cm historical review program
STUDIES IN
INTELLIGENCE
A collection ol articles on the historical, operational, doctrinal, and theoretical aspects ot intelligence.
All statements of fact, opinion or analysis expressed in Studies in Intelligence are those of lhe authors. They do not necessarily reflect official positions or views of the Central Intelligence Agency or any other US Government entity, past or present. Nothing in lhe contents should be construed as asserting or implying US Government endorsement of an article's factual siatemenis and interpretations.
Direct and analog methods offoreign countries spend on atomic energy for rnSttary and peaceful uses.
COSTING NUCLEAR PROGRAMS Alan B. Snrth
How much has the Soviet Uruou. Conununistron Its nuclear program? What il the coat of the Frenchplant at Picrralatte or of the nuclear test rite InIs the allocation of funds for these mstaJlatkrrison schedule? How much has West Cermany spent onof nuclear research and development? What would itIsrael, or Japan to convert its present program forto production of nuclear weapons?
Tbe intelligence community Is frequently called upon to rupply answers to questions such at these for two primarygauge tbe burden nuclear programs impose oo the economies of the countries concerned, and to compare the sizes of different countries' programs.
Attempts to measure the economic burden are usually related to the question whether cost is apt toation from undertaking oreapons program. Analysis for this purpose of the pattern of spending also reveals much concerning the nature and probable rate of developmentrogram. Cost and rate-of-ei-penditure studiesseful approach to these problems.
Comparison of the sire of different countries' nuclearess cogent reason for estimating costs, and cost comparisons of this kind must be interpreted with great caution. Comparison of probable capacities for production of nuclear materials Is the direct and more appropriate way to get at the relative size of nuclear programs. Size can be measured in megawatts, quantities oforr numbers of weapons without involvement in complicated problems of monetary conversion. Conversionextensive studies of materials, manpower, wages, andin the nuclear industries of the countries compared, and the requisite data, as well as the time, for these are usually larking.
Two Methods
Thcte are two distinct approaches possible in estinuting tho cost of st foreign nuclear program. Both are quite useful if their limrta-tions are kept tn mind and they are not used to answer tbe wrong questions. One is the straightforward "documentary" method of examuung data nude available either by open sources in tbe country nbudgets, press releases, fournxl' articles,through dandestine procurement This method is particularly well suited to gauging the burden the nuclear program places on the economy. The costs thus obtained, being stated in the country's own currency, can easily be measured against native yardsticks such as gross national product and national income to determine the share of national resources being devoted to the nuclear effort. It is not well suited to comparing the lire of tbe foreign program with that of the United States, because of the monetary conversion problem. And it is not always praetJcable: the required documentary data may not be available.
Tbe ether method is to estimate by analogy, ia, to start from what it would cost tbe United States to build and operate the facilities known to exist in the foreign country. This method, if carefully applied,asis for comparing the size of the foreign nuclear program with that of the United States; ft is not well suited to detrsmsining the burden imposed oo the foreign economy. One of its obvious difficulties is imperfect knowledge of what is inside foreign plants protected by strict security" measures. Photographic and other typos of technical mtelhgence are useful in identifying tbe nature, and perhaps the capacities, of the plants, but estimates of their internal layout, equipment, and processes can at best be educated guesses. But these guesses must remain the basis for estimating cost by VS. standsarrs.
Moreover, even an accurate figure tor what it would cost. technicians, working at the present level. scientific and technical knowledge and with the resources. industry at their command, to reproduce and operate the foreign facilities may have little relevance to the question of what It Is costing, in terms of man-hours and material, for foreign technicians to construct and operate them in their ownwithifferent level of knowledge
aod industrial support. Tbe problems are well illustrated in tbe ease of the French gaseous diffusion plant at Pierrelatte. Tbe oSnal French estimate of the cost of this plant isillionnixlion at the current rate ofnd unofficialhave placed itillion or more'2his is about one-half of3 billion tbe United States spent for three gaseous diffusion pi* its, catch ofmuchlazgerthe Rearelatte rnstallafioo.
It ft true one should take thro account the huge ecooomies of scale achieved when the initial problems have been solved and unit sizes are increased. This can be attesnpfed by using the cost of early US faohbes roughly equivalentrrelatte. In thee0 roil lion into such facilities, adjustment to present-day prices would bring this up to the neighborhood0 million. So even with this adjustment Pierrelatte will cost fromoercent rnore than the. plant, not counting savings for the latter that would result from improvements in technology since. It is evidentrice tag put oo the fterrelattc plant on the basis of what it would cost the United States to constructacility today, at the present level. technology, would be so low as to be very misleading.
