a/public works J Xl -SeVerely saline land recovered J X2 .. Moderate land' recovered, the JWegresa10n gaves ~. 41,160',238 of ,3,180 e 79X1 + 2,06S. 27X2 R • O.9Sg (1.229) (2.287) Figures in parenthesea are t ratios tor testing the null hyPothesis of th~' coefficients not being different tra. aero.  - 29 - The persistcn,cy or these unfavourable l-eturns despite our attelTlpt to make the investmerrt in Yaqui similar in all respects to that, in the other districts is to be explained by the relatively higher yield levels already achieved in this district prior to the commencement, of rehabilitation narks. There was -. thus less to be gained through simultaneous reclamation and application of improved practices. The analysis so far suggests that it would have been better not to invest in the district at, all, but such a conclusion must depend ... upon the as-yet-uruneasured benefits arising from the prevention of further deter~cration in the absence of the project. 6. Alternative Investment Patterns 67. We now turn to answering the points raised in earlier sections about what would have been the likely outcome had the rehabilitation not been undertaken. In an effort to identify production changes that would have occurred with rising salinity levels, data was sought for two adjacent areas in which virtually no rehabilitation t04:>k place. These were the district known as Colonias Yaqui north of t1'hl Yaqui river and the southernmost adminis- trati ve unit of Culiacan. Unfortunately, despite the best efforts of S.R.H. personnel involved, the data have proved to be too incomplete or, in one sense or another, irrelevant for a comparisoll with the areas included in the project. As a result it has been necessary to draw on the limited evidence of how salinity problems were advancing prior to the project implementation and extrapolate these trends into the ±uture. This, while fraught with oppDrtu- ni ti'es for erroneous conclusions, does have the advantage of being a procedure wlrl.ch would have been possible at the tinle of appraisal had the detailed sur- veys then been utilized. cl;80 It seems quite implausibl\9 to view the "do nothing" policy of' cont'tinuous deterioration as a real t\J.ternative. Rather, it was a real policy alt~rl1ati ve to undertake the project at a later date. Likewise it might have been desirable to undertake it even e(~lier though from an operational view- point, this is an alternative of such academic interest as not to concern us here. The question to be answered, then, is "how long could the project have been postponed?" The salinity surveys were most detailed in the Mayo district and, through linear extrapolation it ha\S been concluded that by 1967 the approximately 48,000 hectares of' units 2 and 3 would all have been affected by salinity to sane degree. On the bas:la of general topography and the location of' saline areas up to 1962, it is believed that these two units enclose all the lands that would become saline with '~he then existing irriga- tion and drainage system. Table 19 sho~l similar assumptions made for each of the districts and compares the maximum susceptible areas with those saline (all classes) in 1962. It is interesting to note ·that one implication of ... ,. 7, these guesses at the progress of salinity is that in 1962 the areas then saline varied between 65 and becoming saline. per cent of"'~he maximum' area susceptible to 69. The distribution of this total e)~ected saline area between the two broad classes "moderate" and "severe" was determined for each district on the basis ot what was expected in Mayo taldng into account the io1 tisl differences among districts. On the basis that each district would have all its susceptible area affected to some degree by 1967 the areas shown in Table 20 were deduced.  - 30 - Table 19 •. Over-all Salinity Incidence by ~stricts Are,a .~ .' -. . Umts of District District Saline ,1962 Area Susceptible Total Area Susceptible to Salting .- Q~:LiaQan: has. 38,,000 60,000 102,000 1, 3, 1/2 of 2 % 37 59 100 :FUerte: has. 129,,300 195,000 213,300 1, 2, 3, L, 5 % 61 92 100 Mayo: has. 3L,LOQ L8,OOO 92,000 2, 3 % 37 52 100 ~~qui: bas. 100,300 ,'US,OOO 220,)00 1 % 45 5L 100 W9tal has. 302,000 L21,OOO 627,000 % h8 67 100 Table 20. PrQjec'lie.dS-Uni ty Cond! tion in 1967 without ." "H' U . Rehabllftatiorf (Hectares) a.tim ty Clas, _ Total District M9dsrate ·S'ever8. Saline Culiacan ,30,OQ.() 30,.000 60,000 ~erte 1l0,OOO as"QOO 195,000 !iayo 16,000 )2,,0<)0. L8,OOO I_qui 77,000 l.al,OOQ, 118,000 411 Districts 233,000 , 1.88'00.0 " L~l,OOO  - 31 - 7C. The reason for not considering the possibility of postponement of rehabilitation works beyond 1967 was the belief that with such large proportions of the total area affected, the socia-political pressures for remedial action to preserve the property of those with vested interests in the area would far outweigh any strictly economic criteria for continued postponement. 71. So much for What might have happened while the rehabilitation was put off. To permit construction of a series of costs and benefits to compare with what actually took place" it will be necessary to project the rate and pattern of rehabilitation that would have taken place'with a delayed project. While allowing proportionally more time depending on the delay and consequent deterioration, it has been assumed that the same general pattern of rehabilitation would occur as did in fact, the end product being a return to the level ofsalini ty observed in 1968. SU,ch a reconstruction for the case that the project is delayed 5 years, nothing being undertaken until 1967 is shown in Table 21. Similar reconstructions were generated for the cases of 1, 2, 3 and L year delays. They are shown in Annex VII. The effective life of the pro j act, once completed, was assumed to be unchanged. 72. In what follows, the analysis done for the Yaqui district differs from that for the others only in that we do not start from what actually occurred, but from the hypothetical situation described in Section 5 above. The reason for this is that, since what happened in Yaqui was essentially a holding action to prevent further salinization, a sj.milar but postponed pro- jeet implies a permanent reduction in the irrigable area of the distrct. Wh:rle this may indeed be the most desirable course of action, the data available do not lend themselves to a thorough analysis of the implications of such a policy. Thus, the question that wi.ll be asked about the Yaqui is "if a policy more like that used in tte other districts had been applied, when shOUld it have been initiated?" 73. The procedure followed in determining the optimum timing of these investments was directed towards one quite simple end. This was to show hnw the discounted value of production in the districts, net of all project costs varied with the timing of the project. As with all such values, they also depend upon the discount rate of interest used. The idea is to show, at different interest rates, which year was the ri~ht one in whicn to start (has the highest present value) and what the cost of a wrong decision was in the case that 1962 was not the optimal year. These concepts are illustrat.ed for a simple choice case in Figure 1. 7L. The tables of Annex VIII show all these costs for interest rates from 5 tr1I'ough 25 per cent. They are summarized very briefly in Table 22. In the case of Culiacan, it can be seen that 1962 was the best year to start .,It so long as the o~portunity cost of capital exceeded 5 per cent. The invest- ment in Fuerte would only have been better postponed (by one year) if the appropriate di scount rate was outside the 10-20 per cent range. The year 1962 was optimal, or 'nearly S0;i in Hayo for all interest rates less than l~ per cent but for rates higher than this level 1967 (or later) would be prefE·rable. In the case of Yaqui, had the proposed alternative investment pattern applied, it would have been best to start it in the 1963~65 period~ It will be noted that 196L and 1966 never came out to be optimal years.  ~ Table '21. ASSUNED PATl'ERN OF DETERIORATION AND REHABILITATION IN THE EVENT TH.4.T REHABILITATION BE POSTPONED m~'l'IL 1967 Actual . Thousands of Hectares Assumed Deterioration\a) Rehabilitation to Actual'1968 Levels(5) District Salinity Class 1962 2963 196h 1965 1966 1967 1968 1962 1970 1971 1,972 1973 197h 1975 Culiacan: Unaffected 64.0 59.6 55.2 50.8 46.h 42.0 45.6 49.2 52.8 56.4 60.0 65.0 70.0 75.0 Hoderate 23.0 24.4 25 •.8 27.2 28.6 30.0 30.0 30.0 30.0 30~0 30 .. 0 27.2 2l~.421.6 Severe 15.0 18.0 21.0 2h.0 27.0 30.0 26.4 22.8 19.2 15.6 12.0 9.8 7.6 5.4 Total Area 102:0 Fuerte: Unaffected 84.0 70.8 57.6 44·4 31.2 18.0 . 39.0 60.0 81.0 102.0 123.0 135.6 148.2 160.8 I-foderate 86.2 90.8 95.6 100.4 105.2 110.0 96.4 82.8 69.2 55.6 "42.0 }5.1 34.1 30.2 Severe .lV. 1 51.4 59.8 68.2 76.6 85.0 77.7 70.4 63.2 55.9 48.6 39.8 31.0 22.2 , Total Area 213.3 w N }layo: Unaf.fected 57.6 55.2 52.4 h9.6 46.8 L~4.0 50.0 56.1 62.1 68.2 74.2 75.4 76.6 77.3 . }'Ioderate 14.4 14.4 14.8 - 15~2 15.6 16.0 16.1 16.2 16.4 16.5 16.6 ]5.4 14.2' 13.0 Severe 20.0 22.4 24.8 27.2 29.6 32.0 25.8 19.7 13.5 7.4 1.2 1.2 1.2 1.2 Total Area -92.0 Yaqui: Unaffected 130.6 J.24.4 118.8 113.2 107.6 102.0 105.6 109.2 112.7 l16.3 119.9 123.4 127.0 130.6 I-loderate 66.2 68.2 70.4 72.6 74.8 77.0. 75.6 74.3 73.0 71.6 70.3 68.9 67.6 66 .• 0 Severe ..l.l:i 27.4 30.8 34.2 37.6 41.0 38.8 36.7 3h.5 32.4 .)0.2 28.0 25.9 2).7 Total Area 220.3 (a) The assUmed deterioration is a linear trend from the actual levels for 1962 (1961 in the cases of Culiacan and Yaqui) to a level for each salinity class based on a projection of the better documented case of ~o.· (b) Since~ relative to the rehabil~tation between 1962 a."1d 1968, the delayed work liOuld involv::! about }-to% larger arGas, the tL'T.e lapse has been expanded to eight years~ tcrnrL'1atine lr! 1975. For CultactL.'1, Fucrt.e and I'fuyo, fi!::-urca fo:'" 1972 reflect the relatively faster proeress achiov('d fron 1962 to 1~/66. Sincc nc ;,rOLTCG!1 l'W.S rrl~do :L'"1 Y~qui, onD linear interpolation has been used to regain 1968 conditions. , ; I I  - 33 - Table 22. Discount Rate ~t1mal StartW Date (Cost ot Wrons Decision) (% p.t,~".) Cu acan .. erte Mayo Yaqui* 'lJ 1965 1963 1965 1965 (91,661) (93,116) (958) (160,,15) 10 1962 1963 1962 1963 (6,660) (21,288) 15 1962 1963 196, 1963 (L93) (l,LOL) (13,092 ) 20 1962 196) 1967' 196$ (h ,101') (13,361.,) (lL ,118) 2$ 1962 1963 1967 1967 (8,LL2) (21,118) (2,3,07L) * The Yaqui project analysed is an hypothetical one. .see text. _, if  - 3L; - Project I Optimal Cost of Wrong Decision • ab Present Value Project I I Optimal Cost of Wrong Decision • cd n Discount Rate Figy.re 1. ~ivation of the, C~8t of a W:r~ng D~c~sion fro lJl Present Values  - .35 - This has arisen trom the discontinuities involved in the use of whole years as time periods. It seems unlikely that the refinements necessary to eliminate this problem would be worthwhile since the general nature of the desirable investment pattern is already clear. 