Since gauging tbe burden on tbe economy it the principal reason for estimating costs, the analog method should be used only when lack of docxrrnentary material makes ft necessary. Failure to keep in mind the limitations and proper orientation of Uie two methods bas sometimes led to confused interpretation and unfortunateof their results.
1 Doc, Assemble*HnVnOrdinaireUpport Fa* an Norn de I* Ccarimissioo don le Prctet da Let de Financesnnexeapport sur It* Credits do Muuitrre de*Annexe ao rarecea-vcxbal de la teener
U
When some documentary information is available but is anbasis for an estimate, analogy may be usedn the country under examination may resemble facilities of known cost in some other country. Tbe known costs, adjusted for evident difference in size or conditions, provide at least something to go on in the absence of hard data. The effectiveness of this mixed
method depends on the ratio of docuxnentary mfonnauoo to analog derivatives and on the comparability of the analog countries inscientific, and industrial development Circumstances, bow-ever, frequently make it the only practical means of estimating the burden imposeduclear program.
These different methods can be illustrated in their application to tbe nuclear programs of different countries.
The Documentary Method: France
The French case will illustrate the documentary method and alsoumber of problems encountered in analyzing the cost of nuclearof the military part of the costs, the allocation of joint costs essential to both rnilitary and peaceful uses, and the forecasting of probable future expenditure. Work was begun on the case soon after tbe first French nuclear test inith the objective of gaiiging the burden the program imposed en the French economy.
A wealth of scattered docuxnentary material was found to beData pahxstakingly assembled from open sources,by occasional documentary material clandestinely procured, haveeasonably dear picture not only of total annual and cumulative costs but of the allocation of funds to different kinds of activities withm the program, to various individual installations, and to capital investment and operating expenses.
the more [inportant source rMbarials were the anneal reports of theEocrgie Atorniqne (tbe French atomic energy commission) and of EJceOtcil* de Fiance (the nxtionslired powerfida! press releases, budget data pobUshed to tbe Journal Offidol de la WpuWiqveorn-mi tu* reports contained In official doeoxneats of the French National Assembly, press coverage of parliamentary debates on appropriations, and articles in nmnej-ons professional and trade journals.
ew franc equalld francs svai btrconced. Ccnt data ia old franc* were coe.ertcd to new bancs at thn rate. Dousx oocopanWco ate the official exchange rateew
Summation of published historical data indicated that by the end4 France had expended somealhontulon at the official exchange raten its nudcar program since it beganbe annual expenditures grew fromill Wo francs6 (all tn the budget of tlsenergic Atomique
at that time) to moieillion francs from all sources4 The sources of funds for the whole period break down as follows:
Facofr
CEA budget, ioaxa from tho Fond for Economic and Socialand Income from rale* of rwuJenrSt
Arasaafsrfataons for The Atom" tn the Defense
Investment by ElectricitG de France in nuclear power programs ., ther: Opea-attoaal expense' of EDF cossneetrd wtth rarcsear power ptograma, budget aQorations to tnteroatiooal axeacies, tnjuferi from tbe Ministry of Public WoiVj. and tn vestment by private
Onlyercent of this "estimate" of cumulative eipenditure involved any estimation whatever. That amount was necessary to fill in gaps in some series in the "other" category. The rest isurnrnarion of published data. But theillion figure must be regardedonservative estimate because it does not include some expenditures for mtcrnational cooperation. exper>ditures by theservices from their operating budgets, or moremall traction of the investment by private industry in new materials and equipment We know that such expenditures have been made but have no adequate basis for quantifying them.
The method by which the total expenditure was built up isby Table I. covering the period since The program has grown progressively more expensiveesult of its expanding size, generally rising costs, and greater emphasis on military aspects-Table 1
Experadunres on the French Nuclearmultoos of franca)
i960 1 2 3 4 5
Payments frees CEA
Payment* from Defense Budget*)
Other'
Total
tn CEA tad Defense bodgets. "Other" funds earned atna tdy tbe tamo rate asith allowance for punned Increase fn erpeoditare by EDF indicated In Lt
Footnote* continued on
CEA, Rapport5 Budget, Le Utmde,
Casts
o Moada.ate. Fans,p.
At-
ssemble* NaKocaL. Kffoft Fear aa> Horn,cmmtmtom dm Fmentm/Voter do Let do FmetumWfWfauace on
' EDF liiwtwul. from: Ambanade d*ervice ie Fresaa at dixAmeuaa. NT, Faaaoc aad Tho Atom. Job. OS,DF Itrpport d'Aettotli. Cemptm do Ceasson.. ILnd EDF Iror-sua
Inb-inatbmal Cooperation from: Asnltassade de Fiance, Francetale, Fans,, ISomte, tflfieUl.; Tho Norntate, OutfAif, Aa-grana9lStale, Vacnaae,.7
TrasMSar from Msnirtry of Fabric Works: State, Park,, ISbe smallorooa of tbe "other* erpendirurra cam* bans ^cattarad references pertaining to eipenditurta by private Industry.
These annual costs were then rotated to the French gross national product at current marVet prices, ar fn Tableo provide an Indicator (admittedly Imperfect) of the burden oo the economy. It wuthat In spite of sharp increases In costs the nuclear program ts well within the capabilities of the French economy.
Table f
Coats of French Nudcar Program Related to CNF
1 2 - 3 4 5
Cross Nataoaal Product at Current Market Prioat (anauions of
banes)*0
Total ripcnditurn on Nuclear Program
(DtdUosu of francs) LSS1 6 nrnd.ru.-es am Ha-cteae Feojraa, as
fwrotl ol
ItcpabLOua Frssscaasa. Miniitfcre de* FtaaDce* et Amlnrutut Nataoaal de la Sutjstsoue et des Etadesta Paris, tofiranealcPrehmfaary) State.
'v Stoi
Intelligence bis been asked what portion of these French expeodi-tures went (or the development of nuclear weapons Theillion total4 includes, besides expendituresurely irulrtary character, funds spent on research for peaceful uses, oo International cooperation, oo electric power production, and on activities essential to both the military and non-military portions of the Program- We start with the obviously military appropriations for "the atom" in the defense budget, which we have seen to. be aboutercent of the total, orillion francs. But this ta not the entireshare. Tbe GEA annual reports described the defense budget funds as intended "to cover the expendituresost immediate, (or direct) mihtaryhat this does not apply to all expenditures for tbe military nuclear program Is confirmed by the fact that investments in facilities fawn, to be exclusively military exceeded total approprUtfoos for "tbe atom" as of the end
Then bow much of5 bulico francs from doo defense sources can appropriately be regarded as military? One canillion expended for clearly non militaryhis figure includes funds for international cooperation In nuclear development and mvestment by Eectricite de France In equipment for nuclear power stations. (Exclusion of the latter might beoo the grounds that such stations could producehe remainingillioo francs must be regarded as joint costs of military and ncm-military projects.
Allocation of these funds to military and oon-mihtary categories was extremelytudy was made of allocations to such categories as administration, research centers, caploration and mining, ore concentration plants, feed materials and fuel eterrtent fabrication facfliKes, the gaseous diffusioo plant, and development of new reactors and chemical separation facilities. The allocations to specificwere derived primarily from monetary and manpower data available fn French documents. As in any attempt to allocate jointreat deal depended on assumptions concerning each type of activity.
La tbe end, it was considered thatillion francs might properly be charged to military aspects. Adding this to the amounts from the defense budgetotal ofillion francs, approximately two-thuds of the total expenditure, associated with
the military fide of (be nuclear programhis figure includes fundi for both capital Investment and operating expenses.
Ofillion francs spent on military aspects of the programomeillion were spent during the, after tbe successful test in0 spurred plans for the Creadontrategic nuclearayments diarged tobudgets accounted for aboutillion of this, as shown Innd the militsry share of Joint costs during this period has been estimated atlDioo. The acMttiotud btTbon fn the higher figure it designed to take care of erpeoditures from tbe regular operating budgets of the armed services that arewith the development and testing of nuclear weapons but not Identified as expenditures for "thehe cost, for example, of army, naval, and air transport of personnel and equipment, the salaries and maintenance of military security details, and tbe cost of military participation fn testing or in tbe development of weapons or propulsion systems cannot be quantified precisely but should be counted.
Future Expenditures
Future expenditures on nuclear programs have In generaloo the basis of tbe past trend fn total annualis known of plans for investment in Dew facilities,their probable operating expenses, trends fn operatingexisting faciLtics, and past relations hips between capitaland operating expenses. Early3 figures for futureon the French nuclear programhole wereprojecting2th or!or theid Ctrtaiw EaMipmerUi UJUatrte at 8
bich scheduled funds daring theas come to be called the "Protoaror Tint Pi op am Law" nowSecond Programpproved by the French Parliamentmiies fortha eWveaopcoect of tbe ittelesac oordear force during tbe. LsrteBlgeaorreeprendr envjparc tbe period of tbe Erst pro-gram law with the forme.