7,. Leaving Yaqui aside for the moment, we can see by adding the costs of wrong decisions given in Table 22, how much was lost by the decision to start all the work in 1962 rather than choose optimal starting dates for each -~ district. These costs are given in Table 23 and expressed as a percentage of the total public works costs for the thr.ee distpicts. It is clear that so long as something like a return on capital in the private sector of about 15 per cent is regarded as the appropriate discount rate, the project costs were not greatly increased by the non-optimality of the timing. If, however, ~ ,social time preference rate in the 2 to 5 per cent range were thought appropriate, the extra cost is very aigniticant. It has been the Bank's policy to use something like private sector returns for these purposes and on this basis there are no strong grounds for having postponed the project in these three districts. 76. The Yaqui valley project prevented further salinization in the aggregate. Of course, there were farmers who benetitted and others who lost their land in this process, but we will continue to deal with the total etfect. The projected salinization process outlined in Table 21 suggested that, by 1967 a further 10,800 hectares would have become moderately saline and an extra 17,500 hectares would have become seriously saline. The public outlay of 33,L,2 thousand pesos can be regarded as having prevented such an occurrence. Assum- ing the more serious salinization to have cost more to prevent (as in the case of recuperation) we can ascribe a cost of about 1,000 and 1,400 pesos per hectare respectively, to the prevention of moderate and severe salinity. Such a cost would seem reasonably cheap were it in fact the end of the matter. The rapidly rising cost ot operation and maintenance in the Yaqui valley, shown in Annex IV Table L, suggests' that it is far from the end and a very much more detailed study is needed to determine, plot by plot J which parts of the valley warrant the costs of reclamation and which do not. The data available for the present study have not been at a su.f.fi~dently disaggregated level for this purpose, but, in the record cards maintained within the districts there is a wealth of information that could make such a study feasible on a large scale. Table 23. Ex-post Cost of Undertaking the Project in 1962 for CUliaean, FGerte and Mayo Mscount 1961 Present Value or As Percentage of .. " , Rate (% p .a..:) 10 Decision (lbousand p~sos) 186,395 6,660 Pnhlic Works Cost L3.0 1.5 15 1,897 o.L 20 2, 'l:7.,'L65 L.O 29,560 6.8  - 36 - 1. Conclusions 17. For a project that has been widely regarded as a great success, the northwestern rehabilitation has be'en shown to have had some quite interest- ing features. In general terms, it is true that there was a large measure of success in moving to the desired objectives and that the probable costs and beneri ts were well specified. It also appears that there was little to be gained from delaying the proje.ct. On the other hand, three aspects of the project do "Tarr'ant one last word. 780 By the time the project was begun, the appraisal report was out of date with respect to cropping patterns, intensity of land use and yields. The best of current Bank appraisal practices would have avoided this situation up to a point. Sensi ti vi ty analysis should pe used to determine the elements· of the appraisal whose values are of critical importance for its estimated impact. These, at least·, should be not ni6re than one year out of date when the project becomes effective. 79. The pricing of inputs and outputs at administered prices may be desirable under circumstances in which there are long-term bases for the continuation of the policies involved. The Mexican agency, CONASUPO, has, on the contrary, proved a most pragmatic policy organ and internationally competitive crops have been the stated objective of this policy for the irrigated areas. In such a case, the real social cost of output subsidies and input taxes should be taken into account. In'the Mexican case, the shadow pricing used in this study resulted in a slight improvement in the value of net farm benefits sinc'e product subsidies were found to be sli ght1y less than the impost on inputs through taxes, import restrictions and other policies. 80. The last of the thr'ee points. concerns pri vate comp1eni~ntary invest- ment. It seems doubtful whether a similar Bank appraisal today would have made any serious attempt to quantify the value to farmers and society of this work. Were the project specifically to finance this and nothing else it would be given better treatment. Thataat that virtually none of the on-farm invest- ments took place indicates the importance of our learning better how to measltre benefits to micro-level investments. It would seem quite inadequate to simply measure an internal rate of return to an assumed cash now. This is the "better" trleatment referred to above and typically accorded 'any agricultural crecli t project. This class of' project i,sone in which the absolute scale of the project is of primary interest. For this reason, a measure such as net present value has a valid role to play. Farmers may be expected to undertake on-farm works if, after taking account of all costs (including interest and repayment of principal) the added output, valued at farmgate prices will give them a nO'l:,iceable increase in residual income which can permit them to improve their living standard, or prevent it trom decll·ning ,so much as would otherwise have been the case. No rate of returncalculati'On elm quan.tity such absolute magni tud es and so other measures should be ueed. In the case of Mexi co, the information necessary for measuring this private financial: impact will shortly be available for a few a,r'eas where S.R.H. has undertaken subs.idized drainage work for individual farmers.  - 31 - 81. Apart trom private profitability, the Jllajor concern of app:-aisal is to show that the investment is also in the social interest. At, least two questions need to be resolved betore one can tell if the social c~iterion is met. .First one Ilust decide how goods and services are to be valuf)d in the eyes of society as a whole. This study used r .o.b. export and cJ:.i.f. import prices. Secondly, one must decide which technology is appropriate to give a "best" distribution of resources and incane. For this study, th6 relatively' capital intensive approach adopted towards conatruction, maintenance and operation seemed appropriate in light of relatively high wages and the sunk costs ot the existing irrigation works. ~th respnct to the private on-farm works, a careful analysiS of the farm l~or cal~ndar may reveal times of year in which a more labor-intensive approach would be optimal. 82 • One final word on the Yaqui valley. Here, when one takes an aggregate view ot the district, there is no reduction in the saline area over the period studied. The intial cost of preventing further salinization are about equal to the net value ot a year's production which makes them appear an attractive proposition. On the other hand, operating and mainte- nance costs continue to rise and it will be through observation of these shifts in the future that the program or 1I.TOrk in Yaqui must be judged • ..  - .38 - III. 'mE NORTHEASTntN PROJECT 1. An Ex~Ante Review J' 83. This chapter deals with the Northeastern districts - Rio Bravo, Rio San Juan and Delicias. It presents a method of revising the original appraisal which would correct or improve upon the latter's deficiencies, on1t utilizing data that could have been made available prior to the start of construction. BL • The methodology used in the original appraisal was essentially the same as that used in analyzing the Northwestern project di.scussed in Part II. The only important difference concerns thf" way in which the value of prevent- ing further salinity was included among the benet! ts. Whereaa, under Loan 275 in the northwest, these benefits were simply not quantified, tbe project return being based on the value of the reclaimed land, in this project average yields before and after reclamation ware assumed applicable to all the area. The change ,in production over the entire area was then counted as the benefit to rehabilitation. Of course, it would be possible to choose the yield changes in such a way as to make the beneti ts so obtained equal to the benefits as measured in any other way but" methodologically it 1s incorrect to ascribe the same benetitto both the presently unsalina and saline areas still in production. Our revision ot the appraisal includes an alternati va specific- ation of the beneri ts in which seriously saline areas are treated as if they were out of production while moderately saline areas are assumed to have yi.elds 75 per cent of the yl.elds on unatfected areas. It was found that the measured benefits were not It'tlUch atfected by this modification. 8$ • The most impOlrtant data deficiency in the appraisal was its fail- ure to use an appropriate cropping pattern. Cotton was assumed to playa big part in the agriculture ot the area but aa Table 23 shows, its share of the harvested area had fallen to an insign1.t:1.oant level over the past decade. Table 24. ·The ChaniinsRo1e Qf Cotton in'fhe Northeast Appraisal Average Item District Report 1962/L 1967/8 Cotton Area Bravo 63 2h 1.3 (% of total) San Juan 37 21 1.7 Deli.ctas 71. 30 20.7 Cotton Price Bravo 22S0 21.53 2577 (Pesos per ton) San Juan 22S0 21aL.B 2553 I:=:-' Deli.cd.'as 22,0 1999 2500 Cotton Yield Bravo 1.1 o.L2 1.68 (Tons per Ha.) San Juan 0.9 0.6L I.Lo Dellciaa 1.1 1.29 2.)8  - 39 ... This decline was apparent before the loan was authorized but the appraisal was not revised. Sensitivity of the project return to the assumed cropping pattern depends upon the assumed relative profitability of individual crops. In this case, cotton was assumed to be most profitable and to playa big role~ Higher than expected yields and cotton prices only softened the blow and could not affect ,the virtual dieappearance of cotton from Bravo and San Juan. Only in Delicias has cotton remained a significa.nt crop and it should be noted that the price of cotton has, in this one area, risen sharply towards the national average over recent years. 