'Much of tha work denen the cocti of tbe French prograni was baaed oo autbnriratforai rather than actual erpendinnrj beoauae data oo expenditures then ivallabW did notreakdown aa to either source or allocation of fundi. Inanaocb ax uriallocaled auifaorixatioiis are amply carried over to the neat year, the lag la expcadituret wasarlotaiin measuring eostiairly long spaa of years. Since tbe raidilrowei er, evtxmatn have beat bated ob expenditure data; even me cfaaiaw tor earlier yean have been recalculated on the basis oftuJ eipenJ ifiire data now availeblc for those yean.
a cent increase per year; tbe average annual increase la au-ejtorirabooi82 had been aboutercent The figures thus obtained were found to be roughly comparable to those derived by adding up probable investment in planned projects and probable increases in operating espeoses.
Now that what was future at tbe begirming3 has become past history, we have an Interesting opportunity to compare these projections with what happened We find that34 fall to the range between tbe actualand the expenditures for those years, as indicated In the following tabulation (in billions of trances):
Actual .ueWntirm
Preliminary. Based en budget data.
Revised at berimdog4 aitflioa.
The authorizations projected56 will msdcubtedry prove leu accurate. Bodge* data indicate that expenditures5 will probably be orth/ aboututton francs, and6 ft appears tbey wiD be fn tbe rangeillion francs. The margin of error oo any forecast tends to increase as the projection moves farther Into the future, and II was recogrilzed from the outset that the projected levels of eapenditure might not be achieved until later. They have, however, been useful as an approximation for tbe latter half of the decade
For the iniutary part of the future nuclear program one begins withillion francs appropriated in5 defense budget. To this, tf it is assumed that tbe military share in joint costs wiD be as high5 as tbe estimated annual average, can beillion tootalnilhon francs. This estimate based on appropriations Is probably conservative, because in recent years espenditures have tended to run higher than initial budgetary appropriations. Moreover, as the military programIn size, the military share of tbe joint costs should really rise over the average of the past five years.
rom what is known of plans for investment and weapons development and past relationships between mvestment and Operating expenses, the military expenditures should continue to rise. Complctioo of tbe Pierrelatte plant and the test site In the Pacific
Actual
alone account for somebillion francs, according to authoritative French sources. Adding the heavy expenditures for weapons tasting, construction of facilities for production ofnd tritium,developmentuclear submarine propulsion system,and operation of the plutonium separation plant at Cap de la Hague, and increased operational expenditures in general, it was estimated thatillion figure5 would increase toiuioniving by interpolationillion francs
With the completion,umber of important facilitiesnnual expenditures on the military program could conceivablyIt was estimated, however, considering the cost of operating the new facilities and further development of weapons and propulsion systems, that they are more likely to remainevel ofillion francs per yearhe ram of the annual amounts thenotal9 billion francs, or at the official conversion rate7 billion, for the ifac-year.
To7 billion, rounded toillion, which has become the central figure of the intelligence community's estimate, wasargin of error of plus or minusbillion, or nearlyange which is considered sufficient tor aboutercent confidence. Tho lower limit ofOhon would assume very little increase in annual expenditure above5 level Some increase is almost certain. Tho upper limit ofillion allows for an excess ofover authorizations in the Second Program Law, possibletn the military share of the joint costs, and service expenditures that are conDOCted with the nuclear program but not soudget accounts.
Analog Method: China
The costs of the Communist Chi nose nuclear program have been estimated entirely by analogy, because very fettle useful documentary information is available. Documentary materials and official public statements have helped to identify and describe some of tbe facilities, particularly some of the carry research facilities, but have given no indication of cost Most valuable in identifying and describing the nuclear installations has been Nationalist aerial photography.
Once the Chinese instaDations are identified and described, they have been related to roughly comparable facilities fneans of estimating costs. For example, photographic evidence suggests that the Chinese reactor at Pao-t'ou is very similar tot Mareoule in France. Irxforrnation released by the French
CEA0 mdicated that the original cost ofeactorUlioo (old) francs,njllion, so the one at Pao-t'ou was estimated tonillion.
Estimating thus on the basis of roughly comparable Westernit has been concluded that by4 the Chinese Cocrununists
been aboutillion.uess, the operating expenditures4 could haveillion.