85a. Apart fram the absence of any price incentive for cotton that might have counterbalanced the guaranteed support price program for maize and other crops, cotton picking costs have risen with recent elimination o:r the "bracero" program. Whereas, prior to 196L, many agricultural workers would pass from Mexico to the United States to pick fruits and vegetables, stopping off in northern Mexico for the cotton crop, there are now no such migrant workers. This poses an organizational problem and also means that a different class of worker will be willing to come for the cotton picking alone. Li ving conditions for such workers are qui ta abysmal. Yet another factor working against the production of cotton is the relative riskiness of the crop as compared with maize and sorghum. The cost of cash inputs is considerably higher and the variability of yield resulting trom climatic variation is also greater. 86. The original appraisal costs, benefits and full-development rates of return are presented in Table 2L. The costs include public works and private, on-farm, investments. The net farm benefits are the expected in- creases in the gross value of farm production net of farm production costs and operation and maintenance cost increases. Expressed as a percentage of the total investment outlays, these net farm benefits give the full-development rate of return (FDR). While these returns are all in a generally acceptable range, they fail, by the nature of their derivation to account for the timing of costs and benefits. Following our general reevaluation methods, the first step is to recast these costs and returns in a time dimension and calculate their internal rate of return (IRR). The result of doing so is shown in Table 25 in the columns headed "Original". Table 25. Original AEPraisal Costs, Benefits and Rates of Return (Values in millions ot peaoe) rlISTril:CT Rj.o Rio Ciudad Total Item Bravo San Juan Del1cias Loan 336 "" . Costs: Public 250.0 90.0 10.0 ~lO.O Private 41.0 19:6 21.8 82.L Total ~91ftO 109.5 9LB" ~2.4 Net Fram Benefits 18.1 16.3 29.1 l2~ .1 F.D.R. 27% 1,% 32% 2,~  Table 25 t ORIGntAL AND REVISED EX-ANTE EVALUATIONS OF THE PROJECT BY DISTRICT AND TOTAL. (Values in l.ftllions of Pe!.2!) DIS T Jl leT Total I T E M Bravo San ~Tuan __ Del.tciB;S Loan ll€ _ ___________________.... Or iOsfil:& ..... "' _ _Re...-...,v......_ _ ~. ~!."Q ~_..Qe~ Rev. cr!g. _!p. Cost of PUblic Works 250.p 90.0 70.0 410.0 Qn-£~ L~vestment8 41.0 19.6 21.8 82.4 Total . 291.0 109.6 91.8 492.4 Che.nge in Annual Cperating and ~~tenaace C03tS~ 1964 on. 6q3 2.7 2.2 11.2 Change in AP..l'lual Net Farm Income at full. develop..1llent (a) 85.0 46.2 19.0 '-13.1 31·3 19 . 1 135.3 7h.O .-.......----~ _. -- ~ - --..,..,.-.------- ...... ....... Project Net Bene£it 1962 ··.0. 0 0 0 0 0 0 0 r::- 1963 -17.66 -17.66 ... 6.51 -6.51 -5.18 -5.18 . . 29035 · -29·35 0 1964 .;60.ho -64.28 -35.27 -35.86 -25.30 -28.43 -120 .. 97 -126,.56 1965 -50.34 -58~10 -26.82 -27.98 -16.36 -18.60 -93.52 -loh . 90 1966 -40 . 30 -51.94 -17.82 -19.55 -13.08 -16 .. 74 -71.20 -38.27 1967 -20.30 -35.82 .99 -1 . 31 5.97 1.09 -13·34 -36 . 10 1968 27.70 8.30 2.89 -0.06 9010 3.00 39.69 111124 1969 36.20 12.92 4.• 79 1.2, 12.23 4.91 53.22 19.08 1970 53.20 26.04 10.60 6.47 19.71 11.17 83.51 43.68 1971 61.70 30.66 12,,50 7.78 22.84 13.08 97.04 51c52 1972 70.20 35 . 28 14.40 9.09 25.97 14.99 lio.57 59.36 1973 - 2004 78 .. 70 39.90 16.30 10.40 29.10 16.90 124a10 67 .. 20 Internal Rate of Return (% p.a.) 20 .. 55 11.19 ll.44 7.35 22.49 13.BO 18.88 10.92 (a) Deriva.tion of revised estimate is given in Table 9.5  \ r ., Table 27 I REVISED EX-ANTE MODEL OF BENEFITS TO THE NORTHEASTERN PROJ~'l' AaEA Yl];L.u PRICE DISTRICTI VALUE COOT NET VALUE OF PRODrN w/o with Percent of wlo with CROP (Pesosl wlo with w/o nth wlo with rehab. rehab. total !I rehab. rehab. ton) rehab. rehab. ha. ha. % rehab. rehab .. rehab. rehab. (tons / Ra.) (Pesos / Ha.) (Pesos / Ha.) (total mill. pesos) BRAVO cotton 51,250 52,500 25 0.72 1.2cJ2/ maize 77,900 79,800 38 2,453c/ 1,766 2,944 1,075 1,888 35.4 55.4 1.46 2.75 68r 997 1,878 sorghum 15,850 17,100 31 2.24 732 1,475 20.7 32.2 TOTAL 3.75 517£1 1,158 1,939 669 1,)00 .c::- 205,000 210,000 100 35.1 49.1 J--I 21·2 IJZ·!! Increase in net value of prodrn Pesos 46.2 mill. SAN JUAN cotton 18,160 19,040 28 0.96 1.4~/ 2,453!:1 2,355 3,434 1,35# 2,125 maize 35,510 36,040 53 1.62 3.00 18.8 24.9 sorghum 12,730 683 1,106 2,049 125 1,575 13.6 12,920 19 1.79 3.75 517£1 17.0 TOTAL 67,000 68,000 100 925 1,939 554 1,.300 4.7 8·3 JZ·1 2,0.2 Increase in net value of prod' n Pesos 13.1 mill. DELICIAS cotton 17,860 20,520 38 1.29 1.61£1 maize 13,630 15,660 29 1,999c/ 2,579 3,218 2,119~ 2,475E1 8.3 15.2 wheat 15",510 17,820 1.50 2.75 683c/ 1,024 1,858 800 1,50~ 3.1 5.6 33 3.00 4.00£/ 81~ 2,448 3,264 TOTAL 47,000 54,000 100 1,100 1,55 20.9 30.6 J2.J 51.4 Increase in net value of prodrn Pesos 19.1 mill. a/ Based on 1962-63 averages b/ Adjusted to reflect actual levels in :rears 196C-63 E/ f.o.b. export or c.i.f'. jmport price  - L2 - 87. Moving from FDR to IRR results in consistently lower return figures. In particular, that for San Juan is reduced to 11.L per cent. ~lhere is an even greater proportional change for Delicias, though in absolute terms J the return still appears respectable. Still .following the pattern set in Part II, under the columns headed "Revised" in Table 25 are to be round rates of return computed on the IRR principle but incorporating the best of data and methodology that might have been undertaken at the time o.f appraisal. It remains a truly ex-ante appraisal. In this revision all returns drop drastically. 88. The revised ex-ante model incorporates average yields for the three years prior to construction (1960-1963) and uses cropping patterns for the year preceding construction (1962-63). Prices used in the re~nsion were those for the same crop year and inputs and outputs were shadow priced in the same manner as described for the Northwestern project. It is important to emphasize that, at some cost, the data used in this revision could all have been made available to those making the original appra.isal. Current Bank practices would ensure that a good deal of this information would indeed come to light. The practice of appraisal review in the case of delays in implementation could have helped detect the trends concerning cotton, while almost all Bank project~ are now evaluated using some form of the IRR concept. 89. Tho estimated returns of 11 and 7 per cent for Bravo and San Juan would undoubtedly have given rise to second thoughts and a restructuring of the project so as to orient it more appropriately towards the cereal crops that were coming to dominate the pattern of farming in those two areas. 2. Ex-Post Evaluation 90. The extent to which this section can do for the Northeast what Sections 2, 3, 5 and 6 of Part II did for the Northwest will be severely limited. by the fact that much of the construction 'WOrk is only recently completed and~ the loan was, at the ~ime of writing, still being disbursed on account of works :rn Rio Bravo. On the other hand, Rio Bravo provides the only case in which outlays and yields efrects can be correlated at the level of inter- nal administrative units. This extra disaggregation of the data would have been most desirable in other areas as it allows one to separate the impact of factors varying across unit~ within districts such as the level of salinity. 91. It is notable (See Annex II, Tables 6, 7 apq 8) that actual expen- ditures by the public authorities were closer to the ms planned at appraisal than in the Northwest. Both S.R.H. and the Bank had improved their costing procedures it would seem. On the other hand, the role or private investment was just as summarily assumed essential here as in the first project, and just as little was actually undertaken. Up to a certain pOint the absence of private investment has been compensated tor by enlargement of public drains using contingency funds. • 92. The approach or this section is to deal exclusively with the case of Rio Bravo for which very complete information was available. As compared  - 43 .., with either San Juan or Deliciaa, Bravo has the added advantage that rehab- ilitation took place at different times in each administ.rative unit (see Annex II Table 10) Uni t 3 of the Bravo District DB bare1y started on fl before 1967-68. This delay has served to pronde tl basis on which develop- -e ment without rehabilitation could be inferred for the other units. Experience in units 1 and 2, on the other hand, was us_d to project the likely future course of' investment in unit 3. The overall performance projected for Bravo has been keyed to the actual achievement,s in nearby San Jua.n. By working in this way, it is believed that we he;Y'e genera.ted an ex-post example of develop- ment with and-without rehabilitation that is indicative of the general order of' the value of' rehabilitation in these three di8trict~l. 93. As in the Northwestern pro.ject, a key question to be resolved in assessing benefits, ha~ been the yield differential as b6.tJtween saline and unaffected lands In the Northeast, available records have only permitted 4' us t.o class the land in this dichotomous way ""·'sw.ne, or ullB.ffect:.ed. Using available data on yields a.nd sa.linity for t..he Bravo district" maize an.d sorghum yields on saline and unaffeoted areas were estimated. The effect of production year was also detennined and finding it nonsignif~ca.nii data were pooled for all years and available subunits of the district, giving 12 obse~v­ stions. The regression equation used was ot the following· form: y n a + bIAs + b2Au + e Where Y • total production of the crop As ,. Area. saline in has e Au • Area unaffected in has ~ e - Norma.lly d:i. stributed random error -term The following results were obtained: Crop Maize 1177.7 2415.7 (l 61)* f0 (~.L2)* Sorghum 1331.6 2750.0 (2 .. 10)* (5.82)* * The numbers in parentheses are tbe t .... ratios, with tOtlOS ,.. ". 1.833 for 9 d.f~, sign considered~ In both cases, yield on saline land was about 50 per cent of yield on unaffected areas and this is our best estimate of tha percent~ge fo~ the Northeastern areas. 9~. Using similar projection techniques to those employed for the North~ western project, it was estimated that without any reclamation program salinity would have increased rapidly in Rio Bravo up to 1970 When only areas with natural dratnage advantages would have remained unaffected. Since the actual rehabilitation is still incomplete a terminal date (1971) had to be assumed and expenditures out of the remaining budget allooatod Qocordingly. Sind.larly  - 44 1M -the Sal.ilUty chMgoS e).~tic1paijed wenl specified. Table 28 shows the details Qf +.hese asaumptiofJ.6 Using historically obselved cropping patterns, ave rag a tI yield.s and cropping iIl't~ellsities together with. (shadow when appropriate) costs of production $nd priQes as in the case of the Northwest, the net .- ~conomic ,,,nJ.ues of prod:l.cticn shown in Table 29 have been derived. 95.- The cost and benefit streams of Table 29 have been used t.o develop till~e ostima:cea of the retum to thG project vis ...a-vis a. "de nothing" situation w~thout the project. ~1~ three cases are distinguished by differing assump- t:tans concel:"lrl:o.g the avere.gc rela:t.ionship 'between the yields and costs on oill.ine and unaffected aroa.B. 11be closer SJ~ the two yields to being identi ... GU (their r.