These cost estimates, made on the basts of very sketchy information, are less precise than those on any other country's program. Moreover, the dollarndoubtedly an inadequate measure of cost to tbe Chinese economy in terms of scarce technical talent, materials, and industrial capacity. It ts, however. In line with costs elsewhere in the world; our estimate of what France spent in tbe period before its first nuclear test Isillion.
Mixed Method; USSR
Estimates of the cost of the Soviet nuclear program have been madeixed method. In the, when it was first undertaken
with VS. costs.
A description of Soviet facuities was obtained from the taterrogation of former prisoners of war who bad worked in or near them, and particularly from Cerman scientists and technicians taken to the USSR5 and used in the nuclear program until the. Papers delivered tn8 Geneva conference on peaceful uses and displays at the Soviet exiubrbon in New York9 also provided some [united information. Photographic and other technicalcontributed further to tbe description.
As installations were Identified and described, their construction costs were estimated from Soviet data on the cost of Other Industrial
construction Known Soviet electric power costs were applied lo estimates of their power cowumptxxi. Soviet cost data werefor tome of tbe chemioab used in tbe program. Tbe cost of uranium, which accountsubstantial portion of Soviet operating costs, was estimated from information on the cost of extracting and muling other minerals and from trade data on imported ore and concentrates. Soviet wage data were helpful in estirrating personnel costs. But mvestmentacihties could not be estimated from intelligence data; It was therefore assumed to have about the same relationshipestment In production faculties as in the United States.
One of the major deficiencies was in data oo the cost of equipment for the production installations. To fill this gap, US Atomic Energy Commission contractors were given descriptions of the plants and asked to estimate the cost of equipping them, breaking this down in considerable detail These detailed estimates in dollars were coo-verted into ruble costs by comparing US. price lists with available Soviet tuts of prices and specifications for wellifferent commodities and activities related to the nuclearhe different ratios were then weighted according to the relativein the program of the commodities or activities in question.
Early1 it could thus be estimated that the Soviet nuclear program bad required throughumulative expenditure ofill.on 1SSS rubles, aboutillion tn capitaland aboutillion for operating expenses.
1 additional inteUigence has made possible Unproved estimates of the capacities of both previously existing and newplants. The gap in information on RcVD facilities has been partially filled by analyzing fragmentary information on the number of personnel employed atenters. ecent estimate
' ResearchndkandBS5 ralxU-doBai ratio ofas appropriate for capital costs, reflecting averagef about il for labor aod naWrkJ*or equipment. (For certain types of ecuipcwM the ratio was as lowad for, butsafor portion It wasor operating costs aa average1 was derived from widely varying .aliosumber of inputs.5 tubWdoCar rank, tor wantons contrarratea produced to tbe USSR was esumatad atbe ratioIndustrial wages -as. with labor peoduc-twsty aa-eaaed to be about half that tn the Carted States. On the bask of comparative rats jehedrdea the ratio tor electric power was IS.I, though tkk probably maderstates rha difference inerjcy. The ratio fee cbeasscal prcdocts was)
places cumulative expenditures on the Soviet nuclear program throughtillion, new rubles (the new ruble equalsboutillion for plant and equipment andillion for operating purposes. Current spending is estimated atillion rubles annually, orercent of CNP.
Tbe estimate ofillion rubles expended4 isconservative. Recent rrdorrnaboo indicates, though notthat the ruble-dollar ratio used for estimates of capital mvest-merit may haveittle low. Mceeovex, the estimates themselves probably do oot make adequate allowance for cost of modernization, conversion to new processes with Improved technology, or complete replacement of facilities. An offsetting factor, however, is that In recent years improved mining and refining methods may have reduced the cost of Soviet feed materials somewhat more than estimated. The annual distribution of capital expenditures is difficult to determine, estimates have been based primarily oo observed construction tunes and oo analogy with experience In Western countries. In spite of these deficiencies, the estimates are considered good indicators of the magnitude of outlays for the Soviet nuclear program.
Presentation and litlerpretation
As we have pointed out, the costs of foreign nuclear programs are best studied in the indigenous currency, so they can be related to units of national accounting and expressedercentageative measure of the economy- In US. intelligence studies, however, they must be expressed in dollars. If told that West Germany spent aboutillion marks oo its nuclear program4 or that the annual nuclear expenditure in Japan as47 billion yen, tbe reader immediately asks, Tfow much is that inor this reason costs derived in indigenous currencies are frequently converted to dollars at official rates, as in tbe following tabulation: Esrrsaanm Torax.