~-).tj.o approaches unity) J the less wil.l be the gain from a i'~"3CJ.as5ific ation in favor of ull£1..f£eoted arena.. Benefits and costs are assumed. c.ons i;..ent from 1972 onwa:rds" The r€lsul ts are as follows; .aasumed R.8.tio of Yit~lds Interllal Rate ____ ~~1~ll6 : un!,:i!'!c:le~,..!!!..~ __t_u_m_....:I(_P_e"'c_e_n_t o_r_He __ ....... ~_ar_t, p_e_r__ 16.5 12 .. 5 9. 0 _________ F~_~.~~_____________.____----------------------- .~  - us- Table 28: Progress of Salinity with and without RehabUitation in Rio Bravo . District 1964 - 1270 ~ .- Areas Affected :!?Z Salt ~Hasl • Actual H.v DO the §i '§Id· - ~hout R~habiWai.oD with Year Sub Units o:~:·:. :.·. . ~ Total Rehab. Control V. Herao8o R. Bravo District District 1964 4J l ,,095 55,650 9,422 108,167 108,167 1965 47,'000 61,00'.) 10,692 118,692 92,176 1966- 54,000 10,000 11,963 135,963 16,186 1967 61,000 80,000 16,871 157,871 78,737 1968 61,000 80,000 18,000 159,000 62,052 1969 61,000 80,000 21,000 162,000 57,000 1970 61,000 80,000 24,000 165,000 53,000 Table 29: Rio Bravo. Net Economic Value of Fann Production and Public exzenditure on Rehabilitation Thousands of Pesos) With Rehabilitation Without Rehabilitation Net Economic Public Net Economic Value of Production Year Value of Production Expenditure Yield Ratio I Saline as fraction Total p~ Inv. 0 & M of Unaffected HectarO " 0.4 : o.S 0.6 1964 85,469 OQ358 30,299 8,,469 85,469 85,469 1965 175,235 0.734 49,637 160.. 037 163,176 166,028 1966 60,879 0~255 26,655 49,567 .51,810 53,889 1967 157,091 0:658 34,657 '118,,141 125,915 133,098 1968 66,608 0.279 3~,759 47 .. 38l 51.. 064 54,534 1969 111,969 0.469 29,899 77,482 84,009 90,191 1970* 112.082 0.469 29,881 75,6{)9 82,426 88,910 1971* 112,454 0.469 29,88J: ·'5,609 82,426 88,910 1972*and 112,702 0.469 2,739 75,609 82,426 88,910 later years * Based on projeoted rehabilitation  - 46 - 96~. As has already been shown in paragraph 93, a ratio of 0.5 appears ~o be the moat }>robable one· and this giVes a returh of 12.5 per cent over the lr 'S'i~~a~1on that as projected 'without a project at a~. We are now i:1 a p:bS"l~,:-on to sUml\~arize tp~ returns that have been est:mated for the R20 Bravo ) -Di;;strJ.ct within !~the, Northeastern Project. These estl.mates are presented ~th ~ brief' des;pription of the evaluation models used in Table 30 What is It tnttre 'evident f'rOl\~ this summary is that, had the original appraisers kept ... 'st'rictly up-to-date with the cropping pattern they would have been able to " obtain a'n internal return very close to that which is the best estimate we have been able to obtain on the basis of hindsight. Various offsetting factors w~ld have been overlooked in the process, the most important being the extent ,or 'on-farm investment. Thus, les8 was invested and less was achieved but the i:ntonnation available 'co date does not, provide the answer as to whether ~~ther investment at the micro level in drains and leveling will have as ·gteatan impact in the future as had originally been expected Further trialsII with 'careful documentation of the results will be needed to ans'Wer this question" 'Table .30. Comparison of Eqonomic Return Estimates for R:!;o Bravo Rehabilitation MOdels Type of Value ;~scription of Model Return (% p.a.) EXlooant'e, original appraisal FDR 27.0 lJei_ .d~~Jo.t.o."ts and pric(jS IRR 20.5 and an historic cropping pattern. Ex. . ante, revised apprais1l11l 11.2 Usihg shadow costs and pri oes and ~ost recent cropping pattern~ h-post, reevaluation. IRR 12.5 U'sirtg actual ( shadow) costs and prices, actual cropping patterns and speoifi- cation of a "without" project m6df!l. , , • , ...... ' r .. · , ~ 91. While limited to making a rull scale "With and Without" analysis in the case of Rio Bravo only, there is fio roason to expect that the other dirttricts would have performed vary different1y. That is to sa.y, on the eVitiel1ce available, it is conoluded that the reVised ex-ante es,timated re- tuttis 'gi.ven in Table 25 are clo'se to those aotually realized. Thug, Bravo, ... Sari JUan and Delic:ias are estimated to yieid internal rates of return of' 11, '1 atld lb per cent respectively. Furthermore, the total investment was 83 pel" cant of that originally intended since hO significant part of the private, dh ... ?arm investment has been ll:h.det'taketl.  - 47 - )11 Conclusions 98. While every effort has been made to evaluate the northeastern rehabili tation projects in the same way as was done for the northwest, a few differences in 'the approach should be noted. First, apart fram a revision of the ex-ante model, analysis has been restricted to the Rio Bravo district. Although this was the one district in which works were still in- complete, the form in which data were available made this the most feasible to study. The sequential nature of the construotion from one subunit to another combined with very detailed cost and benefit estimates made for a very useful data base. Secondly, in setting up an alternative investment pattern against which to measure the impact of the project a "do-notmngn alternative has been used. This is in contrast to the "do-it-latern model used for the northwest and reflects the environmental differences between the two areas. Substantial rainfed production would still be possible in the northeast even without irrigation while such is not the case in the northwest. 99. The principal conclusion of the study is that had cropping patterns used in the appraisal reflected those current immediater! prior to undertaking the 'project, no serious error of judgement would have been made. The move out of cotton and into maize and sorghum was already well documented at the tim~ of the original appraisal, but these changes were ignored or considered unusual aberrations. Whenever individual crops show substantial differences in their net social or private values, sensitivity analysis should be used to examine the response of the return on the project to marginal swings in the cropping pattern. Though nobody would have been likely to predict the complete elimination of cotton that has taken place, it has been shown that simply the most recent cropping pattern would have been adequate for a good guide to the projects' profitability. Current Bank practices would probably have picked up this aspect ot the project. 100. As in the northwest, projected private, on-farm, investments have not so far been realized. Similarly, there is no empirical evidenc~ to sup~ort the contention that undertaking any such inves'inaftt will now be in the soCial interest. Considering the project as a whole, however, the low revised returns tor San Juan and Rio Bravo (7 and 11 per cent respectively) suggest that great care should be taken in studying the impaot of any fUrther investment in these two regions.  ANNEX I -'- Meaicali10 --~ UNITED STATES " ", t .... " " " ._.-. -. :-._,. ., .J Chihuahuo o " '" " "\. "' •.1" ' ........... .'", I O C/IJDAD DELle/AS \ , ."\ )" , \.... BAJO "/~~'~(llamorot SAN • ...-.' ....'-'~ BAJO RIO .RA.fQ7I1 RIO CULIACAN Gulf of M~xico MEXICAN IRRIGATION ~EHABILITATION PROGRAM IRRIGATION DISTRICT LOCATIONS • NORTHWEST o NORTHEAgf o 100 200 500 4()0 &00 Kill. I , I I ' I  ANNEX II Table 1: CULIACAN - PUBLIC INVESTMENT IN REHABILITATION (Values in thousands of pesos) ProEosed Actual Percent change in -.. Item No. Value No. Value unit cost DRAINAGE ",'"' Improve existing system 207 Ian 4,882 282 1an 7,160 +23 Structures existing system 3,000 6,098 Construction new drains 930 Ian 40,959 643 lan 21,273 -10 Structl~es new drains 20,000 16,8.36 CANALS Improve existing canals 270 km 6,933 714 1an 13,703 -56 Structures existing canals 14,200 11,198 Construct new canals 191 Jan 3,913 Structures new canals 4,954 Canal lining 1,000 arHER ~RKS Road improvement 146 Ian 4,380 138 km. 6,554 +58 Telephone network 1,000 964 Metering and weather sta. 200 223 Housing 50 900 16 628 +118 EQUIPMENT iftnor 500 For maintenance 5,000 4,761 DIRECTION AND ADMIN. 5,946 9,499 CONTINGENC IES/ERRORS* 11,100 +867 TGrAL COOT 120,000 108,631 * Actual values do not. aa4 IUP since several data sources ,were utilized.  ANi~EX II Table 2: FUERTE - PUBLIC mVESTMENT IN REHABILITN~'IOl; (Values in thousands of pesos) FUERTE Proposed Actual Percent change Item . . No • Va4:u.e lTo7 Value in unit cos t. DRAINAGE Imp rove en sting system 1,182 km 11,477 1,891 38,413 structures existing system 6,484 13,812 Construction new drains 1,783 km 74,538 1,352 39,909 -15 Structures n-sw drains ~ 12,665 16,042 CA.L\fAtS Improve existing canals 664 kIn 11,110 711 12,937 Structures existing canals 8,713 4,636 -9 Cbnstruct new canals 558 Ian 14,352 573 12,103 structures new canals 12,595 6,417 -33 Canal lining 1,000 2,932 OTHER WORKS Flood defffnse 9,600 .Ro ad impro vemen t 15,000 602 18,672 Telephone network 1,600 2,154 Metering a~d weather eta. 1,000 Housing 3,000 64 3,118 EQUIPMENT Minor 1,800 19,OO07Ht- For maintenance 18,540 17,468 DIRECTION AND ADMIN" 5,211 21,241 OONTINGENCIES/ERRORS* 23 z000 +110 TOTAL COST 253,000 228,964 * Actual values do not ad.d ,up since several data sources were utilized. ** Listed as "materials".  ANNEX II Table 3" MAYO - PUBLIO INVESTMENT IN REHABILITATION (Values in thousands of pesos) Pro~~ed A"tm~J Percent change , Item No. Value No. Value in unit cost DRAINAGE Improve existing systf!m 241 10n 8,111 206 Ian 6,058 +45 Structures existing system 666 4,791 Construction new drains 813 Ian 51,229 506 lan 19,671 - structures new drams 11,579 13,415 -15 CANALS Improve eXisting canals 89 Ian 2,010 515 kIn 8,514 Structures existing canals 8,2S0 17,810 Construct new canals -56 66km 1,623 Structures new canals 1,114 Canal lining 1,461 2,334 OTHER WORKS Road impro vemen t 2,000 88 ion 2,961 Telephone network 500 858 Metering and weather sta. 200 199 Housing 23 494 24 1,150 +123 EQUIPMENT For maintenance 5,500 5,177 DIRECTION AND ADMIN. 2,000 9,785 COlm NGEl'lCI ES/ERRORSr.- 1°1 000 +198 TOTAL oo8T 104,000 95,658 * Actual values do not " add up since several data sources were U'\~:llized. .: '.  AHN.I!.:X II Tabl~ 4: YAQUI- PUBLIC INVEsrMENT IN R.EHABILI\rArrI01~ (Values in tho~ds o:t pesos) Proposed Actual .... It.em- No. Value No. Value DRA.mAGE 26.000 25,644 Improve existing system 60 km 5,890 306. kIn Structures extsting system 1,110 Construction new drains 16,743 342 lan structures new drains 2,257 EQUIPMENT For maintenance 5,200 4,885 DIRIDTION AND ADMlN. 320 2,923 COOT INGENC IES/ERRORS* 348 TOTAL COST 35,000 33,452 * Actual values do not add up since ~everal data sources were utilized. -..  ANNEX II Table 5: PUBLIC INVjSTMENT IN REHABILITATION. LOAN 275 (Values in thousands of Pesos) Percentage Item pro po seq ~ Actual % Dif.ference (E.XtU.Conttng. ) ~~ DRAINAGE 21.1,660 61 229,122 50 -15 CANALS 81,624 19 104,188 20 +27 OTHER WORKS 39,874 9 56,481 12 +42 MAINTENANCE EQPT. 36,540 8 32,291 7 -12 DIRECTION AiID AD~rrN. 13,477 3 43,448 9 +222 CONTINGENCY/ERRORS* 44,448 1,175 TOTAL COST 487,623 466,705 -4 ~~ Resulting .from use of several data sources for actual values • .... ,  ANNEX II Table 6: SAN JUAN - PUBLIC INVESTMmT IN REHABILITATION - (Values in thousands of pesos) Item No. Value No. Value DRAINAGE ~oving existing system 484 km 9,640 ConStruction 327 laD. 10,890 9 .. Structures 1.200 21,730 CANALs IMprove existing canals COn'sttuct new canals structures 4,540 C:anal l:ining 27" 32,1 0 orimR WORKS SUB TarAL (WORKS) 53,870 70,098 MAC'Ii'INERY AND EQUIPMENT 7,830 3,372 nIRECT:tON AND MAINTENANCE 19,300 11,502 CdN''ifINGENCIES 18,000 TOTAL COST 90.000 84,972  ANNEX II Table 7: RIO BRAVO - PUBLIC lNVESTMENT IN REHABILITATICll (Values in thousands of pesos) - '" Item Pro;eosed Actual Percent change in No. Value No. Value unit cOlt DRA:mAGE Improve existing system 415 km 13,300 862 lea 27,887 +0·7 Const:-uction 630 Ian 31,320 593 laIl 25,391 -13.9 Structures 42,220 43,790 CANALS Improve existing canals 260 lorn 3,700 152 loa 3,700 +73.8 Construct new canals 70 lent 1,980 28 km 606 -23.5 Structures 15,800 11,934 Canal lining 166 km 49,400 69 kIl 28,701 +39.8 OTHER WRKS* 2,770 SUBTOTAL (WORKS) 1$7.7$0 142.069 MACHINERY AND EQUIPMENT 14,000 5,300 DI~TION AND ADMINISTRATION 28,250 30,57~ CON'l'INGENC IES .,0,000 TOTAL COST 250,000 180,715 * Road improvement, cleaning and c,ontro1 of rivers, etc.  ANNEX II !!ble 8: DELICIAS - PUBLIC INVESTMENT IN REHABILITATION (Values in thousands of pesos) Item Pro;eosed Actual No. Value No. Value DRAINAGE Improve existing system Conatruction 110 Ian 6,100 Structures 2,700 ~ .. CANALS Improve existing canals Construct new canals Structures 9,200 Canal lining 224 Ian 25,400 SUBTOTAL (WORKS) ~3,400 55,254 MACHIN:ERY AND F4JIPMENT 4,900 4,725 DIRECTION AND ADMINISTRATION 7,700 8,657 CONTINGENC IES 14,000 TOTAL COST 10,000 68,637  ANNEX II Ta.ble 9 Public Expenditure in Loan J~6 Irrigation Areas (Thousands of Pesos) Delicias Rio Bravo HioSan Juan ° & 1'1 Pro,ject O&M P:rojj~ct ° & 1\01 Project .-~ 1960 4,667 10,935 4,533 1961 h,583 - : 11~147 4,179 .. 1962 L~,360 ": 11,482 4,361 1963 4,688 10,060 3,656 1964 3,975 8,590 9,718 30,299 4,022 9,104 1965 3,904 26,627 11,964 47,093 4,320 23,323 1966 h,591 16,138 10,566 26,509 3,968 34,184 1967 5,599 17,281 13,480 31,59'1 5,150 18,361 1968 6,458 12,838 36,341 4,478 1969 27,160 To~eAL 68,636 198,999 84,972 Table 10 Project Costs z Rio Bravo bZ Units UNIT Year (!, (~, (~) 'l'OTAL -- , -"- 1964 26,126 4,173 1965 ?5,32l 30,299 18,346 3,426 L~7 ,093 1966 6,872. 15,707 3,930 1967 6,648 26,509 22,473 2,476 31,597 1968 7,816 28,525 1969 36,341 6,193 20,967 27,160 TaL1AL 64,967 74,708 59,324 198,999 ....  ANNEX III ··iII I i "'[- Table 1 COMPARISON OF PROJECTED LAND RECUPERATION Al'J"D ACTUAL RECUPERATION 1961-2 TO 1968 NORTHWESTERN PROJECT Area RecuEerated District Land Class ApEra1sal Projection Actual Culi8.can: Moderate 28,000 1,350 Severe* 7,000 9,650 Fuerte: Moderate 100,000 56,019 Severe 50,000 20,878 Hayo: }loderate 8,000 1,363 Severe 14,000 18,803 Yaqui: Moderate 22,000 130 Severe 8,000 -130 7~ "Severe" is here equated with "Heavily" or "Ve17 Heavily Affected" areas. 1/ (' ANlJEX III Table 2 DE~.LINZA:.'.Cl:ON • ~ ., ~-,~ ......, .. ..,....,. ....... CULIACAN .. Hectares Gain 1961 to Salinity Cla.ss 19b1 19b6 19b8 1968 J Has. a.. Unaffected 64,000 71,000 75,000 lJ.,POO In production 23,000 23,000 21,6.50 1,350 Out of production 15,000 8,000 5,350 9,650 Total Area 102,000 102,000 102,000 Table 3 DESA.LINIZATION. FUERTE Hectares Gain 1962 to Salinity Class 19b2 ~96.F'· . "19b? ~9b8 1968, Has. > •• j Unai'fected 83,953 74,508 150,640 160,8.50 76,897 Light 86,239* ) 40,737 30,220 56,019 )138,804 Heavy 4J,120* ) 21.2.22 ~22, 242 20,878 "A.......... ". I .. , Total Area 213,312 213,312 213,312 213,312 * Tbe total sa1t-affec'oed area of 129 ,3.5'9 ha~ .• wa~ distributed in accordance with the appra.isa~ prop.ortions of 2·::L formoderate1y an.dser1ously aft ected :t.aIld~ .•  ANNEX III u. Table 4 DESALINIZATION. MAYO -' .liec"Gares Gain 1962 to Salinity Class 19~0 1902 1 [ 19 4 19~~ 19b5 1968, Has. ,.r. Unaffected 64,106 57,616 66,286 75,609 77,782 20,166 Light 9,149 12,002 1l,916 10,.301 8,777 3,225 Heavy 4,659 2,429 5,560 4,824 4,291 -1,862 Very heavy 14,086 19,953 8,234 1.266 1,l50 18,803 Total Area 92,000 92,000 92,000 92,000 92,000 Table 5 DESALINIZATION. YAQUI Hectares Gain 1961 to Salinity Class* 19~1 19 4 12?b 19~B 1968, Has. Unaffected 1,30,600 130,600 130,600 130,600 -2.... Light 60,501 61,113 60,215 60,411 90 Medium 5,675 6,671 5,192 5,635 40 Heavy 23.220 21,912 2),689 23,650 -130 Total Area 220,296 220,296 220,296 220,296 .! * The figures in this table are as given by SRH except that the additional class lien recuparacibn u (being rehabilitated) for 1964 on has been dis- tributed among the classes l:l.ght, medium and heavy according to the pro- portions given by SRH tor 1969. While the derived figures for 1964 and 1966 may be in·error, -the over-all change from 1961 to 1968 'is unchanged.  Annex IV . Table 1. Culiacan. Financ::ia1 return to S.R.H. from rehabilitation (Values in thousands of pesos) Year Water Charges Operation and Excess Collections Collected Public Works Cost Ma.in·~enance Total Due Rehab. Cost L 1960 4,827 4,850 -23 0 1961 6,203 7,322 1,119 0 1962 8,349 8,273 -76 0 26,819 1963 8,383 8,817 -434 -434 29,359 1964 9,377 7,323 2,054 2,054 40,607 1965 10,415 9,164 1,251 1,?51 11,846 1966 12,500 9,674 2,826 2,826 1967 14,578 12,245 2,333 2,333 1968 16,720 13,707 3,013 },013 1969 2,721 onwards 2,721 ~O ;zear life 6o-zear life Internal Rate of Return (%p.a.) 0.26 1.51  Annex IV Table 2 Fuerte. Financial return to SeR.H. from rehabilitation (Values in thousands of pesos) Year Water Charges Operation and Excess Collections Public 1,riorks Collected Maintenance Cost 1'otal Due Rehab. Cost "- 1960 12,346 9,037 3,309 0 1961 12,762 10,060 2,702 0 1962 12,405 8,499 3,906 0 57,241 1963 14,420 9,720 4,700 1,398 57,241 1964 15,874 11,123 4,751 1,449 80,137 1965 15,456 13,092 2,364 -938 34,345 1966 19,563 15,282 4,281 979 1967 19,530 15,763 3,767 465 1968 19,168 16,785 2,383 -919 1969 onwards 3,474 172 40-year life (jJ •.year life Internal Rate of Return (% p.a.) Negative Negative ______________ "!. _.....-...................... ' ........... :l\,;;..r'I_ _ _..-...-_ _ _ _ _ _ _ _ _ _ __ '.  Annex IV I able 3 Nayo. l Financial return to S.R.R. from rehabilitation (Values in thousands of pesos) Charges '\va ter Operation and Excess Collections Public Works Year Collected l.faintenance Cost Total Dle Rehab. Cost ..1- 1960 4,230 6,416 -2,186 0 1961 6,415 6,280 135 0 1962 7,468 6,056 1,412 0 23,914 1963 4,903 5,472 -569 ... 569 23,915 1964 9,138 3,634 5,504 5,504 33,480 1965 6,281 7,483 -1,202 -1,202 14,349 1966 9,617 6,588 3,G29 t 3,209 1967 9,375 7,933 1,442 1,442 1968 11,615 9,611 2,004 2,004 1969 on~ds 1,723 '1,723 ~O . ~ear life 60-'l.ear life Internal Rate of Return (% p.a.) Negative 0.47  Annex IV Table 4 : Yaqui. Financial return to S.R.H. from rehabilitation (Values in thous~ds of pesos) Year Water Charges Operation and Excess Collections Public ~Jorks Collected Maintenance Cost Ifotal Due Rehab. GOSii -- 1960 9,109 7,283 1,286 0 1961 11,078 7,164 3,914 1962 1l.,768 7,861 3,907 ° 0 7,859 1963 13,891 9,201 4,690 8,224 1964 16,743 10,719 6,024 ° 2,L~hG 11,)13 1965 14,786 10,714 4,072 488 6,356 1966 14,515 10,319 4,196 612 1967 18,098 11,856 6,242 2,658 1968 22,241 18,199 )-1-,042 458 1969 5,142 1,558 onwards L~O-~ear life §O.-yep.r life Internal Rate of Return (% p.a.) 3.60 }, .30 .. -"" .,,"'" .. ".  Annex IV Table.5 N. W. Project. Financial return to S.R.R. from rehabilitation (Values in thousands of pesos) Years \vater Charges Operation and Exdess Collections Public ~rks .r Collected ~~intenance Cost Total Due Rehab. Cost 1960 30,512 27,586 2,926 0 1961 36,458 30,826 5,632 0 1962 39,990 30,689 9,301 0 115,333 1963 41,597 33,210 8,387 395 118,739 1964 51,132 32,799 18,333 ~1,447 1('5,737 1965 46,938 40,453 6,485 2,003 66,896 1966 56,195 41,863 14,332 7,446 1967 61,581 47,797 13,784 6,898 1968 69,744 58,302 11,442 4,556 1969 onwards 12,613 6,174 gO-lear life 6O-"y:'ear life Internal H.ate of Return (% p.a.) Negative Negative  Annex IV Table 6 - DESALINIZATION. RIO BRAVO Sallnity Clua Hectares - 1964 1966 1967 19§tL -- Ar£~eted 108,168 76,186 78,737 62,052 Unaffected 125,572 162,.5.54 153,811 176,688 Out of Production 5.000 _AI - Total Area 238,740 238,740 238,740 238,740 Table 7 - DESALINIZATION. SAN JUAN S~ity Class Hectares 1963 1265' 1966 1967 1968' Affected 8j.752 5,560 3,813 2,h13 2,345 Unaffected 6~,.605 65,797 67,544 68,944 69,012 Out of Production 1.000 Total Area 71.,357 71,357 71,357 71,357 71,357 '-  ANNEX V Estimating Yield Changes free of Area Effects in Northwest 1. Yields on the principal crops in Mexico's rehabilitated irrigation districts have changed over time for three distinct reasons: (a) Technological advances have tended to increase yields. J (b) Substantial changes in the area of a given crop tend to have an inverse effect on average yields. As the area shrinks it is the best of the farmers who remain in that crop, or at least those with a Aomparative advantage. Rapid area expansion or contraction in response to price and cost incentives has occurred in several crops. (c) Weather variability has produced many fluctuations. The following model is nominated as a means of isolating these three infiuences: (1) Where Y is yield in kilograms per hectare, t is time (1959-60 • 1), A is area harvested in hectares and b o' b1 , b 2 are coefficients estimated by least sq~ares from the logarithmic form of (1). In this model, b l is an elastic! ty of yields wi th respect to time and may be evaluated as a -time trend rate of growth at the mean percentage change in the time variable. For the data used, 26.8 b l , is the average annual rate of change of yield attributable to general sliifts in technology and other, unspecified, time related phenomena. The coefficient b2 is an elasticity with respect to area and should have a negative sign. Random weather effects would be accounted for by the regression residuals (Yt - It). 2. Tables 1 to L show the results from multiple regression estimation of (1) for all important crops in each of five west coast irrigated areas. Although the regression was not constrained to give such a re'sult, none of the crops examined demonstra.ted exponential growth rates. This is to say, all yield growth net ot area effects was seen to be taking place at a decreasing rate over time. There were a very few cases of siRnificant but negative yield growth rates. Sesame in Culiacan and Mayo was the only notable one. 3. The next step in the analysis. is to isolate the sources of this "technological" change and determine the extent to w.hich it would have occurred without the contemporaneous rerhabili tation' project. The most obvious source of improvement has been ris'ing use of fertilizers. Table 5 summarizes the available information at the district level. In all cases the level of fertilizer use was quite high by 1965-66, though there was more room for improvement in the southern areas of Culiacan and Fuerte than  ANNEX V Page 2 in either Mayo or Yaqui. The rather startling shifts in application rates are probably due to changes in the form of fertilizer used as well as year to year changes in water ·availabili;ty. The declining proportion of land fertilized in Yaqui refiects a big increase in the area pla';~ted to soybeans (unfertilized) as a second crop. The increase in second crops is to be regarded as a direct result of the rehabili tation program but, in view of the quite advanced technology employed in the area y it seems likely that as much as 80% of the yield increases that took place in "winter" crops would have occurred even without the project.  Annex V Table 11 Yield Analysis s Culiacan MOdel, Yt. bo t b1 Ab2 u CroE Coefficients (tit" values in parentheses) . ~ bo bl b2 Time Trend R No . . . Name (Time) %p.a. (Area) (or r) Notes 1. Ajonjoli 8.531 (Sesame) -0.138 -3.7 -0.197 0.608 (2.078) (2.049) 2. Alfalfa 11.650 -0.022 D.S. -0.142 0.839 (0.283) (2.203) 3. Algodon 7.049 (Cotton) 0.082 n.s. Ca) (0.700) 0.256 h. Arroz 7.425 0.278 (Rice) 1.4 (a) 0.953 (0.033) 5. Cana 13,779 (Sugarcane) 0.223 6.0. -0.300 0.878 (3.910) (0.948) 6. Cartamo 7.431 (Safflower) -0.295 -7.9 (.) 0.598 (1.975) 7. Frijo1 6.861 0.052 (Beans) 1.4 (a) 0.390 8 years'data (1.121) 8. Garbanzo 7.795 0.106 2.8 -0.198 (Chickpeas) (1.004) 0.727 7 years' data (2.694) 9. Jitomate 9.409 (Tomato) 0.192 5.1 (a) (1.132) 0.393 10. Haiz 8,293 (Maize) 0.175 h.7 -0.146 (6.539) 0.927 (3.133) 11. Melon 8,366 0.282 .7.6 (a) (1.861) 0.575 12. Sandia 14.403 (Watermelon) -0.445 -11.9 -0.779 0.513 .... (1.136) (1.816) 13. Sorgo 9.174 (Sorghum) 0.725 19.4' -0.266 0.881 (4.491) (2.95S) .r 15., '!'rigo (Wheat) 7.116 0.318 8.5 (b) .. (3.331) (a) wrong a1gn when included (b) Not signj1icantlY different from zero when included Crops ~ (Soya) and 16 (Linseed) had less than 7 observations •. ,,. , . ;l :~::_ : ~ t .. , . ~ .  Annex V Table 2. Yield Ana1ls1s R10 l Fuerte .!bdelr Yt • b o t~~ Ab2 u ...... ¥i Crop Coefficients .cltt" values in parentheses) . .. bo b1 Time Trend b2 i Np. Name ('lime) %p.a. (Area) (or r) Notes 1. Ajonj011. 6.929 .0.219 -5.9 {a} 0.497 8 years (Sesame) (I.S14) 3. Algodon 12.734 0.492 1).2 -0.557 o.8~9 8 years (Cotton) (2.862 ) 1.146 4. Arroz 7.601 0.079 n.s. (a) 0.318 8 " (R1ce) (0.888) 5. Calla 11.607 0.195 5.2 (b) 0.919 8 (Sugarcane) (6.157) " 6,. Cartamo 7.582 .0.• 327 -8.8 (a) 0.550 8 It (Safflower ) (1.741) 7, Frijol 15.777 -0.078 n.s. -0.944 0,,654 8 (Beans) (0.421) (2.031) " B" Garbanzo· 7.581 -q.~S7 -6.9 (b) 8 (1",321) 0.447 " 9 .. J1tomate 8.300 O.P6S 17,,8 (a) 0.433 8 II (T~to) (1.272 ) lQ._ Hriiz 1,0)8 0.257 6.9 (a) 0.774 8 It (Maize) (3.229 ) 12.~ Sandia 9.609 -0.216 -S,,8 (b) 0.481 7 years I data (Watorme1on) .(1.:452 ) 1)~ Sorgo 8.167 0.419 11.2 .0.087 0.828 7 years l data (~gh1Dll) (3.095) (0.993) IS, '1l'igo 7.77Q ;0.1(6 2.e (a) 0.478 8 years (Wheat) (i.•439 ) ... 16,f Linaza 6.881 0.248 6.6 (b) 0.724 7 years' data (LiDaeed) . (2 ;77.5) (a~ Had wrong sign when includ;p. (p~ Was ~s1gnif'icant~ d1rrer.~t fr~ zero wP~ iAc,l:'ll,ded Crf1PS 2 (AlfaJ1'a Verde), 11 (Melpn..) ap,d 14 (.SPraJ had less than 1 o.b'servations available.  Annex if Table 3 z Yield Analysis. Rio Mgo CroE COS!.t""'gnt;. -' Ift".l~ :!:a lW"enthes.es l ..... bo bl Time Trend No • Name b2 Ii (Time ) %p.a. (or r) Notes I. Ajonjoli 7.875 -0.137 -3.7 -0.ll7 0.784 (Sesame) (2.653) (2.117) 3. A1godon 7.459 0.074 D.S. Ca) 0.300 (Cotton) (0.831) 8 years' data 6. Cartamo· 7.574 0.289 7.7 -0.070 0.759 (Safflower) (2.607) (O. 709) ·7. Frijol 6.637 0.504 (Beans) 13.5 (b) 0.961 (9.212) 9. Jitomate 9.090 -0.195 n.s. (a) 0.276 (i'enato) (0.759) 10. Maiz 7.255 0.492 13.2 -0.040 0.912 (Maize) (5.908) (O.785) 15. Trigo 7.641 0.175 4.7 (a) 0.731 (Wheat) (2.834) 8 years' data 16. Linaza 7.107 08096 2.6 (b) 0.488 (LInseed) (1.460) (a) Had wrong sign when included (b) Not significantly different from zero when included Crops 2" 4" 5, 8 .. 11, 12, 13, 14 had less than 7 observations. () ...  Annex V I Table 4, Yield Analysis! Yaqui Valley \ . arenth~ses CroE bo R ~.. No. Name (or r) Notes 1. AjOn~Oli 6.175 OC.21B Sesame) 5.8 (b) 0:872' (4.704) 3. A1godon 9•.350 0.213 (Cotton) 5.7 -0.176 0.961 (4.774) (0.995J 6. Cartamo (Sarf1o'Wer) No significant relationship (b) 9. J1tomate 7.351 0.258 6.9 {b} 0.946 7 years' data (Tomato) (7.736) 10. Maiz 7.474 0.321 B.6 (Maize) (a> 0.B18 1 (3.761) 13. Sorgo 7.537 0'.449 12.0 (b) 0.690 (Sorghum) (5.174) 14. Soya . 7.121 0.409 11.0 ~Q.O.3S (Soybeans) 0.825 (l.841) (0.743) 15. 'rl-igo 7.762 O.122 c (Wheat) 3•.3 (a) 0;S7il (1.874) 16. Linaza 7.318 0.030.- n~~s·. (8) 0.1:33' (Linseed) {O".3S6} : '" ~ry \.I1."1'1'fI"l.,'f#' ",:'Jj: .. II-. ~4~ , A '. .. "'~<'* (a) v.rong sign when included (b) Not significantly different from zero when 1D'c-l.uded Crops. 2" 4# 5J 7, 8, 11, 12 had fewer than 7 observaUon'S•. ,;.  ANNEX V . t .  ,', ~. "-. C- ~able l, ~ost and Benefit Streams, Rate of Return to Rehabilitation, Culiacan (Values in thousands of Pesos) Public Expenditure Net Value of Increase in OutEut Year Project ~hange in Total At Domestic Cost At Shadow At Shadow Costs and Prices, Cost O&H and Prices Costs and Prices fruits and vegetables excluded 1962 26,819 26,819 0 0 0 1963 29,359 3,555 32,91L. 3,90L. 4,095 2,077 1964 40 ,605 805 41,412 7,274 7,732 5,004 1965 11,8!~6 1i670 13,516 9,11+2 9,872 8,436 :).966 2,380 2,380 8,589 9,719 9,485 1967 6,741- 6,741- 27,443 28,683 15,112 1968 8,116 8,116 32,951 33,877 16,321 1969- 7,428 7,428 30,197 31,280 16,046 2006 Internal Rate of Return: 16.25 17.04 7.13 ~ ~ trJ ?< .q H  Table 21 Cost and Benefit Streams, Rate of Return to Rehabilitation. FUERTE (Values in thousands of Pesos) Public Expenditure Net Value of Increase_~ Output Year Project Change :in Total At Domestic Cost At Shadow At Shadow Costs and Prices, Cost O&M and Prices Costs and Prices fruits altdveg:etables excluded 1962 57 1 241 57,241 0 0 0 1963 57,241 693 57,934 12,427 12,626 13,486 1961.,. • it ,~ 80,137 2,49~ 82,632 21,332 21,410 21,20) 1965 -?4,345 4,237 38,582 70,492 72,333 48,672 1966 6,930 6~930 176,451 175,379 59,900 1967 8,213 8,213 l75,658 182,394 74,761 1968. 9,840 9,840 98,976 103,787 L.1,568 1969- 9,026 9,026 13~,317 143,090 58,164 2006 Internal llate of itet urp: 44.36 h5.23 21.56 >. :z: ~ t=:i >4 .q H ," t, ," . 1  \, ... .~ ~ab1e 3~ 80st and Benefit Stre~~, Rate of Return to Rehabilitation, MAYO (Values in thousands of Pesos) Public Expenditure Het Value of .Increase in OutEut Year Project Change in .'otal At Domestic Cost At Shadow At Shadow Costs and Prices, Cost O&H and Prices Costs and Prices fruits and vegetables excluded 1962 23,914 23:91h 0 0 0 1963 23,9l5 -9l4 23,001 2,916 2,971 2,865 1964 33,480 1,356 32,124 13,29< 12,527 12,465 1965 14,349 1,719 16,068 10,707 10,972 10,340 1966 -658 -658 14,516 15,426 14,772 1967 2,034 2,034 14,458 15,333 14,788 1968 4,299 4,299 28,172 27,431 26,570 1969- 3,166 3,l66 2l,315 21,382 20,679 2006 Internal Rate of Return: 18.39 18.47 17.81 ~ ~ ~ -== H  Table 4~ Cost and Benefit Streruns, Ra.te of Return to Rehabilitation, YAQUI (Values in thousands of Pesos) Public E!Penaiture Net Valu.e of Increase in OutEut Year Project Change in Total At Domestic Cost At Shadow At Shadow Costs and Prices, Cost 0&1,1 and Prices Costs and Prices fruits and vegetables excluded 1962 7,359 1,359 0 0 0 1963 8,224 3,950 12,174 3,611 2,,805 2,793 1964 n,513 6,067 l7,580 -1,963 -1,013 -1,012 1965 6,356 4,295 10,651 923 988 985 1966 4,104 4,104 h,917 5,614 5,589 1967 6,596 6,596 -9,)10 -8,619 -8,600 1968 13,019 13,019 -12,612 -12,559 -12,416 1969- 9,508 9,808 -10,961 -10,589 -10,508 2006 Internal nate of rte turn: ~~egative negative .~egative ~ 2;. z: tr: >< <: -; t '. ~  ~ I' II., I • Table 5~ ~ost and Benefit Streams, Rate of Return to Rehabilitation Loan 2 'Values in thousands of Pesos' Public Expenditure rre-=CV-aIUe~rmcrease J.Il OUtput Year Project Change in Total At Domestic Cost At Shadow At Shadow Costs and Prices, Cost 0 &M and Prices Costs and Prices fruits and vegetables excluded 1962 115,333 115,333 0 ° 0 1963 118,739 7,284 126,023 22,858 22,497 21,221 1964 165,737 8,011 173,748 39,935 40,656 37,6&J 1965 66,896 11,921 78,817 91,264 94,165 68,433 1966 12,766 12,766 204,413 206,138 89,746 1967 23,584 23,584 208,249 217,791 96,721 1966 35,274 35,274 147,487 152,536 72,043 1969- 2~,,429 29,429 177,868 185,163 84,381 2006 Internal Rate of Return: 29.97 28.84 12.43 ~ ~ x <: H  J\Ni.JJ~X Vl Table 6, Heconstruction of performance in YAQUI under hypothesized alternative investment •. ,,,1,, P&sica1 Area ~Th. HasJ Gropping Het value Year Salinitl class inte:r;sity of Productiun Unaffected Saline Out of prod r n /' ('fh. Pesos) '\. 1960 130,600 66,176 23,520 112.13 151,el~r( 1961 130,600 66,176 ?3,520 131·57 2(39 J~Y7 1962 130,600 66,176 23,520 129.89 3TI,ub? 1963 139,h66 59,893 20,937 117.80 . ' 37' 1 3'u.::>, .. 1964 148,332 53,610 18,354 1?6.80 1!39,llJ5 1965 157,200 47,326 15,770 119.76 33Li.,L40 1966 J..66,066 41,043 13,187 103.00 ?90,3t~ 1967 174,932 34,760 10,604 138.79 395 ,6:4~ 1968 183,800 28,476 8,020 14L. ·3? 439,))'/0 Table 7 Cash Flow' for hypothesized alternative investment in '(AQUI (Thousands of Pesos) PUblic ~!Eenditure [Jet Value of Lncrensc Year Project Change Fa 'rotal in Output at 6hndow Costs Cost o & M a) and Prices j 1962 44,269 J.~4,269 1963 44,269 478 l~, 747 4,897 1964 .61,976 1,721 63,697 10,751 1965 26,561 2,923 29 Jl484 5,6<.15 1966 4,182 4,782 3,325 1967 S,667 5,667 7,730 1968 6,790 6,790 9,131 1969- 6.228 6.228 8,if20 2006 TarAL 4~balOO· 361,779 " (a) 69 percent of the c:qange observed in F'U]m:rE where proport iona tely more recuperation was achieved in a district of abou.t the same size  ANN~ VII Table 1. - Culiacan Value Cost Net Value Increase in Het 'Value over 1960-62 Average Area Year (Th. Pesos) (Th. Pesos) (Th. Pe~,os) (Th. Pesos) ./ - .1!!:.. 1960 78.,206 278,402 177,080 101,322 1961 102,254 283,680 241,004 42,676 1962 107,842 )24,906 206,145 118,761 1963 98,280 364,556 166,050 198,506 110,919 1964 105,371 435,986 191,347 244,638 157,052 1965 120,093 536,902 220,651 316,251 228,664 1966 142,745 556,135 248,304 307,831 220,244 1967 129,863 735,788 254,915 480,873 393,287 1968 156,797 963,)13 301,330 661,983 574,396 .' Fyrte 1961 163,376 438,622 243,117 195,504 1962 153,997 397,978 233,733 164,244 1963 146,847 411,436 234,779 176,657 ··3,217 1964 156,376 469,840 279,736 190,103 10,229 1965 158,527 647,427 328,073 319,353 139,479 1966 181,204 973,962 375,160 598,801 418',,927 1967 163,999 926,857 395,806 531,051 351,177 1968 172,266 799,748 442,456 357,291 177,h17 !!!l2. 1960 56,173 166,699 92,01l 74,688 1961 80,188 246,299 124,544 121,755 1962 71,652 191,567 103,726 87,841 1963 68,162 173,408 95,374 78,034 -16,727 1964 68,798 237,324 115,009 123,315 27,553 1965 79,706 203,021 129,926 73,094 -21,667 1966 88,265 239,893 155,358 84,534 -10,226 1967 84,555 246,892 162,268 84,624 -10,136 1968 106,680 381,500 224,442 157,058 62,296 ... Yaqui 1960 220,648 492,083 340,235 151,847 .,( 1961 258,916 652,146 362,608 289,537 1962 255,610 745,532 368,444 377,087 1963 234,853 657,7 68 353,586 304,182 31,357 1964 256,079 84.3,468 410,040 433,427 160,60,3 1965 244,953 736,290 409,020 )27,270 54,445 1966 213,325 684,228 402,736 281,492 8,667 1967 291,041 895,973 512,669 )83,303 110,479 1968 306,,366 1,031,126 611,807 419,319 146,494  Table 2 Culiacan: AsSUllled Out co_ of De~d Rehabilitation Reeuperation Starti!!ll96~ Recu;eeration Startins 1966 Crop Net Public Crop Net Public Phlsical Jrea ~Th.Has.2 Ha. Value Inv. P Ha. Value Inv. ~ !!!! & 22l I !'Ii. Pesos Th. Pesos ~ Th. Pesos ~sos 1963 59.60 24.ho 18.00 ll7 196,777 29,757 1966 46.40 28.60 27.00 190 296,,626 40,471 ~964 62.67 24.40 14.93 121 242,915 37,197 1967 49.56 28.60 23.84 166 464,,029 53,961 1965 65.75 24.40 ll.85 133 3lh,~02 44,6)6 1968 52.72 28.60 20.68 192 639,.546 67,452 1966 68.82 24.40 8.78 153 306,1,,30 22,318 1969 55.89 28.60 17.51 169 551,)68 53,961 1967 70.37 23.70 7.93 138 478,580 14,878 1970 59.05 28.60 14.35 163 553,207 40,,471 1968 71.91 23.00 7.09 165 658,,715 0 1971 62.21 28.60 ll.19 157 554,918 13,490 1%9 73.46 22.30 6.24 149 567,274 0 1972 66.47 26.27 9.26 154 559,693 0 1970 75.00 21.60 5.40 148 568,672 0 1973 70.74 23.93 7.33 151 564,274 0 1974 75.00 21.60 5.40 148 568,672 0 Recuperation Sta,rting 1964 Recu;eeration Starting 1967 .!!!!: ~ 1964 55.20 25.80 21.00 1.30 240,,414 37,825 1965 59.01 25.80 17.19 141 311,573 47,,282 1967 42.00 30.00 30.00 180 457,547 46,522 1966 62.81 25.80 1).)9 161 304,018 56,738 1968 45.60 30.00 26.40 207 631,63) 62,<129 ~967 66.62 25.80 9.58 140 475,105 28,)69 1969 49.20 )0.00 22.80 180 545,3)5 77,,537 1968 68.71 24.75 3.54 167 654,,775 18 .. 912 1970 52.80 30.00 19.20 173 547,786 62,,029 1969 70.81 2).70 7.49 151 564,581 0 1971 56.41 30.00 15.59 1.65 550,060 46,522 1970 72.90 22.65 6.45 149 566,649 0 1972 60.01 30.00 1l.99 159 552,152 15,507 1971 75.00 21.60 5.40 1L8 568,,612 0 1973 65.00 27.20 9.80 155 557,921 0 1974 70.00 24.40 7.60 151 563,422 0 Recuperation Starting 196$ 1975 75.00 21.60 5.40 148 568,672 0 !2!: • Pesos Recuperation Total Public Investment Starts (Year) (Thousands o~ Pesos) 1965 50.80 27.20 24.00 153 307,822 JL,420 1966 54.20 27.20 20.60 175 300,495 45,893 1963 148,789 1967 57.61 27.20 17.19 153 469,798 57,)67 1964 189.129 1968 61.01 27.20 13.79 177 647,142 45,893 1965 229,469 1969 64.42 27.20 10.38 156 557,635 34,420 1966 269,809 1970 67.06 25.80 9.14 154 560,506 1l,u7) 1967 310,149 1971 69.71 24.hO 7.89 152 56),)00 0 1972 72.35 23.00 6.65 150 566,021 0 19{3 75.00 21.60 5.ho 148 568,672 0 ~ ~ <: H H I " ~ .j  J, .... " Table 3 Fuerte: Assumed Outcome of Delayed Rehabilitation Rec:!!!eration Startini ~963 Rec~ration Startin~ 1966 Crop Net Public Public ~sical Area ~Th. HaS~ Ha. Value Inv. !nv. !!!!: ~ sat 2 .L 'fli. Pesos Tli. Pesos ~ ~sos 1963 70.86 90.96 51.48 90 169,880 54,979 ~966 31.44 105.24 76.62 132 510,668 61,818 1964 95.43 73.70 44.17 92 185,600 68,723 1967 52.38 92.45 68.47 113 467,922 82,504 1965 119.99 56.45 36.86 89 315,905 82,468 1968 73.32 79.67 60.32 112 323,502 103,131 1966 ~.56 39.19 29.55 98 598,680 41,234 1969 94.26 66.88 52.16 104 416,456 82,504 1967 148.64 36.94 27.71 88 529,067 27,489 1970 115.20 54.09 44.01 99 429,606 61,878 1968 ~52. 73 34.70 25.87 91 35h,709 0 1971 1)6.13 41.31 35.86 94 439,731 20,626 1969 156.81 32.45 2h.04 88 446,891 0 1972 144.39 37.60 31.31 92 442,785 0 1970 160.90 30.20 22.20 87 448,457 0 1973 152.64 33.90 26.75 90 445,690 0 1974 160.90 30.20 22.20 67 446,457 0 Recuperation Starting 1964 Recu~ration Starting 1967 Crop Net Public E&siCal Area (Th. HaS~ Ra. Value Inv. ~ Free Silt _ .L 'fn. Pesos Tn. Pesos ~ 1964 57.72 95.72 59.86 101 175,608 64,154 1965 85.73 77.11 50.46 97 302,719 80,192 1967 18.30 110.00 85.00 127 437,4)6 66,760 1966 113,74 58.50 41.06 105 578,701 96,231 1968 41.31 96.40 75.59 125 )06,831 91,680 1961 141.75 39.90 )1.65 90 526~203 48,ll5 1969 64.31 82.80 66.19 ll4 401,204 114,600 1968 146.5h 37.h7 29.29 93 352,987 32,077 1970 87.32 69.21 56.78 107 415,293 91,680 1969 151.33 35.05 26.93 90 44h,952 a 1971 110.32 55.61 47.37 101 427,785 68,760 1970 156.11 ,32.62 24.56 89 446,726 a 1972 133.33 42.01 37.96 95 438,936 22,920 1971 160.90 30.20 22.20 87 448,457 0 1973 142.52 .38.07 32.71 93 442,294 0 1974 151. 71 34.14 27.45 90 445,462 0 RecuE!ration Startini 1965 1975 160.90 )).20 22.20 87 446,457 0 Crop Net Public P&siCa1 Area ~Th. HaS~ Ha. Value Inv. !!!:: Free Silt ~ ..L 'rn. J5esos 'rn. Pesos Recuperation Total Public Investment 1965 44.58 100.48 68.24 109 283,908 54,997 Starts (Yearl (Thousands of Pesos) 1966 68.17 85.51 59.62 117 5h4,1132 73,329 1967 91.76 70.54 51.00 101 496,278 91,661 1963 274,895 1968 115.35 55.57 42.38 100 341,691 73,329 1964 320,772 1969 1.38.94 hO.60 33.76 93 440,507 5h,997 1965 )66,648 1970 1971 14U.43 11,9.92 38.00 35.ho 30.67 27.98 92 90 4h2,';89 44h,606 18,332 0 1966 1967 h12,524 458,401 ~ 89 h46,561 0 ~ 1972 155.41 32.80 25.09 <I 1973 160.90 30.20 22.20 87 448,457 0 H H  Table 4. MaYo: Assumed Outcome o£ Delayed Rehabilitation Recuperation Starting ~963 Recuperation Starting 1966 Crop Net Public Crop Net Public ~ ~ ~ ~ ~ ~ ~sos Physical. Area (Th. Has.) Ha. Value !nv. Physical Area (Tn. Has.) Ha. Value Inv. !!!!: !!.!£ @ ..!.. Th. Pesos Th. Pesos Th. Pesos 196) 54.88 14,72 22.40 97 77,856 21,686 1966 h6.72 15.68 29.60 141 82.,637 21,966 1964 61.89 14.73 15.38 89 122,795 27,107 1967 52.35 15.69 2).96 124 82,947 29,288 1965 68.89 14.74 8.37 95 73.,221 32,529 1968 57.98 15.70 18.32 144 154,221 )6,610 2966 75.90 14.75 1.35 97 84,588 16,264 1969 63.61 15.71 12.68 120 116,692 29,288 2967 76.37 14.32 1.31 93 84,547 10,843 1970 69.24 15.72 7.04 112 117,092 21,966 1968 76.85 13.88 1.27 117 156,671 a 1971 74.87 15.73 1.40 105 117,442 7,322 1969 77.32 13.44 1.24 105 118,228 a 1972 75.84 14.82 1.33 105 117,755 0 1970 77.80 13.00 1.20 105 118,378 a 1973 76.82 13.91 1.27 105 118,066 0 Recuperation st~in~ 1964 1974 77.80 13.00 1.20 105 118,378 0 Net Public Recuperation Starting 1967 ,,'alue !nv. Crop Net Public ~ ~ ~ - ,---- ---- - ( "010 .Q!!! l Th. Pesos Th. Pesos Physical Area (Th. Has.) Ha. Value Inv. 1964 52.16 15.oh 24.80 102 121,778 24,220 __ Year __ Free __ Salt __ OUt __ it Th. Pesos _ __ Th. Pesos 1965 59.96 15.05 16.99 106 72,754 30,275 1967 44.00 16.00 32.00 140 82,153 23,866 1966 67.76 15.07 9.18 106 84~166 36,330 1968 50.10 16.01 25.88 161 153,037 31,822 1961 75.55 15.08 1.37 93 8h,]6o 18,165 1969 56.21 16.02 19.77 132 11,5,963 39,778 1968 76.12 14.56 1.32 117 156,360 12,110 1970 62.31 16.04 13.65 122 116,490 31,822 1969 76.68 14.04 1.28 105 118,022 1971 68.42 16.05 7.53 113 116,941 23,866 77.~l 0 1970 13.52 1.2L 105 118,200 0 1972 74.52 16.00 1.42 105 117,331 7,955 1971 77 .80 13.00 1.20 105 118,378 a 1973 75.62 15.oh 1.34 105 117,680 0 Recuperation Starting 1965 1974 1975 76.71 77.80 14.02 13.00 1.27 1.20 105 105 118,029 118,378 0 a Crop Uet Publin Phvsica.l Area (Th. -iClS.) aa. Value Inv. Year Free Salt OUt !Z rh. Pesos th. Pesos -- -- -- -- -- Recuperation Total Public Investment 1965 49.1,h 15.36 27.2<1 123 72,058 20,065 Starts (Year) (Thousands 0:£ Pesos) 1966 55.aS lS.37 20.75 123 83,1:21 ('6,7!ilJ 1967 62.33 15.38 14.29 lOG 6),665 33,442 1963 108,431 1968 68.77 15.40 7.84 126 155J L50 26,754 19611 121,101 1969 75.21 15.1:1 1.38 105 11'; ,:>51. 20,065 1965 133,772 1970 75.86 11.81 1.34 10, H'I,760 6,688 1966 11,6,""2 !i 1971 76.51 J..h.20 1.29 105 117,966 0 1967 159,113 tzj 1972 '11.15 13.60 1.25 105 118,172 a x 1973 77 .80 13.00 1.20 105 H8,378 ,. S -.! ~ ..  ~ .,l r J" ~ Table 5. Yaqui: Assumed Outcome of De~ed Rehabilitation ternate Investment Model Recuperation starting 1963 Rec~eration St~in~ 1966 Crop Net Public Physical Area (Th. Has.) Ha. Value Crop ~ fr!! Inv. Net Public 2!.!i Q!!! % ~n. Pesos ~h. PesOB P2lsical Area {Th. Has. ~ Ra. Value Inv. ~ ~ ~ ~ 1.96) 124.94 68.36 27.00 121 302,936 l Th. Pesos ~sos 1964 136.98 60.27 23.05 39,757 1966 107.96 1965 129 437,558 49,697 74.8h 37.50 n6 275,746 149.03 52.18 19.. 10 121 334,216 1967 117.29 69.20 33.82 39,589 1966 161.07 44.09 59,636 1968 126.61 156 379,429 52,785 1967 15.14 103 290,685 29,818 63.55 30.14 161 419,469 166.75 40.19 13.36 140 398,920 1969 135.94 57.91 26.46 65,982 1968 172.44 36.29 19,878 1970 145.26 154 405,057 52,786 1969 11.57 146 437,814 0 52.26 22.77 151 408,793 178.12 32.40 9.79 141 1971 154.59 46.62 19.09 39,589 1970 183.80 420,908 0 1972 148 412,393 13,196 28.50 8.00 140 423,058 164.33 40.58 15.39 0 1973 174.06 145 416,076 0 34.54 11.70 143 419,629 Rec~eration starting 1964 1974 183.80 28.50 8.00 0 140 423,058 0 Crop Net Public Rec~eration Startin~ 1967 Phlsica1 Area (Th. Has~ Ha. !!!!: !t!! §!!f _ i Value fn. Pesos Inv. Tn. Wsos Crop Net Physical Area (Th. Ras.~ Public 1964 119.28 Ra • Value Inv. 7(\·52 30.50 134 ~ ~ ~ 1965 132.49 61.99 25.82 125 429,740 328,907 4h,100 55,125 ~ .L Th. Pesos ~sos 1966 145.70 53.46 21.14 1967 102.30 77.00 1967 107 286,573 66,150 1968 41.00 162 373,)24 158.91 44.93 16.46 142 112.33 71.09 36.88 42,846 1968 165.13 395,179 33,075 1969 167 413,)58 57,128 40.82 14.34 143 435} 029 22,050 122.35 65.19 32.76 159 1969 171.36 36.71 J:~.23 1970 132.38 59.28 399,710 71,410 1970 143 418,438 0 1971 28.64 155 403,901 177.58 32.61 10.11 142 420,771 JJ.,2.40 53.37 24.53 57,128 1971 183.80 2B.50 0 19'12 152.43 152 407,915 42,846 B.OO 140 423,058 0 47.46 20.41 149 1973 162.89 h1.14 4n,764 14,282 16.27 146 415,681 Recuperation Starting 1965 197h 173.34 34.82 12.14 0 1975 183.80 143 419,442 0 28.50 8.00 140 423,058 0 PhYsical Area ('I'h _ ~"'., , "'''-U}J n_ I'lel; .PUblic !!!!:... 1965 113.62 72.66 3L.OO 131 Recuperation 1966 121,.40 322,530 36,332 Total Public Investment 65.95 29.94 112 280,579 Starts (Year) 1967 135.18 59.23 25.89 hB,443 (Thousands of Pesos) 1968 145.97 149 366,372 60,551, 52 So 21.83 154 IJ27,441 1963 1969 156.75 115.77 17.7e 48,443 198,789 1970 163.51 147 1113,014 36,332 1964 220,502 h1.46 15.33 145 415,615 12,110 1965 1971 170.2·{ 37.14 12.69 242,216 1972 177.01, 144 1J16,154 0 1966 263,930 32.82 10.44 142 1967 1973 183.80 28.>~ 420,6JL a 285,644 ~ B.oo 140 423,058 0 :.:: ~ < ~  Annex VIII Table 1. CULIACAN: THE COST OF CO!1MENCING REHABILITATI01~ OTHEH. THAN IN THE OPTIMUM YEAR. (THOUSANDS of 1961YESOS) Discount. Year Rehabilitation Starts Rate (%) 1962 1963 1964 1965 1966 1967 "" ,.J ;;; 91,661 51,092 1.1.4,933 Optimum 20,380 35,784 6 13,064 4,349 13,764 Optinrum 30,478 .J 55,906 7 Optimum 9,998 28,183 32,249 67,027 96,687 8 Optimum 20,560 43,363 57,172 92,923 123,548 9 Optimum 26,104 50,976 69,645 104,520 134,305 10 Optimum 28,580 53,983 74,592 107,608 135,614 15 Optimum 24,42? 44,516 59,964 80,951 97,675 20 Optimum 16,332 29,819 37,442 49,394 58,035 25 Optimum 10,053 18,b$9 20,605 26,818 30,621 J  Table 2.• FUERTE: THE COST OF GO¥Jlv'lENCING filllLu.BIJ.JI'l'ATION QTHm? THAN IN 'rilE OPTIMUM YEAR. (THOUSAJ.1 DS of 1961 PES~) Disc 0'(%1 Rate . Ye~~ Rehabilitation Starts 1962 1963 1964 1965 1966 1967 5 93,776 Optimum 22,322 39,688 94,684 147,15h .~ 6 58,403 Optimum 32,hL,2 70,119 150,973 207,725 8 21,379 Optimum 39,157 91,176 168,31L 222,876 10 6,660 Optinmm 37,850 89,019 157,255 ?04,525 13 537 Optimum 31,552 73,583 127,669 163,248 14 302 Opti.mum 29,217 67,807 117,60? 149,687 15 493 Optimum 26,911 62,134- 107,918 136,774 20 4,101 Optimum 16,773 37,655 67,445 B3,907 25 8,442 Optimum 9,166 20,055 39,211 L~8 ,036 r  AnneK VIII Table 3. MAYO: THE COST OF COMMENCn~G REHABILITATION arHER THAN n~ THE OPI'IMUM YEAR. <:TI-IOUSANll3 of 1961 PES'OS) Year Rehabilitation Starts DiSCO(%~ Re.te ~b 1962 1963 1964 1965 1966 1967 5 958 12,392 11,630 Optinnlm 23,616 5,084 .; 6 Optimum 12,891 14,630 8,462 31,839 16,703 8 Optimum n,747 15,155 13,367 33,959 22,349 10 Optimum 8,735 11,846 10,623 27,577 17,566 14 Optimum 2,707 3,789 886 11,363 2,872 15 1,404 2,854 3,403 Optimum 9,159 937 16 3,5L~8 3,851 3,887 9 7,970 Optimum 17 6$367 5,624 5,173 854 7,728 Optinllun 18 8,934 7,238 6,330 1,606 7,497 Optimum 19 11,261 8,699 7,360 2,271 7,272 Optirm.un ~O 13,36h 10,012 8,273 2,855 7,055 Optimum 25 21,118 14,730 11,376 4,810 6,046 Optimum '. '  Al1Il8X V.I I r il Table 4. YAQUI: THE COST OF COMMENCING dEI-fABILITA~eIOr~ OTHER THAN IN THE OPI'IMUM YEAR. (THOUSANDS of 1961 PESOS) Discount Year Rehabilitation Starts Rate (%) 1962 1963 1964 1965 1966 1967 c~ ~ 160,515 56,092 46,517 Optimum 19,294 20,210 6 84,254 13,107 14,859 Optimum 17,815 26,865 8 35,901 Optimum 11,269 16,148 37,931 51,987 10 21,288 Optimum 12,817 22,208 42,023 55,046 15 13,092 Optimum 8,143 12,917 23,995 30,317 20 14,118 Optimum 2,801 1,495 6,151 7,654 21 14,895 386 2,278 Opt:i.mum 3,713 4,536 22 17,028 2,104 3,158 Optimum 2,854 3,.078 ?3 19,028 3,697 3,962 Optimum '2,077 1,767 24 20,893 5,160 4,682 Optimum 1,375 598 25 23,074 6,941 5,772 439 1,187 Optimum * !he model used here· is one ~ieh considers postponement of a pr.oject similar in seo})!' to tho~·~, in the other qistrlcts and not the. project actually undertaken. See paragraph 76. of the: text. ,. \.  ANNEX IX Cropping Patterns, Yields and Prices before and after Rehabilitation 1. The data in Tables 1 and 2 are from S.R.H. records and show averages, weighted by area and production as appropriate for two years prior to the rehabilitation program and for the last two years for which complete data were available. A heading such as "61-62" should be interpreted as indicating the average for the crop years 1960-61 and 1961-62. Yields are in metric tons per hectare and prices in pesos per ton. In the case of fruits and other crops, an arbitrary I-ton yield is assuned and the price given is actually the average gross value of production per hectare. 2. In Table 1 the crops are numbered not named. Correspondinp, names are as follows: No. Name No. Name 1 Sesame Ie Maize 2 Alfalfa ( green) 11 Cant loupe 3 Cotton 12 Watermelon L Rice 13 Sorghum 5 Sugarcane 1L Soybeans 6 Safflower 15 ~meat 7 Beans 16 Linseed 8 Chickpeas 17 Fruits 9 Tomatoes 18 Other crops -.- . ...  .. ~ "" ,) ~ Table 1.. District lata before and after Rehabilitation - Northwestern Project Distri.21.! - Ouliacan District 3 - :Mayo Per Gent .Area Yield Price Per Cent Area Yield Price Crop 61.-62 67-{;Jj 61-62 67-{;Jj 61.-62 67-68 Crop 61-62 67-68 61-62 67-68 61-62 67-68 S!SAHE 9.1.0 2.99 0.746 0.749 2365 2695 SFS.AME 15.82 8.05 0.754 0 .. 761 2447 2726 ALFALFA 0.10 0 .. 48 60.000 46.251. 100 ll2 ALFALFA 0.00 0.00 0.000 0.000 0 0 COT'IDN 6.19 1.20 1.4:13 1.455 2376 2n8 COTl'ON 24.94 15.61 1.728 2.336 2407 2590 RICE 16.58 20.41 2.211 3.107 899 ll44 RICE 0.00 0.00 0.000 0.000 0 0 SIDARCANE 20.25 l6e41 56.396 76.262 56 61 SUGARCANE 0.00 0.00 0.000 0.000 0 0 SAFFLO'Wm 10.97 8.17 1..224 0.947 ~1.61 1499 SAFFLOWER 3.44 8.17 1.352 2.050 1226 1471 BEANS 8.24 7.53 0.991 0.919 1749 1958 BEANS 1.50 0.63 1.088 2.275 1800 2584 CHICKPEAS 0 .. 21 0.92 0.887 0.536 1226 1813 CHICKPEAS 0.13 0.10 2.345 2.315 1248 1748 'lOHA'lO 1.2 .. 60 6.23 12.437 20.349 332 2342 TOMATO 1.80 0.29 16.200 7.030 958 1131 CORN 2.58 h.17 1.412 1.622 BOO 939 CORN 3.22 6.37 1.327 2.895 800 826 MELON 1.37 0.4.1. 6.652 8 .. 248 1.499 1272 MEIDN 0.00 0.04 0.000 4.773 0 998 WATmMELON 0.69 0.25 6.976 3.788 799 546 WATERMELON 0.00 0.09 0.000 8.341 0 1249 SOHGHUH 5 .. 39 19.81 l.538 3 ..188 677 625 SORGHUM 0.00 7.99 0.000 6.330 0 624 SOYBEANS 0.00 3.48 0.000 1.313 0 1599 SOYBEANS 0.97 7.83 1.500 1.957 1300 1598 liHEAT 0.39 3.14 1.427 2.334 9ll 826 WHEAT 37.49 37.54 2.562 2.767 865 822 LINSmD 0.00 0.00 0.000 0.000 0 0 LINSEED 3.47 1.68 1.101 1.558 1448 1544 PRUIT~ 0.42 l.04 1.000 1.000 5131 9390 FRUIT TREE) 0.00 0.10 0.000 1.000 0 68544 OTHEFS 4.92 3.35 1.000 1.000 9347 7784 OTHERS 7.23 5.51 1.000 1.000 3204 6519 Ifistrict 2 - Fuerte District 4 - Yaqui Per Gent .Area Yield Price Per Cent .Area Field Price Crop 61-62 67-68 61-62 67-68 61-62 67-68 Crop 61-62 67-68 61-62 67-68 61-62 67-68 SFSAME 8 .. 79 2.92 l.l13 0.719 2132 2233 SFS.AME 7.96 3.40 Q.630 0.772 2154 2612 ALFALFA 0.00 0.00 0 ..000 0.000 0 0 ALFALFA 0.00 0.00 0.000 0.000 0 0 COTl'ON 22.35 31.<S4 1.250 2.174 2000 2530 roTTON 23.40 20.29 2.097 2.533 2231 2452 RICE 24.97 2.34 2 .. 428 3.239 1031 1072 RI~E 2.82 0.00 2.307 0.000 916 0 SOOARCANE 7.63 6.70 126.48 166.97 49 59 SUGARCANE 0.00 0.00 0.000 0.000 0 0 S.A.FFIDWm 7.20 2.27 1.659 1,.714 1299 2316 Sm'IDWER 0.40 1.86 1.4ll 1.537 1265 1507 BEANS 10.51 6 ..1.7 0.7l.4 0.987 1199 2153 BEANS 0.22 0.01 1.054 1;5n 1976 2477 CHICKPEAS 0.32 0.59 1.899 l.369 669 1~60 CHICKPEAS 0.00 0.08 0.000 1.541 0 1957 IDMA.'lO 2.52 l.41 16.000 30.331 300 2954 TOMA'ID 0.03 0.01 1.864 3.000 2049 2800 OORN 4.l8 5.63 1.364 1.964 849 828 CORN 3.lU 10 .. 62 2.583 3.000 300 805 ~ tz:j MELON 0.00 0.13 0 .. 000 3.824 0 1004 MELON 0.00 0.00 0.000 0.000 0 0 ~ WL."'ERMELON 0.40 0.30 ll.382 10.363 677 400 WATERMELON 0.00 0.03 0.000 30.000 0 675 !;! SOIDHUM 1«21 10.58 2.206 4.043 662 567 SORGHUM 1.07 1.95 2.lU8 4.522 525 618 SOYBEANS 0.00 8.90 0.000 1..6n 0 1637 SOYBEANS 7.22 18.26 1.692 2.045 1274 1525 6.07 16.31 2.783 2.807 912 830 2~ WHEAT WHEAT 43.77 1l1.12 2.717 3.197 794 ~99 LINSEED 19~~1 0.70 0.38 1.156 1.706 O.~ 0.3~ 1.858 1.781 La~ LINSEED FRUIT TREES 0.06 0.26 1.000 1.000 000 FRUIT TRmS 0.0 0.0 1.000 1.000 OTHFES 3 ..ll 3.48 1.000 1.000 3394 5048 OTHERS 4.ll 1.94 l.OOO 1.000 254 3129  Table 2. District Data before and after Rehabilitation - Northeastem Project District 2. - Bravo District 6 - San Juan Per Cent ~ Yield Price Per Cent Area Yield Price Crop 63-64 67-68 63-64 67-68 63-64 67-68 Crop 63-64 67-68 63-64 67-68 63-64 67-68 - - SESAME 0.01 1.176 2499 ~TroN 23.74 1.35 0.418 1.685 2453 2577 COT1l",- 21.53 1 .. 69 0.643 1.402 ~h48 2553 BEANS 1.83 0.53 0.272 0.32) 2499 1796 (lWnfed) 2.82 0.477 2403 ToMATO 0.01 1.125 1544 BEANS 0.03 0.50 0.395 0.169 1.800 1799 MAIZE 36.43 36.16 1.862 1.872 889 790 (Rainfed) 0.18 0.649 1799 SORGHUM 34.64 61.13 2.079 2.362 563 612 TOMAro 0.02 0.00 8.250 3.000 400 1000 WHEAT 0.04 2.550 897 MAIZE 40.15 64.02 2.236 2.642 915 796 OTHE&S 0.42 0.79 (~ed) 4.32 0.12 0.519 0.750 899 800 (Reinfed) 2.94 WATERMELON 0.05 0.08 9.232 7.512 205 269 (Rainfed) 0.01 3.700 208 SORGHUM 14.58 32.66 2.025 2.920 576 613 District 7 - Delicias (RaL~ed) WHEAT 14.34 0.04 0.30 0.964 1..731. - 1.766 577 693 699 Per Cent Area Yield Price (Rainted) 0 .. 02 1.639 690 - FRUIm 0.01 0.10 561 4777 Crop 63-64 67-68 63-64 67-68 63-64 67-68 OTHmS 1.26 0.51 (Rainfed) 0.62 COTroN lWZE 30.22 23.19 20.73 16.35 1.290 ' 1..500 2.377 2.451. 1999 799 2500 939 TOTAL Irrigation TiT9 w.mr ~ SOEGHUM 1.43 7.95 2.119 3.000 603 624 22.31. 0.12 ~ Rainfed H WHEAT 26.24 36.52 3.000 4.000 883 912 >C PftUIm 1..20 1..03 7200 3998 oTHERS 17.72 1.7.42 2481. 948 r" , ~ \.  ANNEX X Shadow Costs of Production 1. An examination o,f the costs of production and their proportion to the value of production was made for the period 196L to 1969. Six crops were selected for the districts Rio Culiacan, Rio Fuerte, Rio Mayo, and Rio Yaqui. These crops, enumerated in Table 1, were those for which CO$ts were available and values of production proved to be eonsistently important over the years. 2. Costs were weighed by the areas harvested for each district and year to give the indexes shown in Table 2. These costs excluded taxes and rehabilitation quotas. Growth rates were derived from a comparison of the average of the indexes for the last two years, (1967-68), with the average for the first two years, (196L-65). 3. In Rio Culiacan, the cost of production rose from 10C per cent in 196u-65 to 109.6 per cent in 1967-68 or an annual growth rate of 3.12' per cent. Growth rates were hiRher in the three other districts. There was an A.5 per cent annual rate of increase in Rio Fuerte, 7.0 per cent in Rio Mayo and 5.5 per cent in Rio Yaqui. L• Costs of land preparation and fertilizer were singled out from the total. These items are of particular interest, not only because they account for a large portion of the total production costs but also, because they are the ~ractices for which the domestic prices of materials tend to be high relative to "world" prices. Table 3 gives the percentage of cost due to land preparation and fertilizer by year. Cotton was analyzed separately because of its relatively lower cost requirements as compared with other crops. These costs represented about 31 per cent of the total for cotton and about u6 per cent for other crops. There were no marked changes in land preparation and fertilizer costs from year to year. Whereas, in the case of cotton, land preparation represented 55 per cent of the combined land preparations and fertilizer cost, it was 60 per cent of this sum in the case of other crops. In general, these costs were split as follows: Per Cent of Total Cost Cost Item Cotton Other Crops Land Preparation 17 Fertilizer and Application til 31 ,. '" ... "  TABLE 1 ANNEX X Major Crops by District AREA HARVESTED DIsrRICT CROPS (PERCENT OF TarAL) r " Rio Culiacan Sesame, Cotton, Maize 44 and Rio Fuerte v/heat, Sorghum, Rice 69 Rio Mayo Sesame, Cotton, Maize 88 and Rio Yaqui Wheat, Sorghum, Safflower 88 T:ABLE 2 Cost Indices by District, 1964-65 and 1967-68 Cost Indices, All Crops IDIS l' aIC T ANNUAL GROWl'H IN TarAL COST 1964-65 1967-68 Rio Guliacan 100 109.6 3.12 • Rio Fuerte 100 128.8 8., Rio r~ayo 100 123.0 7.0 ./ Rio Yaqui 100 117.2 5.5 ) TABLE 3 LAND PREPA..l?ATION AND FERTILIZER COSTS AS A PERCENTAGE OF TOTAL COSTS BY DISTRICT AND YEAR