1>ockam
4
19G4
to Cur-
Prices
Cerrrjany
8S0
'
tso
Although tbe official exchange rates may not accurately refled differences in the purchasing power of money spent on tbe nuclear Industry in the United States and in the country concerned, the dollar figures at least have the merit of being readily understandable and can be reconverted to the original currency with ease. And ennver-sion to dollars for purposes of presentation does not affect our measure of the burden oo the economy, the percentages of CNP in the table were computed entirety tn the indigenous currencies-*
Along with the desire to have the cost of foreign nuclear rsrograms stated in dollarsendency to compare the results of theIt should be remembered, however, that when the cost-version has been based oo official exchange rates rather than ratios derived from study of relative productivity, the comparisons can be only rather crude measures of relative sire.
Even if comparative costs could be accurately expressed in the sameurther difficulty in comparing size is that different programs are not homogeneous, aggregate costs do not reflect differ -ences in tbe nature of the programs- West Cermany, for example, has spent nearly as much oo Its nuclear program as France had by the time of its first test, or nearly as much as we estimate forChinese expenditureset West Cermany has no nuclear weapons program at all Its program has been oriented strongly toward education, research, and technological development aimed at developing low-cost, high-quality equipment, particularly reactors for electric power and for ship propulsion- The table shows that India has spent only about one-fourth as much as West Cermany on its nuclear program. It is true that the Indian program is much
ladJranouscatena! wss avaBabia to pensut fatdy precise cost estimates for the nuclear programs of Wert Ceimany. Italy. Sweden, India, aad Japan, both total coats and alloeatiOof among various MntUttoaa and types of aetretttea (except that iesoaxoaooo eaa la^aexoeat by private oegantra-bons was ks all eases uwdrquata, making all the estimatesor the Israeli nuclear program, however, the eattpaitat want reachedaised saetfaod. taxng some cost data ralaased by Israeli soaroeaskul*rtyberooipOooe of (he UcshOes obtained from observers and visitors, and analogy with the costs of riroJar facilities In other countraea. In estimating tbe costhermeal aeparabOa plant, for example, mould Israel elect to con-ifrurt one, ate wai made of detailed Information India bat released oo (he costs of its chemical separation plant. It wss assumed that Israel could buy tbe equipment for the plant at about (he same price that the Indians paid. Indian coottractson costs were adjiutrd according to the differeoo* In costs for labor and cement in Israel.
smaJIcr than the German ooe, bat although It li thut far oriented toward peaceful uses, it is in balanced that it now has all the facilities needed to produce the fissionable materialsmall weaponsSuch fads as these, which become apparent throughtbe allocation of funds to various types of rnstxfliiioru andare oot indicated in the overall costs.
xsors
As we have seen, the method of estimating the costs of nuclear programs> and to some extent the usefulness of thesually determined by tbe availability or lack of document my maferiaL The rnore documontxry evidence there is, tbe more the Intelligence problem becomes the traditional one of painstakingly collecting, combining, and analyzing the data. Itatter of indifference to the cost analyst whether the documentary materials were overtly published and procured or obtained by clandestinereat deal of documentary information on the costs of nuclear programs isopenly.
Unfortunately, the fact that information has been published abroad in the official reportoreign atomic energy commission or electric power monopoly or perhapsrade Journal does not necessarily mean that it ix available to tbe analyst in Washington. In most of our diplomatic missions abroad the pubbcations procurement officers have that respcaxsibility merely as an addition to other duties, and they must look for publicationside variety of consumers. Even if the analyst knowspecific publication andequest for il, the delay before he actually gets it may be considerable; but often he does not even know that such-andublication exists and therefore cannot request it
If all types of collectors were kept aware of the need forinformation on the cost of foreign nuclear programs they might pick up and forward useful material they ran across by chance in the course of cither activities. Clandestine source materials, both documentary and of an incidental conversational variety, have proved useful both in tilling gaps in overt information and as an aid inovert data. Sometimes they have lent credibility to overt materials that otherwise would have been disbelieved until confirmeduch later date
The supplementary method of estimating cost by analogy, used wben tbe documentary materials are not adequate, depends on col-
Coifs
lection ofdiffoeotofography and other types ot technical intelligence often malce possible the description of facilities that is necessary before attempting an analog estimate.
Collectors should think of the cost analyst as an insatiable sponge who wekornes data from any source on other the over-all costs or particular allocation of funds, for investment or operation, to any installations or activities coamexted with nuclear programs alnaoct the world.
Original document.
Comment about this article, ask questions, or add new information about this topic: