Allocation under Dictatorship: Research in Stalin's Archives · Allocation under Dictatorship: Research in Stalin’s Archives Paul Gregory and Mark Harrison∗ We survey recent research

Journal of Economic LiteratureVol. XLIII (September 2005), pp. 721–761

sp05_Article 2 8/18/05 9:39 AM Page 721

Allocation under Dictatorship:Research in Stalin’s Archives

Paul Gregory and Mark Harrison∗

We survey recent research on the Soviet economy in the state, party, and militaryarchives of the Stalin era. The archives have provided rich new evidence on the eco-nomic arrangements of a command system under a powerful dictator includingStalin’s role in the making of the economic system and economic policy, Stalin’s accu-mulation objectives and the constraints that limited his power to achieve them, thelimits to administrative allocation, the information flows and incentives that governedthe behavior of economic managers, the scope and significance of corruption andmarket-oriented behavior, and the prospects for economic reform.

∗ Gregory: Professor of Economics, University ofHouston, and Research Fellow, Hoover Institution onWar, Revolution, and Peace, Stanford University.Harrison: Professor of Economics, University of Warwick,Distinguished Visiting Fellow, Hoover Institution, andSenior Research Fellow, Centre for Russian and EastEuropean Studies, University of Birmingham. The authorsthank Eugenia Belova, Valery Lazarev, Roger Sherman,and the referees for their comments, and the HooverInstitution and its archives department, the NationalScience Foundation, the Leverhulme Trust, and theBritish Academy for financial and other support of theirresearch.

1. Introduction: A Well-DocumentedDictatorship

The Soviet planned economy was themost important socioeconomic experi-

ment of the twentieth century, and it under-pinned the century’s most elaborate anddurable dictatorship. It was a major obstacleto the study of this experience, however, thatthe Soviet state also took secrecy to anextreme.

When the Soviet Union collapsed in 1991,many historical secrets were laid bare.

721

1 The main federal archives of interest to economistsare the State Archive of the Russian Federation (GARF);the Russian State Economic Archive (RGAE); theRussian State Military Archive (RGVA); and the RussianState Archive of Social and Political History (RGASPI)and Russian State Archive of Contemporary History(RGANI) which together hold the records of the Sovietcommunist party. The Archive of the President of theRussian Federation (APRF) and the Central Archive ofthe Federal Security Service of Russia (TsAFSBR) wouldalso be of great interest were they not largely closed tooutsiders. A gateway to web-based information on for-mer Soviet archives is provided under the PoliticalEconomy Research in Soviet Archives website athttp://www.warwick.ac.uk/go/sovietarchives/archives.

These were documents never intended forpublication or scholarly research: secretplans, reports, minutes, decisions, appeals,and the official and private correspondenceof citizens from the highest authorities in theKremlin to the humblest provincial petition-er. In Russia, these records are now held bythe Federal Archival Service.1 The scale andscope of its holdings are vast and intimidat-ing, running to hundreds of millions of files.The Soviet state intervened everywhere,

722 Journal of Economic Literature, Vol. XLIII (September 2005)


2 Among the landmarks of this revolution see, in partic-ular, Stephen G. Wheatcroft and R. W. Davies (1985),Davies (1989a, 1996), Oleg Khlevnyuk (1993, 1996), E. A.Rees (1997a, 2001, 2004), J. Arch Getty and Oleg Naumov(1999), Yoram Gorlizki (2001a, 2002), Davies andWheatcroft (2004), and Gorlizki and Khlevnyuk (2004).

3 Khlevnyuk, Kvashonkin, Liudmilla P. Kosheleva, andLarisa A. Rogovaia (1995); Lars Lih, Naumov, andKhlevnyuk (1995); A. Berelovich and V. P. Danilov (1998);Berelovich (2000); Khlevnyuk, Devis [Davies], Kosheleva,Ris [Rees], and Rogovaia (2001); Khlevnyuk, Gorlitskii[Gorlizki], Kosheleva, Miniuk, Prozumenshchikov,Rogovaia, and Somonova (2002); Davies, Khlevnyuk, Rees,Kosheleva, and Rogovaia (2003). These include Stalin’scorrespondence, secret police reports, state planning com-mission and ministerial records, and Politburo reports.Another notable series covers agriculture and collectiviza-tion (Danilov, Roberta T. Manning, and Lynne Viola 1999continuing).

4 Forced labor: V. P. Kozlov (2004); the defense-indus-try complex: R. S. Ganelin (2003) and V. A. Zolotarev(forthcoming).

5 Further information is available at http://www.hoover.org/hila/projectsarch.htm.

standing guard over state property, publicmorality, and most aspects of social life; itsrecording of decisions and outcomes wasmeticulous. While political power and eco-nomic organization always rested on abedrock of informal relationships, the degreeto which the exercise of power was expressedin writing is nonetheless staggering. Therewas no question of Stalin ever avoiding for-mal association with difficult decisions orpassing them along by word of mouth.

Although Putin’s Russia has remainedprotective of Soviet secrets of the 1960s andsince, the opening of the archives has stim-ulated a true revolution in the history of theStalinist state.2 Many volumes of documentshave been published to aid those who arethemselves unable to work in the archives.3

While some sensitive documents of theperiod remain classified in the archives ofthe president and the security and militaryagencies, even the most secret institu-tions—the labor camps and the defenseindustry—are now being massively docu-mented by new collections.4 Significantmicrofilm holdings are also available outsideRussia, for example in the HooverInstitution.5

6 R. W. Davies, whose own work towers over the field,and others have surveyed the major findings from thestandpoint of Russian history (Davies 1989b, 1997, 2003;Sheila Fitzpatrick 1999; Khlevnyuk 2001a; GáborRittersporn 2001).

7 The economic journals currently archived at JSTORpublished 320 articles on aspects of the Soviet economybetween 1920 and 1991; half of these appeared in justthree journals, the American Economic Review, theReview of Economics and Statistics, and the Journal ofPolitical Economy. A catalogue is provided under http://www.warwick.ac.uk/go/sovietarchives/before.

This survey summarizes the first fruits ofarchival research on the political economy ofthe Soviet Union in the period of “classicalsocialism” (János Kornai 1992) that endedwith Stalin’s death in March 1953. We covertopics of interest to comparative economists,institutional economists, political econo-mists, and new economic historians. As forfindings of specific interest to country spe-cialists on Russia and the former SovietUnion, we discuss them only in passing.6

In making this survey, we are well aware ofthe debt we owe to previous generations ofwestern scholarship. Despite the obstaclesof secrecy and censorship, the quantity ofthis scholarship was large and its quality highenough to warrant many publications in topeconomics journals.7 As a fresh generationhas grappled with new evidence, there hasinevitably been a certain amount of rein-venting the wheel; it has been one of ourtasks to try to discern what is genuinely new.Following the organization of our paper, wesummarize this as follows.

Part 2 deals with the economic organiza-tion of dictatorship in a hierarchical com-mand system. Before the archival revolution,scholars had almost no direct evidence of thepolitical economy of centralized power; wesaw mainly what reached the public ortouched the lowest levels of economicadministration. In politics, Kremlinologistsmade an educated guess at the shifting influ-ence of moderate and radical factions; econ-omists tended to presume a powerful rolefor technocratic central planners. We had lit-tle idea of what Stalin really wanted or whatconstrained his ability to get it.

Gregory and Harrison: Allocation under Dictatorship 723


The archives demonstrate that Stalin’spersonal dictatorship was real and, from1932 onwards, unchallenged. Through histen-man Politburo, Stalin was continuallyimmersed in economic decisions. HisPolitburo guided the general direction ofthe economy by infrequent major decisionsrelating to investment and innumerabledetailed interventions relating to outputsand inputs. There were no moderates orradicals; the Politburo was unified on gen-eral strategy, but divided by the specialinterests of its members, the expression ofwhich Stalin particularly feared. The econ-omy itself featured extreme centralizationand the priority of vertical over horizontalrelationships. Given that the Politburocould make only a few of the decisions thatmattered, it delegated much of its power tolower levels in complex hierarchies to cre-ate a nested dictatorship of many smaller“dictators.” Since each player distrustedthose below him and those below knew anydecision could be revised by those above,even trivial issues tended to be passedupward. The result was a “dictator’s curse”:despite the intention to delegate, Stalin hadto make many more decisions than hewished.

Part 3 deals with the limits on Stalin’spower to extract a surplus. Before thearchives, we knew the published economicplans, decrees, and speeches that promisedan ever-brighter future for Soviet producersand consumers, but even the best informedcould only speculate as to the economic pol-icy that lay beneath (Eugène Zaleski 1971,1980). Most basically, we did not know theeconomic fundamentals in the dictator’sobjective function nor how he would tradethem against political and other objectives.

The record of Politburo decisions, com-bined with our knowledge of the informa-tion that Stalin monitored most closely,suggests that Stalin aimed to maximize theeconomy’s surplus, defined as output lessconsumption. His horizon was distantenough to make accumulation his first use

of this surplus as long as war did not imme-diately threaten. Stalin saw this surplus asdepending critically on worker incentives;he managed these by giving unflaggingattention to consumption. While keepingconsumption low, he feared that at any timeit could fall so low as to promote workerrevolt. This accumulation model explainswhy investment became volatile. The work-ers’ consumption minimum was not a givenand Stalin sought to reduce it by coercivemeans. The archives have recorded the con-sequences of extensive coercion, many ofthem unintended and perverse. Having cre-ated a surplus, Stalin had to hold rent seek-ers at bay. For the first time, we can witnessthe rent-seeking process. Special interestsformed immediately within the dictator’sown circle, leaving Stalin and a few associ-ates to battle for the interests that theyviewed as encompassing.

Part 4 deals with allocation in the economyas a whole. Before the archives, we alreadyknew a great deal about shortages, queues,and corruption. The relationship betweenthese was hotly debated: were shortages theunintended result of the famous “soft budgetconstraint” (Kornai 1980), for example, orwere they intentionally created for corruptpurposes (Andrei Shleifer and Robert W.Vishny 1992)? By implication, if shortages ledto corruption, was corruption the intended orunintended consequence?

Archival research has shown clearly howthe soft budget constraint operated andwhere it came from. It originated in Stalin’sindustrialization drive and was perpetuatedby his inability to commit to financial disci-pline. The “softening” process is significantbecause the channels through which itoperated help explain the unexpected inter-est of producers in higher prices andmoney, including traceable bank money.The archives provide no support for theproposition that a shortage economy wascreated to increase rent-seeking opportuni-ties. Still, despite extensive shortages,money remained a prized commodity and



producers pushed inexorably for higherprices, contrary to previous stereotypes ofthe command system. There were marketsor market-like institutions everywhere.These markets did not work well, wereseverely circumscribed, and relied heavilyon relational enforcement.

Part 5 deals, finally, with efficiency,growth, and economic reform. Before thearchives, we already knew that the Sovieteconomic system did not work that well. Inearly studies of enterprise managers, DavidGranick (1954) and Joseph S. Berliner(1957) uncovered massive principal–agentproblems. Investment choices were madewithout rational criteria (GregoryGrossman 1953). Prices failed to reflectscarcity values (Abram Bergson 1964). Theevidence of substandard postwar growthwas overwhelming (Bergson 1978; GurOfer 1987; William Easterly and StanleyFischer 1995) and increasingly preoccupiedthe country’s leadership; the system seemedincapable of significant reform (GertrudeE. Schroeder 1973, 1979, 1982; PhilipHanson 1983; Morris Bornstein 1985;Vladimir Kontorovich 1988).

Research in the archives has shown howdecisionmakers at every level actually allo-cated resources: They made it up as theywent along, using intuition, historical prece-dent, and common sense. As befits abureaucracy, there were plenty of formalrules, but the rules were constantly revisedor overridden. Those at higher levels madeformal rules, only to break them. Those atlower levels, unable to live without rules,relied on customary norms or rules ofthumb. The flaws in this planned economybecame apparent to its leaders almostimmediately. Before the archives, webelieved that official proposals for decen-tralizing economic reform began to circulateafter Stalin’s death; the first serious reformexperiment actually dates to 1932 and bearsa strong resemblance to reform proposalsafter Stalin’s death. But, like all those thatfollowed, it was quickly frustrated.

2. The Dictator, Hierarchy, and Economic Policy

Marx, Engels, and Lenin paid almost noattention to how socialism would organizeitself politically or economically. In their pio-neering discussion of socialism, Mises (1920reprinted 1935), F. A. Hayek (1937), Barone(1908/1935), and Oskar Lange (1964) effec-tively set this issue aside by assuming a tech-nocratic Central Planning Board inculcatedwith welfare-maximizing goals that dealtdirectly with enterprises. As a result, the“socialist controversy” (Bergson 1966)became focused on calculation and incen-tives. In terms of the real character of high-level decision making in the Stalinist state,archival documentation has given us com-pletely new knowledge. This regime wasindifferent to calculation, preoccupied bythe need to punish and deter its enemies,and bent on implementing its decisionsthrough a complex administrative hierarchyof agents motivated by threats and promises.

2.1 Stalin as Dictator

Early archival investigations, focused onStalin’s role in the Great Terror, showed thathe orchestrated it to a surprising degree(Khlevnyuk 1995), although he may not haveintended all its consequences (Getty andNaumov 1999; Rittersporn 2001). Stalin’scomplete authority is revealed in how he wasable to turn various major policies on and off,for example, stopping the Great Terror with asingle memorandum (reproduced in Kozlov2004, vol. 1). The same conclusion applies toStalin’s role in economic decision making.

From approximately 1932 until his death,Stalin was a true dictator: he had his way onevery matter and was not afraid to abuse andhumiliate those on whom he depended mostclosely (Davies 2001a; Davies, Melanie Ilic,and Khlevnyuk 2004; Rees 2004; Gorlizkiand Khlevnyuk 2004). As Khlevnyuk (2001a,p. 325: emphasis added) has concluded,“Stalin himself was not merely a symbol ofthe regime but the leading figure who made



the principal decisions and initiated all stateactions of any significance.” Stalin was per-sonally informed about and gave time tolarge numbers of minor matters. Stalin’s pre-eminent role does not mean that he was incontrol of the consequences of his decisions.On the contrary, his correspondence is full ofconcern about “paper fulfillment” and ofangry calls for “implementation” committeesand monitoring fulfillment by placingresponsibility on designated officials(Gregory 2004, pp. 165, 266).

What role did Stalin’s colleagues play? Inthe late 1920s and at the beginning of the1930s, decision making in the ten-personparty Politburo was still collective. Alreadyacknowledged as the senior personality,Stalin still had to bargain and cajole to gethis way (Khlevnyuk 1996; Rees and D. H.Watson 1997). Dissenting Politburo mem-bers were obliged to support majority deci-sions in public while bringing disputes to thefull Politburo for resolution, procedures thatStalin conveniently ignored during his powerstruggle with the “right deviationists”Bukharin, Tomskii, and Rykov. In the courseof this power struggle, Stalin also cultivatedthe support of regional leaders and othermembers of the larger party CentralCommittee, the venue for the ultimatedenouement of the power struggle in 1929and 1930. The archives confirm that Stalinexercised great cunning, patience, and self-control in organizing his Politburo majorityand Central Committee support to oust hislast rivals.

After 1930, Stalin increasingly bypassedthe formal procedures for party consultationand committee decisions; this is reflected inthe declining frequency of Politburo meet-ings, which fell from weekly in 1929 to onlyone in the second half of 1937 (Rees andWatson 1997). While formal meetings fellaway, Stalin met increasingly with his associ-ates in private where he could control par-ticipation and agendas directly (Wheatcroft2004). He reached decisions alone or with adhoc subcommittees that he personally

appointed and scheduled, an arrangementthat continued until his death in 1953 withlittle change. One element of formalityremained: Stalin continued to bind his asso-ciates into complicity by requiring eachPolitburo member to approve his decisionsonce he had made them (Gorlizki andKhlevnyuk 2004).

The erosion of collective rule is consistentwith Hayek’s (1944) insight that the rise of asole dictator is inevitable in such an environ-ment; this leaves unsolved the puzzle of itspartial restoration after Stalin’s death.Stalin’s ascendancy is explained by the needfor a tie breaker within the Politburo and thefact that Stalin was more ambitious, brutal,and controlled than his rivals. But why was atie breaker necessary? Although political sci-entists and historians had speculated aboutideological divisions within the Politburoafter 1930, the archives have revealed nodisagreements on the basic directions of for-eign and domestic policies. There was no“moderate” group after the “right deviation-ists” were eliminated; therefore, there wereno extremists either (Rees 2004, p. 47). Thedivisions that did exist within the Politburowere on lines of narrow self-interest basedon departmental position.

Stalin, who made all top-level appoint-ments personally, was deeply suspicious ofprofessional administrators and technocratsand trusted only a few old Bolsheviks. Heaimed to concentrate economic decisions inthe hands of a small number of “reliable”people; the ethos that Stalin wanted to instillis summed up in the term proposed byWheatcroft (2004): “Team Stalin.” Althoughhe consulted with them regularly as individ-uals, Stalin conceived of his colleagues col-lectively as his instrument, not aconsultative body. When, for example, hesaw political dominance over Gosplanthreatened, he ranted: “It is sometimesworse than that: Not Gosplan but Gosplansections and their specialists are in charge[and are turning the Politburo] into a courtof appeals or a council of elders” (Belova



and Gregory 2002, p. 270). He placedTeam Stalin members in positions forwhich he would hold them personallyresponsible and was dismayed when theybegan to represent narrower interests. Ashead of the party control commission,Ordzhonikidze, a fellow Georgian and for-mer head of the Caucasian party, led the battle against “disloyal” managers; onceappointed minister of heavy industry hebecame their ardent defender (Khlevnyuk1993). As head of the party’s transport over-sight committee, Stalin’s first deputyKaganovich (interviewed by F. I. Chuev1992, p. 61) stood up for encompassinginterests but began to demand more railsand investment immediately after hebecame transport minister.

Stalin railed against rent seeking within hisnarrow circle: “It is bad when we begin todeceive each other” (Khlevnyuk et al. 2001, p.80). He complained bitterly about the “self-ishness” of the minister of heavy industry,who pressed “on the state budget of the work-ing class, making the working class pay withits currency reserves for his own inadequacy”and that the “use of funds must be discussedin the interests of the state as a whole, notonly in the interests of [the ministry of heavyindustry]” (Khlevnyuk et al. 2001, pp. 72, 88).He particularly loathed the deputy ministerof heavy industry for “turning our Bolshevikparty into a conglomerate of branch groups”(Rees and Watson 1997, p. 16).

Examples of Politburo figures choosingencompassing interests over their own arerare. One case was the first secretary of theUkrainian party, who, unlike other regionalsecretaries, did not fight for lower grain col-lection targets with the result that Ukrainelost millions to starvation during the famineof 1932–33 (Davies and Wheatcroft 2004).The practice of top party officials represent-ing narrow interests illustrates a point madeby Khlevnyuk (2001a, p. 325): although thecompetition of interests “contradicted theprinciples of the dictatorship” it could alsolimit “the destructive consequences of

hyper-centralization.” In short, interestswere not less moral for being sectional.

2.2 Principals and Agents

Team Stalin was responsible for bringingdecisions forward and implementing them.The Politburo’s own immediate staff was lim-ited in the early 1930s to 230 specialists;below this layer were better-staffed ministriesand state committees that often prepareddecrees for the dictator. Team Stalin wasassisted by a state planning commission,Gosplan, that employed 900 specialists ofwhom many were technocrats of initiallyquestionable loyalty (Khlevnyuk et al. 1995).Team Stalin issued high-level decrees in thename of the government or its subcommit-tees for defense and the economy, but themost important state decrees were issuedjointly with the Central Committee.

The order of issuing decrees was confus-ing even to the top leadership, and Stalin hadto guard against improper invocation of theparty’s name by agencies seeking higherendorsements (Gregory 2004). Most decreeswere issued to a restricted number of recipi-ents on a need-to-know basis. Alongside thealmost 4,000 decrees published between1930 and 1941 were more than 28,000 secretones, of which over 5,000 were so secret thatthey were known only to a handful of people(Davies 2001a). Rules on secrecy were them-selves especially secret, which sometimescomplicated enforcement (Harrison 2004).

Standard organization charts (Gregory2004) show a vertical organization with thePolitburo and central government at the top,aided by functional and control agenciessuch as Gosplan, the committees for stateand party control, and the omnipresent inte-rior ministry known in different periods asthe OGPU, NKVD, or MVD. The actual“managers of production” were industrialministries and regional authorities thatplanned and supervised production units.The archives dispel the pretence that princi-pals and agents at all levels of the hierarchy



united to build socialism, share information,and forego private profit. Rather, the man-agers of production who were responsiblefor results engaged in disputes with those towhom they were accountable (Belova andGregory 2002; Gregory and AndreiMarkevich 2002; Gregory 2004) that wereusually resolved by informal means,although administrative and legal remedieswere also available (Belova 2005).

Formal links among industrial ministries,regional authorities, or among enterprises,are notably absent despite the fact that hori-zontal transactions are the basis of special-ization and exchange in all economies.Unauthorized horizontal links represented atroubling problem for the dictator. On onehand, a principal gives an order to an agentprecisely because the allocation that theprincipal desires differs from the one thatwould result from the agent’s trading on herown account. Hence, the obedient agentmust be induced to forego opportunities totrade for private gain. On the other hand,frequent planning mistakes ensured thatorders issued from above could not beimplemented unless supported by unautho-rized transactions, even the barter of “anordinary suitcase full of cigarettes” (cited byDavies 1996, p. 266). Such horizontal trans-actions (again to cite Khlevnyuk) “contra-dicted the principles of the dictatorship” butlimited “the destructive consequences ofhyper-centralization.”

2.3 Planning With a Light Touch

When key decisions were taken by TeamStalin, usually on Stalin’s personal authority,what role was left for the planning profes-sionals? Textbook accounts (Gregory andRobert C. Stuart 1974; Alec Nove 1977) sug-gest that Gosplan exercised considerableexecutive power over allocation. For theStalin period at least, this view requires revi-sion: Gosplan was important, but it was notpowerful, and surprisingly sought to limit itsown power.

8 Gosplan’s politicization had less of an impact on prac-tice than one might imagine because, in the early 1930s, itsmethods were still primitive. Gosplan’s most comprehen-sive centralized balances covered only thirty raw materials,eight energy sources, and four types of machinery allocat-ed among large-scale industry, small-scale industry, foodprocessing, and exports (Wheatcroft and Davies 1985,appendices A and D).

Gosplan was established in February 1921to institute and operate a “unified state plan”for the whole economy and harmonize theplans of other economic departments. It wasmodestly endowed for these grand tasks andstill had only fifty specialists in 1925. In1929, Stalin made it clear that he was notinterested in “balances” and “limits” thatrestricted his freedom of action. The purgeof Gosplan that followed famously tested“the civil courage of those specialists who arealready admitting in the corridors that theyprefer to stand up . . . for high rates of expan-sion than to sit [in jail] . . . for low ones” (aplan official in 1929, cited by Edward H.Carr and Davies 1969, p. 938). At the sametime a planned economy was being laiddown, those who were trying to create itsintellectual foundations were being“hound[ed] out of Moscow” in accordancewith Stalin’s instruction (cited by Belova andGregory, 2002, p. 271).8

Gosplan prepared only plans that werehighly aggregated, stating: “Gosplan is not asupply organization and cannot take respon-sibility either for centralized specification oforders by product type or by customer or theregional distribution of products”(cited byGregory 2004, p. 139). Gosplan refused toplan horizontal transactions in detail, label-ing the latter “syndicate work” (Belova andGregory 2002). Gosplan did represent thegovernment in interministerial conflicts andserved as a reluctant consultant despitepleading that “we are simply not equipped todeal with such matters” (Belova and Gregory2002, p. 271). In short, after its purge andsubsequent politicization, Gosplan limitedits exposure by doing as little as possible. Itwas not until the late 1940s that Gosplan



again became an activist organization underits young and influential chairman, NikolaiVoznesenskii.

The prevailing view of Gosplan before thearchives exaggerated its role, at least for theStalin period. Gosplan was not an all-power-ful director of resources; the powerbelonged to the dictator. Stalin did not wanta planning board with immense powers ornumerous staff. Why then did he keep it inbeing? Stalin had need of a relatively smallagency on which he could rely to tell him thetruth. He held most of those around himaccountable for results and for that reasondid not trust them to report outcomes truth-fully. That was the specialized task ofGosplan, which became Stalin’s solution tothe wider principal–agent problem (Belovaand Gregory 2002. Typically, Stalin also gaveoverlapping responsibilities to other agen-cies such as the economic administration ofthe NKVD. Gosplan’s leaders did not have tomake a success of the economy in order toretain Stalin’s trust; that was the job of theministers for the production branches. AllGosplan had to do was to report honestly.This did not necessarily make an independ-ent Gosplan leader such as Voznesenskiipopular with the other top party managers.Voznesenskii was one of Stalin’s favoriteswhile Stalin trusted him. When the othersfound an opportunity to sow distrust withthe boss, Stalin had him shot (Gorlizki andKhlevnyuk 2004).

2.4 Rules Versus Discretion

Hierarchical organizations resort to con-tingent rules, customs, and conventions tomake boundedly rational decisions. Bergson(1966) suggested that contingent rules couldresolve a number of problems, such asopportunism and computation, specific to aplanned economy. Hayek (1944, p. 82)asserted to the contrary that a totalitariansystem “cannot tie itself down in advance togeneral and formal rules that prevent arbi-trariness . . . It must constantly decide ques-tions which cannot be answered by formal

9 Contingent rules that guide decision making featurein two different economic literatures, both of which arerelevant to the present survey. The rules-versus-discretionliterature refers to formal stable rules, for example, thosethat govern taxation. In this context, it is a problem if theagent can too easily ignore or renegotiate such a rule. Thiswas Hayek’s point. The bounded-rationality literature dis-cusses customary rules or rules-of-thumb that are madeinformally and evolve as agents struggle with computationconstraints and uncertainty. In this context, it is a problemif such rules lead to systematic errors. Our point is that theSoviet command system was intolerant of formal rules; onthe other hand, informal rules-of-thumb proliferated.

principles only.” The archives show thatHayek was right.9

While the Soviet economy was managedby decree, there were few formal stablerules; the rules that existed were subject tooverride. Fresh guidelines were issued toplan each new year or quarter, rather thancarry over general planning rules. Ministriesoperated without charters that spelled outcorporate governance (Gregory andMarkevich 2002). The few accounting andloan administration rules were easily ignoredwith the tacit approval of Team Stalin(Gregory 2004); the enforcement of financialtargets and value-for-money was selectiveand arbitrary. Rather than let the courtsenforce legal interagency agreements, TeamStalin allowed and encouraged “administra-tive” enforcement through appeals to verticalsuperiors (Belova 2005).

Hayek’s expectation of the dictator’s aver-sion to formal rules is exemplified by theoperational plans that enterprises wereobliged to fulfill by law (Belova 2001). Theformal procedures were complicated, con-tradictory, and confusing for the actual par-ticipants (Markevich 2003). In the Stalin era,the operational plan for the enterprise wasnot the unified output, input, finance, tech-nology, and labor plan [tekhpromfinplan]that Soviet and western planning textsdescribe. Rather, the enterprise usuallyreceived a few output and assortment assign-ments midway through the plan period,while secondary targets for costs and pro-ductivity were worked out retrospectively forreporting purposes.



A surprising feature of the workingarchives of ministries and enterprises is thenear total absence of final “approved” plans.All plans were labeled “draft” or “prelimi-nary.” The draft plan was no more than aninformal agreement which could be changedsubsequently by virtually any superior. The“correcting” and “finalizing” of plans was anever-ending process; the “final” planremained always on the horizon. Searches inthe ministerial archives have located onlyone finalized annual plan, that for lightindustry in 1939 (Markevich 2003).

The archives provide thousands of cases ofplan revisions. Ministers ordered last minutechanges; factories were shuffled from oneauthority to another; one factory wasordered to increase its production post hasteto make up for production shortfalls inanother factory. Even Politburo commissiondecisions could be changed at the lastminute: Although the first-quarter 1933vehicle distribution plan had been approvedby its own transport committee, the Politburothrew the plan out the window by tripling theKazak party committee’s allocation and dedi-cating 90 percent of the vehicles to “organs ofcontrol over agricultural producers” (ValeryLazarev and Gregory 2002, p. 332).

Each level received a barrage of requeststo intervene or to refrain from interventionin a highly formalized bureaucratic process.In deciding how to respond, Team Stalin diduse some implicit rules-of-thumb, such asthe priority of heavy industry and the mili-tary. Thus the minister of heavy industry:“All orders for the Ministry of Defense mustbe fulfilled exactly according to the schedulenot allowing any delays” (cited by Gregory2004, pp. 160–61). Enterprises cited priorityconsiderations to defend against interven-tions. Military shipbuilders in 1935 ignoredorders from the highest state authority onthe grounds that they interfered with mili-tary objectives (Gregory 2004). Priorities,however, had to be limited to be effective,and the priority statements issued by leadingauthorities were often either too broad or

too specific (Gregory 2004 gives examples).Delegating the right to intervene to lowerlevels was dangerous, however, since it was aproblem to align the interests of loweragents with those of the dictatorship asBergson (1964) once pointed out.

Plan interventions created havoc for pro-ducers. A meeting of industrial ministersheld in December 1946 turned into a repre-sentative complaint session. The minister ofthe electronic industry: “During the quarter,even during the month, there are a greatnumber of changes, modifications, and addi-tions to the approved plan. We do not usual-ly work according to the plan; rather we workon the basis of supplemental decrees, admin-istrative decisions, and the like . . .” The min-ister of the aviation industry: “It is better tohave one plan than to change it twenty times”(Markevich 2005). The most important indus-trial leader of the 1930s expressed his frustra-tion as follows: “They give us every daydecree upon decree, each one is stronger andwithout foundation” (Khlevnyuk 1993, p. 32).

Ministries and enterprises insured them-selves against interventions by holding backobligatory information and submitting theirown plan proposals at the last minute toavoid duplicating this work later; ministriesoften proposed relatively modest targets toGosplan while quietly imposing tougherassignments on their own enterprises(Belova and Gregory 2002). Ministriesfought for generalized plans and tried toavoid divulging enterprise plans to Gosplan;in April 1933, for example, Gosplan com-plained that ministry plans “suffered fromsuch incompleteness that it is impossible touse them” (Belova and Gregory 2002, p.274). The ministries withheld informationfrom Gosplan and financial authorities ongrounds of national security (Belova andGregory 2002; Harrison 2004). For produc-ers, the best plan was either no plan at all ora plan so general that it left all the real deci-sions to them. Gosplan even uncoveredcases of “nonplanning”: “Enterprises [largeenterprises located near Moscow] declared



to our representatives that they had not seenannual plans for a period of years” (cited byBelova and Gregory 2002, p. 275).

Stalin (1937, p. 413) wrote: “Only bureau-crats can think that planning work ends withthe creation of the plan. The creation of theplan is only the beginning. The real directionof the plan develops only after the puttingtogether of the plan.” Team Stalin main-tained the right to override plans at willbecause this provided “resource mobility.”In practice, such procedures served theinterests of principals at every level andcould not have been better designed for theexercise of political influence. Everythingwas tentative and subject to arbitrary changeby someone higher up in the chain of com-mand. No one can know what went onbehind the scenes when petitioners met withsuperiors, but we can guess that savvy politi-cians like Stalin would weigh the politicalbenefits of satisfying an influential regionalor industrial leader.

The aversion to explicit rules reflected thedynamic commitment problem of a dictator.Stalin’s unwillingness to bind himself inadvance cascaded down through the politicalsystem, preventing the emergence of a for-mally rule-based or “law-governed” econo-my. Kornai, Eric S. Maskin, and GerardRoland (2003) suggest two kinds of commit-ment failure: in one, a predatory principalforces agents to break formal rules so as toexploit them; in the other, agents choose tobreak the same rules so as to exploit a weakprincipal. In the Soviet case, producerscould break rules citing the threat to pro-duction from the rule, while superiorsreserved the right to punish hapless scape-goats for breaking the same rules. It is some-times hard to detect whether such rulebreaking reflected the power of officials toforce producers to commit violations or theability of producers to commit violations andget away with them.

Generally, to exploit power to the fullinvolves encountering its limits. Specifically,it was the power to live outside formal rules

that sentenced Stalin and his Politburo tolives of toil, drudgery, and tedium (Gregory2004). Threats of resignation and pleas forlengthy vacations were commonplace. A rep-resentative Politburo meeting, held on March5, 1932, had 69 participants and 171 points onits agenda (Khlevnyuk et al. 1995). The great-est burden fell on Stalin who, in a typicalyear, 1934, spent 1,700 hours in official meet-ings, the equivalent of more than 200 eight-hour days (Khlevnyuk 1996). Virtually everycommunication requested his decision.

On rare occasions, Stalin would explode atthis torrent of paperwork, for example in atirade of September 13, 1933: “I won’t readdrafts on educational establishments. Thepaperwork you are throwing at me is pilingup to my chest. Decide yourself and decidesoon!” (Khlevnyuk 1996, p. 340). A fewweeks later the same Stalin berated thePolitburo for not following his proposed dis-tribution of tractors to the letter (Khlevnyuk1996). Stalin suffered the dictator’s curse(Gregory 2004): his power to decide all gavehis most trusted colleagues the incentive todecide as little as possible. The less theydecided, the less he could blame them whenthings went wrong.

3. Accumulation and Consumption

Before the archives, we could only guess atStalin’s real economic policy. Would the dic-tator foster economic growth (Mancur Olson1993; Edward Glaeser et al. 2004)? To whatextent would he share his rents to build loyal-ty or increase his political power (RonaldWintrobe 1998)? The archives tell us thatStalin was obsessed with accumulation, whichis hardly a surprise. Between 1928 and 1937,Soviet real GNP doubled, but the fiercerepression of private consumption enabled aquadrupling of real investment (Bergson1961). More generally, investment rates insocialist countries were consistently higherthan in capitalist countries of comparable sizeand income levels (Simon Kuznets 1963).

Without information on high-level deci-sion making, the prearchival literature had



10 Bergson’s authoritative Economics of Soviet Planning(1964) introduces investment first in chapter 11 and large-ly puts aside the issue of its total volume, focusing insteadon rational choice of individual projects. The most widelycited paper on investment (Grossman 1953) deals exclu-sively with the choice to be made when a given increase indesired output could be obtained from alternative projects.

little to say about aggregate investmentchoice.10 Bergson argued (1964, p. 241) thatinvestment depends “on the output mix tobe produced from it”; this view percolatedthrough the various editions of textbooks(Gregory and Stuart 1974). Hence, the pre-vailing prearchival view was that the dicta-tor’s primary target was output, from whichinvestment was derived. The archives showthat this view was wrong: Stalin and thePolitburo attempted to control the basicdirection of the economy through the leveland distribution of investment. They wereless concerned about setting physical out-puts in operational plans.

The literature on how a socialist economyought to determine aggregate investmentstems from Marx’s emphasis on accumula-tion and proceeds via the growth models ofPreobrazhenskii and Feldman in the earlySoviet period (Alexander Erlich 1960;Nicolas Spulber 1964) to the later models ofBrus, Kalecki, and Dobb (Peter Rutland1985). The earlier models suggest that thefeasible investment rate depends on the bal-ance of power in the economy and the scopefor expropriating the “former” social classes.The later models suggest that the optimalinvestment rate depends on society’s dis-count rate, but that a labor-abundant marketeconomy may underinvest if the realizedsubsistence wage is above the shadow wage.No Politburo member was trained in eco-nomics, however; nor did they feel the lackof such training.

3.1 The Politburo Accumulation Model

At the core of the Politburo’s strategy to“build socialism” in the first two five-yearplans, were massive programs for the hydro-electric dams, machinery complexes, vehicle

works, blast furnaces, railways, and canalsthat were included on its itemized “title lists”of approved projects. Their purpose was toembody the new society in cement andstructural steel. How were these programsimplemented and managed?

Although Politburo meetings for the1930s left few formal minutes, his deputieswrote to Stalin from time to time to detailkey Politburo meetings and seek approval ofoperational decrees (Gregory 2004). Theyreveal that the Politburo consistently setthree targets in the 1930s: the nominalinvestment budget, grain collections, andforeign exchange. These three control vari-ables all related to investment. The invest-ment budget allotted funds to industrial andregional agencies for construction andmachinery. Despite an original intent to usephysical material balances of investmentgoods to plan investment, ministries andregional authorities were simply given“investment rubles,” and no one appeared toknow the real investment that resulted.Grain collections were designed to con-tribute to a budget surplus through theexcess of state sale prices over purchaseprices. Stalin personally directed foreignexchange to the import of capital goodsrather than the luxury goods sometimesdemanded by the Bolshevik elite.

If Stalin’s goal was indeed to maximizeinvestment, the archives provide two typesof evidence that are, at first glance, confus-ing. First, Stalin was extremely concernedabout consumption, particularly where ittouched upon the productivity and moraleof the industrial workers. Consumer sup-plies were one of the most frequent itemson Politburo agendas; in Stalin’s words, the“provisioning of workers” was one of “themost contested issues” before the Politburoand trade was “the most complicated min-istry” (quoted by Gregory 2004, pp. 93–94).Stalin interpreted declining labor productiv-ity as a sign that workers “were not provi-sioned as well as last year,” and personallyordered the delivery of consumer goods to



Figure 1. The Politburo Model

E(W,C)

Q(E) Total Output, Q

Q∗

W∗

Total Effort, E E∗

cities where labor productivity was declin-ing (Gregory 2004). The Politburo set itselfup as the highest trading organization,deciding retail trade plans, prices, assort-ment, and even the opening of new stores.Second, we have the little-known empiricalfact that the Politburo, which based thesuccess of its program on capital construc-tion, deliberately reduced nominal invest-ment on two occasions, in 1933 and 1937(Davies 2001b) citing the fear of inflationand rising construction costs.

This evidence can be interpreted in twoways. One is that that Stalin’s preferenceswere unstable; fitful humanitarian impulsesled him to direct more supplies to theworkers from time to time; periodic fitsof financial orthodoxy led him to regretoverstraining the economy. Alternatively,Stalin’s preferences were stable and hereallocated consumption or reduced invest-ment according to a consistent rule-of-thumb. Knowing of Stalin’s capacity forcalculation, patience, and self-control, wereject the first explanation and investigatefurther the second.

Figure 1 illustrates the model that thePolitburo appears to have used to setinvestment and consumption. The figuremirrors the Marxian concept of the surplusproduct, the gap between output and con-sumption, as the outcome of a distributivestruggle. The model has theoretical precur-sors in Wolfram Schrettl (1982, 1984) andLeon Podkaminer (1989), and is set outmore fully by Gregory (2004). It belongs toa general class of models in which a ruler’sfreedom of action is circumscribed by social“tolerance limits” (Kornai 1980) or a revo-lution or disorder constraint (DaronAcemoglu and James A. Robinson 2000).By raising investment without limit, thePolitburo risked provoking the workers togo slow, strike, or rebel.

The demand for labor was always enoughfor full employment and besides all able-bodied persons were required to work bylaw (Granick 1987), so the figure takesemployment, N, as fixed exogenously; indi-vidual effort, e, was variable, so total effort,E = e·

—N, was variable although employment

was not. Total output, Q, depended on total



11 The fair-wage concept in figure 1 can be related tothe observation of George A. Akerlof (1984) and Akerlofand Janet L. Yellen (1990) that workers in marketeconomies give extra effort if they are paid a wage per-ceived to be fair, and withdraw it if the wage is reducedbelow the psychologically determined fair level. We sup-pose that effort has a maximum, so payments above the fairlevel do not elicit more effort.

effort, E. Total effort varied with the realwage, w, as follows. The aggregate wage billW, the consumer goods received by workers,is measured along the vertical axis in thesame units as output, and is proportional tothe real wage given that employment isfixed, i.e., W = w ·

—N. There is a reservation

wage, analogous to a tolerance limit or disor-der constraint, below which effort is zero;there is also a “fair” wage at which effort ismaximized.11 As the economy moves fromthe fair wage to the reservation wage, effortdeclines as workers turn up late or drunk, orgo slow or absent without leave; at the limit,unrest simmers and threatens to boil overinto local and general strikes and rebellion.Thus the effort curve intersects the horizon-tal axis at the reservation wage and becomesvertical at the fair wage. Effort also dependson the level of direct coercion, C, to whichwe turn in the next section. To maximizeeffort, the dictator would pay the fair wageand get the maximum output, but this wouldnot maximize the surplus. To maximize thesurplus, Q − W, he would choose the inter-mediate wage, effort, and output levelsdenoted W∗, E∗, and Q∗.

The shape of the effort curve is hypothet-ical; the hypothesis is Stalin’s and can beinferred from his observable anxiety abouterrors in the distribution of consumergoods. An effort curve of this shape makesthe consequences of plan mistakes asym-metric: for the dictator, paying the workerstoo little could be much worse than payingthem too much. Starting from the invest-ment optimizing position, more consump-tion does at least raise effort and output,and this somewhat mitigates the fall ininvestment. A mistake of the same size from

the same starting point and in the oppositedirection, however, not only cuts effort andoutput by more but also risks pushing theworkers into outright confrontation with thestate. An investment-maximizing dictatormust tread a fine line between the pursuitof investment and the triggering of seriousdisorder.

Stalin managed worker morale and effort intwo ways. When investment and consumptionwere about right in the aggregate, detailedplan mistakes could still leave some workerswith too much and others with too little.When this happened, Stalin personallyordered the reallocation of consumer goodsto those left short. But there could also beaggregate mistakes; when too much invest-ment threatened to disrupt the economyand provoke the workers, the Politburopreferred to sacrifice investment. This ishow we interpret the unforced investmentcutbacks of 1933 and 1937. Although Stalindid not express his concern for consump-tion directly, when he advocated “strength-ening the ruble” as a justification for lessinvestment (Gregory 2004, pp. 236–42), itwas the same ruble that workers had tospend in retail markets; when he feared ris-ing construction costs, these were fiscalcosts that had to be covered through taxa-tion or the inflation tax. In this sense,Stalin’s behavior was stable and consistent,given the constraints that he perceived.

Figure 1 suggests other options. TeamStalin could seek to manipulate the effortcurve by offering ideological rewards inplace of material payoffs. Idealists fromRussia and abroad in fact assisted the firstfive year plan, motivated by the idea ofbuilding socialism. The attempt to trans-form homo economicus into homo sovieticusled, however, to a vicious circle of wageequalization and declining productivity(Hiroaki Kuromiya 1988; Davies 1989a,1996). Subsequent mobilizations were limit-ed to short lived campaigns such as forWorld War II, and to cultivate the “virginlands” of Kazakhstan and Siberia after the



war. The most publicized mobilizationcampaign of the 1930s, the Stakhanovitemovement, was driven by progressive piecerates that permitted participating workersto drive up their incomes by overfulfillingnorms (Davies and Khlevnyuk 2002); thearchives suggest that Stalin abandoned itbecause it tended to raise fair-wage aspira-tions among nonparticipating workers, andalso threatened inflation (Gregory 2004).Stalin could also elicit extra effort fromparty “enthusiasts” in the present by prom-ising future promotion. The problem wasthat these promises eventually had to bemade good. Lazarev (2005) has identifiedthe creation of new posts that carried awage premium to reward loyal supportersas an important feature of Stalin’s policy.Relentlessly purging the potentially disloy-al made it cheaper for him to keep hispromises to the actually loyal.

In an extension of the model, thePolitburo could bear down on consump-tion selectively via targeted rationing.Elena Osokina (2001) argues that Stalinwanted to drive down consumption andhence introduced a discriminatory ration.Julie Hessler (2004) argues that, on thecontrary, Stalin approved of nonrationedtrade and a consumption oriented life-style but crises (collectivization, rearma-ment, war, and postwar famine) allowedhim to achieve this goal for only short peri-ods. Gregory argues that Stalin sawrationing as a way to force accumulationwithout a loss of effort of high-priorityworkers according to the principle: “Hewho does not work on industrializationshall not eat” (cited by Gregory 2004, p.98: emphasis added). Rationing carriedother costs, however, as Stalin becameaware (Davies and Khlevnyuk 1999).

Figure 1 has a further ominous extension.The fair wage was set by a mass psychologythat was unpredictable and hard to manipu-late. If workers concluded from the propa-ganda of economic successes that they werebeing cheated, the fair wage would rise,

forcing the Politburo to cut investmentback. Stalin used the vast informant networkof the NKVD to monitor protests, strikes,anti-Soviet statements, and factory-wallgraffiti, and eavesdrop to gauge mass opin-ion (Berelovich 2000). Stalin had obviouspolitical motives to do this, but within ourframework we conclude that wages and fair-ness lay at the cross-hairs of politics andeconomics.

3.2 Coercion: Three Experiments

In principle, efficient penalties are cheap-er than efficient rewards since “a promise iscostly when it succeeds and a threat is costlywhen it fails” (Thomas C. Schelling 1960, p.177). In figure 1, the fear of punishment maycause workers to lower their reservationwage without reducing effort; coercion maynot make effort more productive but shouldmake people willing to supply effort for less.As long as coercion displaces the effort curvedownward while leaving the productioncurve undisturbed, the surplus is increased.In market economies, outside options, suchas alternate employment, leisure, or unem-ployment, cannot easily be limited by forceas the ubiquity of informal sector employ-ment and illegal migration demonstrates. Inthe Stalinist state, however, the idea of con-trolling workers’ alternatives by force provedattractive.

The archives show that Stalin believedthat a wide range of problems could besolved by force. The language with which headdressed obstacles to his rule was habitual-ly violent. Faced with foot-dragging in thetransport ministry in 1931, he wrote: “wemust smash this gang . . . If you can managewithout my help, smash the gang before it’stoo late.” Considering the role of speculatorsin a legalized urban rural market, he wrotein 1932: “we must eradicate this scum . . .The OGPU [secret police] and its agenciesmust, without delay, start training its forcesand studying the enemy” (Davies et al. 2003,pp. 95, 102, 165).



The archives have thrown new light onpast controversies. Historians have arguedabout the extent to which Stalin’s motives inthe Great Terror of 1937–38 were primarilyeconomic or political; thus Roberta T.Manning (1993) speculated that Stalinlaunched the terror in order to solve wide-spread economic difficulties, but thearchives have not yielded any substantial evi-dence to this effect (Davies 2004). Rather,the aim of the terror was political: to isolateand remove a potential “fifth column” fromsociety that, Stalin believed, could endangerhis regime in a time of rising internationaltension (Khlevnyuk 1995). Of more interestto economists, some scholars have arguedthat coercion was the glue that held theStalinist economy together. When coercionfailed, or was abandoned, the economy fellapart (Kontorovich 1986; Olivier Blanchardand Michael Kremer 1997; Harrison 2002).Others have emphasized the huge social andeconomic costs of coercion (RobertConquest 1987; Khlevnyuk 2001b; Daviesand Wheatcroft 2004). Below we discuss theissues of whether coercion paid for itself andwhere the optimum lay.

Stalin conducted three notable experi-ments with the use of coercion to fosteraccumulation: the forced collectivization ofthe peasantry, the criminalization of work-place indiscipline, and the widespread useof forced labor. In each case, the outlineshave been known for decades, but the fullconsequences are only now coming to light.

3.2.1 Collectivizing Peasants

Politically, collectivization aimed toimpose Soviet power in the countryside andeliminate the stratum of richer peasants,the kulaks. It was triggered, however, by agrain marketing crisis that reflected thepeasants’ perceived unwillingness to con-tribute sufficiently to investment-led indus-trialization. Collectivization began inearnest in December 1929, signalingStalin’s victory in the power struggle. The

relatively open discussion in the late 1920ssupported a substantial literature on the col-lectivization decision, the most completeaccount of which is by Davies (1980a,1980b); James W. Heinzen (2004) has nowadded a study of the agriculture ministry inthe 1920s. Naum Jasny (1949), Moshe Lewin(1968), and Erich Strauss (1969) outlinedsome general results. James R. Millar (1974)and Michael Ellman (1975) concluded thatthe investment surplus that Stalin hoped togain from collective agriculture was probablynot forthcoming. But the detailed conse-quences of collectivization and the massdeportation or detention of peasants whowere excluded or resisted were concealedbehind a thick veil of secrecy.

Mark B. Tauger (1991), Davies, Tauger,and Wheatcroft (1995), Wheatcroft andDavies (2002), and Davies and Wheatcroft(2004) have now reviewed the archives onthe immediate aftermath of collectivization,including the famine of 1932–33. The collec-tive farms enabled Moscow to replace localdecision making with its own detailed plans,instructions, and formal, but often transientrules. Stalin was focused on what he couldcontrol: sown acreage and the state’s share inwhat this acreage produced. But Stalin couldnot control the harvest. Acreage expandedbut yields collapsed; the share delivered tothe state increased. Excessive procurements,bad weather, and plan errors combined tostrip the countryside of grain; first the live-stock were slaughtered, then the farmersthemselves starved. They were preventedfrom feeding themselves from their ownharvests by severe punishments includingdeath for petty theft. Davies and Wheatcroftdispel Conquest’s (1987) notion that Stalinmanufactured the famine to kill class ene-mies; rather they show the leadership subse-quently trying to ameliorate the effects of itsown bungling.

While famines usually occur in poor coun-tries with limited statistical reporting, theSoviet archives provide good documentationof the two peacetime famines of 1932–33 and



1946–47. Ellman (2000) has applied the enti-tlement theory of Amartya K. Sen (1981) tothe latter famine, which particularly affectedthe peasants of the Ukraine and other south-ern regions. Sen argued from historical expe-rience that famines are more likely to arisefrom entitlement or distributional failuresthan production failures, and that “no matterhow a famine is caused, methods of breakingit call for a large supply of food in the publicdistribution system” (1981, p. 79). Ellmanmakes two points from the evidence.

First, he argues, in the 1946–47 famine,production and entitlement failures interact-ed. The production failure was not such as toleave insufficient food available to keep every-one alive. In this sense, the famine arose froma failure of entitlement. But the famine wastriggered by a harvest failure; if the harvesthad not failed, everyone would have had suf-ficient entitlements to stay alive. Therefore,the famine cannot be attributed to entitle-ment failures alone. Second, Ellman notesthat in this famine the role of the state wasessentially negative: it selected those who diedby denying them entitlements. Therefore, itcould be argued, concentrating grain stocks inthe hands of the Soviet state actuallyincreased the number of deaths. Davies andWheatcroft (2004) show that, in 1932 and1933, the state intentionally directed food tothose able to work in the fields and denied itto those already hospitalized by hunger.Accordingly, the 5.5 to 6.5 million faminedeaths in these years far exceeded thoserecorded in famines before the Revolution(Davies and Wheatcroft 2004).

3.2.2 Regimenting State Employees

As the 1940s began, Stalin redirectedcoercion from specific class enemies to theentire public-sector work force. A batteryof intimidating laws criminalized workplace violations which had previously beenmanaged by administrative sanctions with-in the enterprise. The laws themselveswere not secret and were described byConquest (1967) and Nove (1969). Their

consequences, however, remained obscureuntil the archives were opened.

The law of 26 June 1940 (Kozlov 2004,vol. 1) made absenteeism, defined as anytwenty minutes’ unauthorized absence oreven idling on the job, a criminal offense,punishable by up to six months’ correctivelabor at work with a 25 percent reduction inpay. Repeat offenses counted as unautho-rized quitting, punishable by two to fourmonths’ imprisonment. Enterprises man-agers were made criminally liable for failureto report worker violations. In August 1940,the minimum sentence for petty theft atwork and “hooliganism” was set at one year’simprisonment. Wartime decrees punisheddefense and transport workers for unautho-rized quitting with long terms in a laborcamp. After the war, a notorious decree ofJune 1947 set the minimum sentence for anytheft of state or socialized property at fiveto seven years imprisonment. Transportwork was eventually demilitarized inMarch 1948 and work in the defense indus-try two months later; otherwise, thesepunitive laws remained on the books untilStalin died.

More liberal governments also took pow-ers to direct their key workers in wartime,but the detail, scope, and degree of enforce-ment of the Soviet measures went to anextreme. A report prepared as backgroundfor Khushchev’s secret de-Stalinizationspeech of February 1956 (Kozlov 2004, vol.1, statistical appendix), shows that, from 1940through June 1955, the regular courts andmilitary, transport, and labor camp tribunalssentenced a total of 35.8 million persons forall criminal offenses. Not allowing for repeatoffenders, this would represent about onethird of the adult population of roughly 100million. Of the 35.8 million, 15.1 millionwere imprisoned and a quarter of a millionwere executed. The annual rate of imprison-ment was one million or more in most yearsup to 1950, and more than half a million atthe time of Stalin’s death. Such convictionrates were about five times as large as in the



12 Given the many factors that differentiated the Sovietjudicial and penal system from others, it is hard to findappropriate standards for detailed comparison of work-place offending. It is simpler to go to the aggregate of alloffenses. From 1940 to June 1955, the USSR imprisoned15.1 million people or, on an annual basis, roughly one per-cent of its adult population each year. In 1958, for com-parison, the United States imprisoned not more than onefifth of one percent of its adult population (U.S.Department of Commerce 1960, pp. 139, 141). ThusStalin imprisoned his subjects at least five times more fre-quently than the United States.

United States at about the same time.12

These totals were dominated by workplaceoffenses (Don Filtzer 2002; Andrei K.Sokolov 2003, forthcoming) and by convic-tions for theft of socialized property (Kozlov2004, vol. 1, statistical appendix).

The archives show that Team Stalin, notan independent judiciary, determined thenumber of convictions. When the laws wereenacted, Stalin had to intervene personallyto force a reluctant judiciary to prosecutetardy or lazy workers (Sokolov 2003).

3.2.3 The Labor CampsCollectivization in the early 1930s, the

Great Terror of 1937–38, the repression ofstate employees in the 1940s, and the arrestsof “national contingents” during and imme-diately after World War II created hugeflows into labor camps. In the 1920s, therewas just one forced labor complex in theArctic where mainly political detainees wereheld. The first major expansion came withcollectivization which threw hundreds ofthousands of peasants into camps or specialsettlements; the Gulag, the interior ministry’schief administration of labor camps, was cre-ated in 1930 to handle the sudden inflow andlabor camps spread across the remote interi-or and the far north and east to house, attheir peak, more than 2.5 million inmates.

Although the Gulag was shrouded inintense secrecy, its human side leaked outthrough the recollections of former inmates.Estimates of its scale and scope, however,could be based on little more than guessworkuntil the first official figures were released in1989. While contemporary estimates of the

number of detainees ranged up to 20 million,we now know that there were approximately2.3 million penal laborers at the outbreak ofWorld War II and about 2.5 million at thetime of Stalin’s death; similar numbers ofdeportees were also confined to labor settle-ments in the remote interior. The forcedlaborers were mostly engaged in forestry,mining, and construction, where they madeup substantial shares of employment, butnever more than about 3 percent of the totalworkforce including farm workers, and lessthan this in terms of the value of nationaloutput (Khlevnyuk 2001b, 2003a).

These new facts must be set beside others,including the very high rates of convictionand sentencing that we now know about:although the Gulag population was smallerthan observers had earlier guessed, it alsohad much higher turnover with very largenumbers entering and leaving to return tosociety. While we are confident of the stocksof Gulag inmates at different points in time,the cumulative total of persons sentenced tothe Gulag in the course of its existence,probably in excess of twenty million, remainsthe subject of debate. We now know that theGulag’s own central catalogs are inconsistent(Kozlov 2004, vol. 2); it appears that eventhe Gulag did not know the correct number.

Internal Gulag documents confirm thatpolitical strategy (collectivization, terror,war) rather than economics dictated theGulag’s development. Before the archives,there was speculation that in the course ofthe 1930s economics eventually took overfrom politics as a motivation for recruitingforced laborers, or alternatively that theNKVD became a lobby for forced labor (S.Swianiewicz 1965). Neither of these hasturned out to be the case, although ourjudgment has to be carefully shaded.

The Gulag had a consistent economic rai-son d’être: to explore and colonize regionsthat were resource-rich but inhospitable,since forced labor could be ordered aroundthe country at will (Khlevnyuk 2003a).Subsistence wages combined with the



enforcement of effort through close supervi-sion were supposed to promote low-costaccumulation.

At the level of the state, the continued pri-macy of politics over economics is shown bythe fact that the NKVD–MVD did not lobbyfor expansion. The NKVD projected ashrinking number of Gulag inmates for thethird five-year plan (1938–42), just as the firstvictims of Stalin’s Great Terror began to floodin (Gregory 2003a). After the war, the NKVDagain planned for contraction and politicsagain overturned this expectation(Khlevnyuk 2003a). In the late 1940s, Gulagofficials proposed to release all but the mostdangerous prisoners from camps (AlekseiTikhonov 2003), but this was unacceptable toStalin. In 1953, within three months ofStalin’s death in March, MVD chief LavrentyBeriia had released one and a half millionprisoners, 60 percent of the Gulag’s inmatesaccording to a plan prepared five years earli-er. In fact, elements within the MVD wereincreasingly alarmed by the Gulag’s econom-ic and social costs. The economic costs werereflected in its growing financial deficits; thesocial costs were measured by high rates ofrecidivism. Although the camps were sup-posed to segregate hardened criminals fromyouth offenders, the camp population was amixing bowl and recidivism soared. The highturnover spread the culture and mores ofcamp life throughout society.

Why did the Gulag fail? Research on themicroeconomics of the Gulag is in its infan-cy but early case studies show a complexlearning process. The Far Eastern camps(David Nordlander 2003) show how earlyoptimism about huge surpluses in gold min-ing was replaced by pessimism as output perinmate fell precipitously. The Kareliancamps (Christopher Joyce 2003) show theexperimental process by which the authori-ties learnt the scope and limits of theexploitation of forced labor. The fact thatthe White Sea–Baltic Canal (MikhailMorukov 2003) was finished on time and onbudget stimulated illusory expectations for

the Gulag’s future until its major construc-tion flaws became apparent. The Gulagleaders willingly undertook the building ofNoril’sk (Leonid Borodkin and Simon Ertz2003) because they underestimated the risksand difficulties that would arise. The opera-tion of Noril’sk (Ertz 2003) helped to exposeillusions about the ease with which theinmates could be coerced into supplyingeffort without economic rewards.

By the postwar years, Gulag officials hadconcluded that the camps failed to generatea surplus. Labor productivity there wasextremely low relative to that of free work-ers, while guarding detainees was veryexpensive; in 1950 there was one guard toten inmates, leading to the widespread prac-tice of “unguarded” prison contingents.Within the Gulag, prisoners formed protec-tive networks among themselves and withthe guards to cover for each other (Heinzen2005). The arsenal of punishments was notsufficient to motivate prisoners and trade-offs were complicated: prisoners placed onreduced rations for failing to meet work quo-tas were no longer able to work effectively.One of the most effective incentive systems,early release for exemplary work, deprivedthe Gulag of its best workers. Material incen-tives played an ever larger role in motivatingpenal labor (Borodkin and Ertz 2003, 2005;Ertz 2005). In the last years of the Gulag,there was a process of “conversion of slavesto serfs” (Khlevnyuk 2003a, p. 57); thecamps increasingly paid prisoners civilianwages and the distinctions between penaland free labor became blurred.

3.3 Coercion Failure?

The collective farms effectively ceased toexist in the mid-1960s when the farmerswere placed on fixed wages like any otheremployee. The Gulag was emptied of massprisoners between 1953 and 1957. The dra-conian labor laws of the 1940s were rescindedin the mid-1950s. These relaxations have twopossible interpretations: either the post-Stalinleadership did not have the stomach for a



system of repression that was working orcoercion was abandoned because it did notwork. The new evidence supports the sec-ond interpretation, particularly the fact thatthe most coercive instrument, the Gulag,was dissolved by its own leaders.

In short, the initial belief that results couldbe obtained more cheaply by raising thepenalty for failure than by increasing thereward for success appears to have been mis-taken. Effective coercion requires that penal-ties be accurately assessed and targeted, andthat the agents of repression are wellinformed about offenders and the costs oftheir crimes. It proved more difficult thanexpected to target penalties accurately; as aresult the relationship between true effortand punishment was “noisy.” The devising ofan efficient coercion system is then furthercomplicated if workers and managersrespond strategically to increase the noise.

The archives show that officials had littleidea whether workers were exerting fulleffort or not; the law could do little morethan ensure that they were physically at workand did not steal too much. Agricultural con-trollers could order the collective farms tosow more land but could not assess whetherthe land was being farmed efficiently(Davies and Wheatcroft 2004). In industry,attempts to pin “normal” effort down toobjective technological criteria proved fruit-less; attempts to maintain effort norms whenproductivity should have been rising oftengave rise to damaging social conflicts (Daviesand Khlevnyuk 2002; Filtzer 2002).

Because the state relied on indirect indica-tors of effort, mistakes crept into theeffort–punishment relationship. The investi-gation of low effort could yield an error ofType I that punished the innocent, and aType II error that acquitted the guilty. Errorsof both types appear to have been present.Numerous Type I errors are reflected in thevery high rates of penalization that con-demned hard workers along with ne’er-do-wells, drunks, and thieves. Such a wide rangeof behaviors was criminalized that virtually

every worker became liable to prosecutionfor something, including one-time and acci-dental violations: a broken-down commuterbus could make criminals of scores of haplessworkers. Rational managers might wish toselect the truly guilty for prosecution, theproblem workers and repeat offenders, butthe laws subjected even petty offenses toharsh penalties and managers who failed toreport offenses were threatened with thesame. As a result, the innocent were bundledthrough the courts and camps along with theguilty in extraordinarily large numbers.

Team Stalin probably knew this but didnot care. In March 1937, Stalin’s chiefinstrument of the Great Terror, NikolaiEzhov, told his officials to expect “someinnocent victims . . . Better that ten innocentpeople should suffer than one spy get away.When you chop wood, chips fly” (SimonSebag Montefiore 2003, p. 194). Stalin’sprime minister, Molotov, (interviewed byChuev 1991, p. 416) also cared more aboutcondemning the guilty than acquitting theinnocent, “never mind if extra heads fall.”

Type II errors were also clearly numerous;this is evidenced by the fact that, althoughpenalization rates were very high, offendingrates were even higher. Filtzer (2002) usedthe records of eight production branch min-istries in 1947 to show that almost one thirdof a million “labor desertions” gave rise tofewer than 55,000 convictions, for a convic-tion rate of 16 percent. In other words, ajudicial system that was supposed to “makethe chips fly” somehow failed to chop thewood. The combination of severe penaliza-tion with low conviction probability for theguilty is consistent with high-cost policingand justice administration (Gary S. Becker1968); the high rate of conviction of inno-cents, however, is more properly seen as acost of dictatorship (Simeon Djankov,Glaeser, Rafael la Porta, Florencio Lopez-de-Silanes, and Shleifer 2003), in the sensethat the dictator’s efforts to achieve a lowerrate of offending than society was willing totolerate had highly suboptimal results.



The effectiveness of the Politburo accu-mulation model rested on the dictator’s abil-ity to create a gap between the civilian wageas a “fair” return for effort, and low subsis-tence in the Gulag as the return to shirking,so that the difference between them was theintended punishment for shirking. Althoughthe Gulag did not generate an internal sur-plus for accumulation, it could still have con-tributed to the surplus in the economy as awhole. The effort curve of the accumulationmodel will be displaced downward if civilianworkers expect the Gulag wage as their pun-ishment for low effort. But the widespreadmistakes that we have described subvertedthis strategy: if workers expect Type I errorsto prevail, they will be punished regardlessof effort; if they expect to benefit from TypeII errors, they can shirk without fear.

Error rates, moreover, were not exoge-nous. They were fashioned by the counterac-tions of those threatened with punishment,who could take steps to reduce their risks.Workers and managers diverted effort fromproduction into mutual insurance: since thethreat was shared among them, they couldagree to cover up each other’s shortcomings.The archives have added detail to theprearchival literature on concealment in thefactory (Berliner 1952). Filtzer (2002) hasshown that postwar managers tolerated late-ness and absence to maintain goodwill, andcolluded with workers to underreport suchviolations, while pursuing quitters whoundermined morale and the factory’s capaci-ty to fulfill the plan. The rural police andcourts pooled risks with the rural communi-ty in sheltering the young offenders who haddeserted factories or technical schools(Kozlov 2004, vol. 1). In all these ways,mutual insurance tended to cut the individ-ual risk of punishment. The archives alsoshow how regional party officials defied eventhe most powerful central organizations toprotect their own (Khlevnyuk 2004). JamesR. Harris (1999, pp. 156–63) has describedhow the Urals regional leadership formeditself into a “protected, mutually reliant”

clique that fended off criticism from belowand investigation from above in the mid-1930s, but was eventually wiped out in theGreat Terror. Belova (2001) has shown thathigh-level patrons could protect the mostegregious embezzlers.

Faced with widespread enforcementfailures at lower levels, officials respondedin two ways. First, they aimed to under-mine horizontal trust by rewarding whistleblowing and informing through organizedcampaigns (Belova 2001). Second, Stalinforced the legal system, local party offices,and the militia to increase arrest and con-viction rates or suffer penalties themselves.The most common method of forcingrepression was to distribute quotas byregion and profession to officials at lowerlevels (Kozlov 2004, vol. 1). In the GreatTerror of 1937–38, local officials had towork feverishly to achieve a set number ofconfessions per day (A. I. Vatlin 2004). Tofulfill such plans, the police officials imput-ed individual guilt from increasingly trivialdifferences in behavior. Whether or notthese measures reduced the Type II errors,they seem likely to have encouraged falsedenunciation and confession and so addedto the errors of Type I.

Team Stalin hoped, we suppose, thatincreased coercion would induce agents totake a lower wage without withdrawingeffort, making coercion effectively costless.The evidence suggests that, in practice,those threatened with punishment raised thechance of escaping detection (a Type IIerror) through mutual insurance, but thisdiverted effort from production. At the sametime, successful collusion against Type Ierrors (false accusation) combined with theincreased chance of a Type II error to bluntthe effect of penalization on the reservationwage. Finally, the heavy direct costs of therepressive apparatus must have furtherreduced the net payoff from coercion.

We have only one empirical study of theaggregate effects of the terror on industrialproduction (Barbara G. Katz 1975), which



attributes the 1937 slowdown in industrialgrowth to political repression. New archivalstudies show the disastrous effects of theGreat Terror on important Gulag operations(Ertz 2003). Finally, the archives have shownthat Soviet leaders themselves saw a directconnection between the terror and fallingproductivity (Davies 2004): the terror elimi-nated a generation of skilled and knowledge-able managers and technical staff, worsenedlabor discipline, and damaged effort.

3.4 Sharing the Surplus

Whatever the mechanism, Stalin’s eco-nomic policy certainly generated a substan-tial increase in the excess of output overconsumption for use in investment, defense,or other public spending. But surplusesattract rent-seekers. To what extent didStalin have to share the proceeds with othersas the price of his power? This is a questionthat could not be asked before the archiveswere available. Although the archives arenow open, definitive answers remain elusive.

The archives show that Stalin presidedover a “nested” dictatorship (Lazarev andGregory 2002; Gregory and Markevich 2002;Gregory 2004) in which power was delegat-ed from one vertical level to the next so thateach organization duplicated the administra-tive control structures of its superior. Eachprincipal in turn acted as the agent of a high-er principal until the top where the great dic-tator ruled alone. This was also a complexhierarchy with nodes from which organiza-tions stemmed with partly complementary,partly overlapping functions. The dictatorruled by delegation, so under him thereruled many smaller dictators who exercisedunconstrained power within their specializedfiefs that Stalin personally allocated to them.

Recent studies of the Soviet nomenklatu-ra under Stalin, which comprised up to theorder of a million posts (Lewin 2003;Khlevnyuk 2003b), show how power cascad-ed downward through branching networksof agents that fractionalized the “aggregate”

nomenklatura into many little nomenklat-uras each dependent on an intermediateboss. With delegation and fractionalization,these would form natural market places forthe exchange of gifts and favors for loyaltyin which distributional conflicts and dealstook place.

In the 1930s, investment was the largestsingle use of Stalin’s rent. It is not easy to tellwhether he distributed it primarily to sharerents or to generate growth. In principle, therent shared would have been the excess ofoutlays in a given use over those which agrowth-maximizing dictator would have allo-cated. Even if rent-sharing took place, how-ever, it would have been rationalized underthe official guise of advancing growth, andthe dictator’s most loyal agents already occu-pied the key positions most likely to receiveinvestment.

Recent case studies throw light on themotivation behind particular transactions. Inthe late 1920s, the Politburo and its agent,Gosplan, fought against regional rent seeking,accusing “irresponsible” regions of “self serv-ing projections based on local interest, lack ofobjectivity, and inexact calculations thatundermine the very foundations of planning”(Gregory 2004, p. 80). Stalin resisted fierceregional lobbying in the second half of the1920s, rejecting projects such as theDneprstroi metallurgy complex on thegrounds that there were better uses of thefunds (Lih, Naumov, and Khlevnyuk 1995).Gosplan’s resistance evaporated with the 1929purge of those who argued for realistic plan-ning. Regional leaders descended on Moscowwith grandiose investment plans (Harris1999). Eventually Stalin set up institutionalbarriers to this kind of behavior: regionalparty leaders required official permission tocome to Moscow and enterprises were pro-hibited from maintaining representativeagents in Moscow (Khlevnyuk 2004).

Stalin’s advocacy of superindustrializationfrom 1928 on promoted his power as well ashis long-run rents. He gained allies amongregional leaders by espousing a program that



13 This study is based on full data on requests for motorvehicles from the dictator’s reserve fund in 1932 and 1933,a process that was quite separate from the wholesale allo-cation of motor vehicles through ministries and regionalauthorities. The researchers had each request, the casesmade for the allocation, letters of support, and the eventu-al outcome. The final decisions were made by a high-levelPolitburo commission headed by Molotov. The researcherscoded the cases made for the vehicle allocation as “eco-nomic,” for example a note from Gosplan explaining whythis allocation was vital to fulfill the plan, or “political,” forexample support from a political patron. In the regressions,only the political variables were significant in explainingapprovals. Economic variables were insignificant and alsohad the wrong signs.

promised unlimited funding. Stalin’sappointments diary is full of meetings withregional leaders who made the pilgrimage toStalin to seek favors. Stalin was uncharacter-istically concerned in 1931–32 that his homerepublic Georgia was “on the verge ofhunger” and of “bread riots,” while he made“feigning hunger” a counterrevolutionaryoffense in other republics (Getty andNaumov 1999, p. 69). Lazarev and Gregory(2003) have analyzed the dictator’s distribu-tion of motor vehicles, the scarcest capitalgood of the time, from his own reservefunds. The strong econometric results showthat Team Stalin allocated vehicles for polit-ical gain and that economic considerationswere not significant.13

Stalin needed a growing military andindustrial base and could not have allowedrent seeking to randomize investmentbeyond a point. Stalin opposed the “unnec-essary spreading of investment” that resultedfrom industrial and regional authorities’competing bids to start up favored projects(cited by Davies and Khlevnyuk 1997, p. 41).A first-hand account shows Stalin advocatinginvestment rationality at a meeting in 1947:“The plan is very inflated and is not withinour capacity. We should give money only toprojects that can be put into commission,and not spread it out among many projects.They are building all kinds of nonsense innew, unpopulated areas and they are spend-ing a lot of money. It is necessary to expandold factories. Our dear fantasists design only

new factories and inflate construction” (citedby Khlevnyuk 2000). In the early 1930s, thePolitburo, in a rare show of opposition toStalin, delayed the Baltic–White Sea Canaluntil Stalin reduced its cost by proposing theuse of “cheap” Gulag labor (Morukov 2003).

The Red Army provides a case study inStalin’s relationship with a powerful specialinterest. Sokolov (2005) has shown that thedemands of the armed forces were signifi-cant in Stalin’s turn to forced collectivizationand industrialization in 1929–30, and themilitary leaders themselves supported hisresort to a command economy to bring thesethings about. Stalin quickly moved to raisemilitary spending in secret in 1931 (Davies1993). On the other hand, he kept profes-sional soldiers out of the Politburo. He alsoconsistently resisted the attempts of militarymodernizers, such as Marshal M. N.Tukhachevskii, to use plans for a mechanizedmass army to justify a military role in manag-ing the defense industry (Lennart Samuelson2000). Evidently alert to the danger that thearmed forces could become just another spe-cial interest, Stalin encouraged rivalry andtension between the army and the industrial-ists who might otherwise have formed natu-ral allies (Gregory 2003b; Harrison 2003b).While Stalin lived, therefore, there was noemergence of a “military–industrial com-plex,” whatever may be said of the SovietUnion under his successors.

Similarly, there is evidence of rent seekingin military R&D but no evidence that Stalinor his subordinates wanted it that way(Harrison 2003a, forthcoming-b). He pun-ished rent seeking where he suspected it,and this helped to make the intermediariesthat he charged with funding this work keento get results.

4. Money, Prices, and the Seller’s Market

A third set of issues addressed by archivalresearch considers how plans and commandsare nested with markets and money, howmarkets work, and how contracts areenforced in the context of a seller’s market.



These issues connect literatures that are notoften linked, dealing with markets and hier-archies, contracts and enforcement (AvnerGreif 1993, 1994, 2000; Greif, PaulMilgrom, and Barry R. Weingast 1995), andthe soft budget constraint (Kornai 1980;Maskin 1996, 1999; Kornai, Maskin, andRoland 2003).

4.1 The Soft Budget Constraint

Kornai (1980) developed the concept ofthe soft budget constraint to explain shortagesand sellers’ markets in Soviet type economies;he attributed the soft constraint to the state’spaternalistic domination and its desire toinsure enterprises against failure. The result-ing moral hazard damaged firms’ incentive toeconomize on inputs and created an insa-tiable hunger for resources that was eventual-ly transmitted to permanent shortages andqueues in retail markets. Kornai (1980) main-tained that such shortages could not be elim-inated by price flexibility. An early critic,Stanislaw Gomulka (1985), suggested thatprice rigidity must be a necessary condition ofshortage, specifically that input prices mustbe less negotiable than firms’ budgets, butthere did not appear to be a convincing expla-nation of why the state should cap inputprices. Others have attempted to deduce thesoft budget constraint from the formal rulesand incentives arising from specific institu-tions, rather than from Kornai’s politicalnegotiation process; for example, M.Dewatripont and Maskin (1995) suggestedthat the softening agent is the sunk costs thatarise when there is sequential monitoring oflong-lived enterprises and their financing iscentralized. Underlying all these variants is aproblem of dynamic commitment (Mark E.Schaffer 1989): the state does not intend tosupport loss-making ventures but is unable tocommit itself not to after the event, and thoseresponsible for the losses anticipate thisbeforehand.

In contrast, Shleifer and Vishny (1992)and Wintrobe (1998) departed fundamental-ly from Kornai, questioning how excess

14 Kornai ought to have been right; he had the advan-tage over others of witnessing the formative years of theHungarian planning system as an insider.

demand could be transmitted from firms tothe retail market and arguing that shortagecould be explained more parsimoniously byfirms’ withholding supplies and pushingprices down in order to collect bribes fromconsumers.

The literature thus offers a range ofhypotheses for archival research. Kornaiwould look for a political decision by TeamStalin to tolerate loss-making activities so asto exploit the economy for political goals.Dewatripont and Maskin would look for therules and incentives that made funding prin-cipals prefer to tolerate losses. Gomulka andothers would expect to see the price con-trols of which shortage is a by-product,while Shleifer and Vishny suggest that short-age is the intentional creation of rent-seek-ing producers who restrict output and lobbyfor low prices.

Archival investigations have shown that allthe theoretical explanations of the soft budg-et constraint have some merit; they do notsupport the Shleifer–Vishny critique. Thedocuments of the formative years tellKornai’s story of a mobilizing state overrid-ing the market to gain discretionary powerover resources.14 In the 1920s, the rulingBolsheviks allowed industry to go over fromself-financing to a regime of price controlsand subsidies for the sake of its plans forindustrial and military mobilization; mili-tary–industrial interests and military lobby-ing were more important in this process thanwas previously recognized (Sokolov 2005).The trusts that were later transformed intoministries were already shifting funds fromprofitable to loss-making enterprises in themid-1920s (Gregory and Tikhonov 2000). Asthe state won control of agriculture throughcollectivization and the first five year planwas enacted, soft budget constraints wereextended to industry and investment as awhole (Harris 1999). Specific legislative



attempts to harden budgets were quicklyabandoned (Gregory and Tikhonov 2000).These episodes show that budget constraintsbecame soft through a direct exercise ofpower that changed the relationshipbetween the state and the economy. To thedictatorship, the loss of financial credibilitywas a price worth paying for the freedom toits pursue wider goals.

More detailed case studies have alsoshown the value of institutions-based theo-rizing. Long-lived R&D projects in defenseindustry, for example, were funded and mon-itored in installments. The resulting sunkcosts made the authorities ready to refinanceprojects that they would not have financedinitially with complete foresight (Harrison2003a). Even in this case, however, the docu-ments reveal an intrinsic element of negotia-tion that is not captured by a model based onformal rules alone. Funding principals couldnot or did not wish to commit to explicitrules. They understood the games thatunscrupulous agents could play with suchrules and they sought to regain financial con-trol in two ways. First, they cut funding backfrom time to time in ways that were to someextent arbitrary. Second, even if establish-ments were never closed for making a loss,they could be put under new managementand the old management could be demotedand punished individually. For self-protec-tion, agents built mutual insurance networksand also engaged in cutthroat rivalry to pro-tect funding. Thus, well after the formativeyears, the softening of budget constraintscannot be understood without close attentionto political processes of vertical bargainingand horizontal rivalry and collusion.

Does the story of the soft budget con-straint stand up as an explanation of Sovietretail shortage? The transmission mecha-nism, doubted by Shleifer and Vishny (1992)and Wintrobe (1998), has been substantiatedfrom unique access to the post-StalinGosplan archives by Byung-Yeon Kim (1997,1999, 2002). According to Kornai (1980, pp.486–88), the soft budget constraint would

lead to retail shortage if firms used claims oninputs to “siphon” general-purpose com-modities intended for the retail market backinto production. Kim showed that siphoningreally did drain off resources that should havebeen available for private consumption, thatactual private saving was persistently higherthan suggested by published budget surveys,and that significant private saving was forcedby firms’ siphoning activity as a result.

Budget constraints were soft but not limit-less; what were the limits on softness? Kornai(1980, p. 211–14) defined three “tolerancelimits” on the state’s readiness to accommo-date inefficient behavior: foreign currency,social unrest, and administrative complaints.The view of Stalin’s economic policy set outabove confirms the importance of all three.Stalin personally kept an iron grip on foreignexchange. The unrest associated with con-sumption shortfalls is represented by the dis-order constraint in figure 1. Finally, thePolitburo carefully monitored the flow ofcomplaints up the vertical hierarchy for signsof general worker discontent.

4.2 Money and Prices in the Seller’s Market

In theory, money was relatively unimpor-tant in the Soviet command system. Firms’soft budget constraints created seller’s mar-kets, the so-called “dictatorship of the sell-er.” Prices were ineffective in incentivizingproducers and signaling allocative needs(Kornai 1980). Instead, the planners’ “visiblehand” was supposed to direct allocation ofproducer goods at controlled prices (Nove1961; Ellman 1972, 1979; Gregory andStuart 1974). The banks supplied money andcredit to “follow” physical plans; under earlySoviet rule this was called “planned automa-tism” (Gregory and Tikhonov 2000). Moneymerely enabled planners to practice “controlby the ruble,” monitoring financial flows todetect departures from physical directives.The money stock was supposed to be strictlysegregated into bank money for interfirmtransactions and cash money for wage pay-ments. Given that firms could use money



only for planned purposes, there would belittle incentive for them to acquire addition-al holdings. The state budget mainly provid-ed a façade behind which the authoritiescould divert resources from profit-makingenterprises to loss-makers and the military(Igor Birman 1981).

A major surprise from the archives is thatmoney played a much larger role than weexpected. Allocation actually began not withphysical supply plans but with nominal budg-etary assignments to investment and othergovernment uses such as military orders.The Politburo gave much more time andenergy to how rubles would be spent than toconsideration of the “control figures” for out-put in physical units (Davies 2001a; Gregory2001; Davies, Ilic, and Khlevnyuk 2004).

Budget outlays usually came first becausebroad-brush supply plans could not fix thedetailed assortment of physical products ortheir final uses. Plans in rubles of outputwere then calculated at “fixed” plan prices.Plan targets had to be fixed in rubles becausemost producers supplied many products andoutput was too heterogeneous to be plannedany other way. Supply quotas binding onindividual ministries and enterprises werealso denominated in rubles (Harrison 1998).While the government might announce aplan target for steel tonnage, the directiveplan for the enterprise was in rubles; intera-gency contracts were supposed to link thetwo, but were usually incomplete or hardto enforce, even in the defense industrywhere compliance was monitored intently.Markevich and Harrison (forthcoming)report the case of an aircraft factory in 1934where the same managers, reprimanded forpoor-quality work, were rewarded days laterfor fulfilling the plan—in rubles, of course.

Stalin-era plans were too aggregated to tieproducers to particular products or users toparticular suppliers; this suited both min-istries and planners. Planners could not dis-aggregate plans efficiently, and preferredthe responsibility for disaggregation andsubcontracting to lie at the ministerial level

or below. Ministries also liked this arrange-ment because it freed them to decide howbest to fulfill aggregate ministerial output inplan rubles.

Final allocations of products wereachieved through contracting between min-istries, ministry main administrations, andenterprises. Research on the late Soviet-eraeconomy showed that this contractingprocess was relatively decentralized (HeidiKroll 1986, 1988). In the Stalin era, supplyplanning set limits on this process but thelimits were broad. The ministry’s annual orquarterly supply plan was only the first salvoin the “battle for the plan,” in which users ofintermediate goods entered bids with pro-ducers for the contracts needed for theirown plans (Harrison and Nikolai Simonov2000; Gregory and Markevich 2002). At thispoint, command-economy allocation becamea market-like contracting process; officialretail markets that traded state goods openlyat state prices and retraded them under thecounter at a premium were just at oneextreme of a continuum of market-orientedactivities that even included bizarre activitiessuch as home production of automobiles(Lazarev and Gregory 2002).

Decentralized contracting generated adegree of price flexibility, and this tells usmuch about the motivations, resources, andconstraints of the agents involved. By devis-ing legal and illegal ways to bid up contractprices, suppliers could fulfill both plans andcontracts with less effort and more financialgain. According to the stereotype, priceswere supposed to be fixed from above on thebasis of initial costs plus an allowance foroverhead and taxes. Longstanding empiricalconcerns about official price indexes (ColinClark 1939; Alexander Gerschenkron 1947;Bergson 1947, 1953, 1987; Peter J. D. Wiles1982; Hanson 1984), including the exploita-tion of new products specifically to free theenterprise from fixed plan prices (Berliner1976; Harrison 1998), suggested that thefixed-price assumption might oversimplify.Kornai (1980, p. 363) specifically noted the



interest of the firm in higher prices.Nonetheless, standard accounts of Sovietallocation took it for granted that the firmwas a price taker (Nove 1958; Edward Ames1965; Michael Manove 1971; Richard E.Ericson, 1983, 1984; Stephen M. Goldfeldand Richard E. Quandt, 1988, 1990; Shleiferand Vishny, 1992).

The archives show that price-setting wasone of the most important activities of Sovietfirms (Harrison 1998); producers’ pressure onprices was relentlessly upward, not downwardas Shleifer and Vishny (1992) predicted.Actual transaction prices were negotiatedbetween buyers and sellers during “contractcampaigns” loosely managed from above(Harrison and Simonov 2000; Gregory 2004).A memo described the consequences of fail-ure to pay an “illegal” high price: “If you don’twant to pay, we’ll keep this in mind when weconsider your next order” (cited in Gregory2004, p. 220). Official prices were supposedto be used in more important transactions butofficial price handbooks were often incom-plete, lagged behind new products, or wereignored. The mammoth metals administra-tion of the ministry of heavy industryemployed only three persons in its pricingdepartment, which set the official prices ofmetallurgical products (Gregory 2004). Thedefense ministry was particularly vulnerableto inflationary pricing because of the rapiditywith which its product requirements werechanging. Military buyers complained ofprices based on “how much it costs whetherthe result of correct work or poor manage-ment.” (Gregory 2004, p. 220). Defense sup-pliers withheld information about costs on thegrounds that it was too sensitive to entrust tothe defense ministry. They would delay set-tlement and hold out for higher prices andillegal advance payments before agreeing toterms (Harrison and Simonov 2000).

Figure 2 illustrates the resulting bargain.The plan of military orders obliged industryto deliver a volume of output Q∗ at a presetprice P∗. Industry’s problem was that to pro-duce Q∗ cost effort; it would be more advan-

tageous if the defense ministry could bepersuaded to settle for a smaller real quan-tity, say Q� in the figure, at some higherprice level P� along a unit-elastic curvethrough Q∗, P∗. It did this by reporting high-er costs relative to the true effort of produc-ing them, and by introducing new productsat a higher ratio of reported cost-to-effort,i.e., by simulating productive effort.

Price renegotiation was costly, however.At some margin, the simulation of produc-tive effort became as costly as productiveeffort itself. Planners monitored the pricesof existing products although they could dolittle to control the prices of new ones(Harrison 1998). Faced with price gouging,a powerful buyer like the army could com-plain through the defense minister to thePolitburo on which he sat (Davies andHarrison 1997). Less powerful customerscould and did complain frequently to thestate arbitration commission (Gregory2004) provided they were willing to risk dis-rupting good relations with suppliers. It isnoteworthy that the Soviet authoritiesapplied the term “signalization” to thescarcity information forthcoming from thearbitration courts.

The widespread evidence on illegal priceincreases raises a number of questions. Ifenterprises had soft budget constraints, whydid they make the effort to push up pricesand incur the legal risks rather than wait pas-sively for an automatic subsidy? The answeris that subsidies were not automatic and soft-ening budget constraints took effort. Theenterprise had to equate the effort costs ofovercoming resistance at several marginsrepresented by the ministry which couldswitch funding from profit to loss makers,the budgetary and credit authorities, and abuyer that had to contend with its own budg-et constraint. Overcoming the resistance ofthe buyer enabled one enterprise that facedmore resistance from superiors to pass itsproblem on to another that faced less. Thisphenomenon expressed itself in illegal butpersistent interenterprise payment arrears,



Figure 2. Price Renegotiation

Price

Industry’s Plan of Gross Output

P�

P∗

Quantity

∗∗

QP P Q

·

Q∗ Q�

Source: for the full reasoning underlying this figure, see Gregory 2003b and Harrison and Kim 2004.

ultimately made good by the state bank aslender of last resort (Gregory and Tikhonov2000).

Since producers had to spend effort andincur risks to extract it from superiors andbuyers, it follows that money was not free.Despite the supposed firewall betweentraceable bank money and anonymous cash,increases in official credits were in factaccompanied by increases in cash holdings(Gregory and Tikhonov 2000). The impor-tance of money is underscored by Stalin’sauditors who uncovered numerous cases offraud and embezzlement that resulted in theaccumulation of caches of illegal money(Belova 2001). Apparently, not all scarcitymarkups were lost in lobbying costs or favorsto third parties.

Unauthorized cash did not necessarily goto line private pockets. Harrison and Kim(forthcoming) argue that the main purpose ofhidden inflation and siphoning was to relieveeffort. Thus Soviet corruption differed froma conventional picture of bribe-taking for

personal enrichment: managers extractedside payments from buyers more usually toengage in siphoning and fulfill the plan, sothat the proceeds of corruption were recy-cled into production. Planners could shiftthe purposes of corruption from embezzle-ment to siphoning up to a point by raisingplan tension, although this helped the plan-ner more than the consumer. It follows,Harrison and Kim suggest, that the lowerplan tension under Brezhnev (Schroeder1985) may have encouraged disloyal or pri-vately motivated embezzlement.

4.3 Commands, Contracts, andEnforcement

To persist, hierarchies and markets mustboth evolve ways to motivate repeated partic-ipation. According to Greif (2000), peoplecontinue to participate in an institution whenit is in their interest to do, conditional upontheir expectation that others have made thesame calculation. When transactions are



sequential and the completion of a contract rests on advances by one agent to another, the institutional equilibriumrequires an enforcement mechanism to pre-vent cheating by some that makes abstentionby others their best option.

The command system combined hierar-chy with elements of horizontal dealing;the archives show that the dictator gave histacit approval to this. When hierarchies andmarkets coexist, they must find a jointequilibrium in the sense that changes inthe marginal returns are likely to induceagents to reallocate effort from one to theother. An increase in market returns, or areduction in the power of principals, wouldshift resources out of the vertical hierarchyinto unofficial horizontal dealing. Thus ver-tical enforcement relies on the ability ofprincipals to reward loyalty and punish dis-loyalty and so offset the potential gainsfrom horizontal exchange.

The problem of horizontal relationshipswas particularly acute given the persistenceof outside options for agents in the hierarchy.Stalin may have aspired to monopolize theeconomy into a corporation that was integrat-ed on both horizontal and vertical lines, butthis aspiration was not and could not be real-ized. Bureaucrats could choose to deal pri-vately with each other rather than throughsuperiors. There was a thriving labor marketthat moved workers from enterprise to enter-prise, even when coercion was at its mostintense. Everyone could turn either to legalor illegal commodity markets. Outside thecountry, citizens could sell secrets to a foreignbusiness or power or defect to a foreign coun-try. In this general sense, the final collapse of the Soviet command system was a failureof vertical enforcement and compliance(Harrison 2002).

According to Greif (2000), the “funda-mental problem of exchange” is the possibil-ity of horizontal cheating in markets. Inhierarchies, there is a parallel fundamentalproblem of command (Harrison 2005): anagent may exploit the principal by shirking

or stealing. The returns to stealing, however,are reduced if what is stolen cannot be sold,and this is more likely when market transac-tions cannot rely on the law and depend onprivate enforcement alone. Therefore theenforceability of vertical commands is likelyto vary inversely with that of private con-tracts. Stalin made use of this trade-off byprivileging socialist property over privateproperty and enacting harsher penalties fortheft from the state than from fellow citizens(Gorlizki 2001b).

When agents resorted to unauthorizedbusiness contacts, the ability to completetransactions depended on a business reputa-tion for plain dealing and keeping one’sword. Belova (2001, 2005) has analyzed the“relational” contracting system that arosewhere unauthorized contracts were costly toenforce or unenforceable in the courts:agreements rested on a handshake and wereenabled by personal contact and friendshipwhich alone could overcome the culture oflow trust.

Even loyal agents, however, had to engagein some unauthorized market trading to ful-fill plans that were incompletely providedfor. Despite the threat to dictatorial power ofunauthorized horizontal trade, the clearmessage of the industrial archives is that a“good” manager was expected to get the jobdone by all means necessary and at any price[liuboi tsenoi]. The minister of heavy indus-try, for example, bluntly relayed this messageto his managers: “We will not listen to thosepeople who say our materials have not beendelivered, but we say that a good manager, agood shop director, a good master technicianknows how to organize things and producethe required results” (cited by Gregory 2004,p. 164). In short, the command system reliedon loyal agents’ unauthorized horizontaldealing on behalf of the plan, despite thefact that disloyal trading detracted from thedictator’s goals.

Principals could not necessarily distinguishthe unofficial deals that agents made to fulfillthe plan from those that lined their pockets.



A study of the last Soviet-era anticorruptiondrive, based on rare access to the post-Stalinjudicial archives (Luc Duhamel 2004), sug-gests that disloyal agents would always claimloyal motivation, and sometimes the investi-gators did not want to know because theyhad a political agenda. Belova (2001) arguesthat, in the Stalin era, principals insuredagainst disloyal corruption by tolerating hor-izontal networks while restricting theirscope. As a result, horizontal reputationcould only accumulate within small groups.This limited private exchange by reducingthe efficiency of relational contracting.

5. Efficiency, Growth, and Reform

The debate over the feasibility of centralplanning occupied leading economicthinkers for a century. Barone’s “Ministry ofProduction in the Collectivist State” (1908,reprinted 1935) was followed by Mises’“Economic Calculation in the SocialistCommonwealth” (1920, reprinted 1935) andHayek’s many writings on the informationproblems of a nonmarket economy (1935,1937, 1940, 1945). Bergson’s two surveys(1966, 1967) remain authoritative accountsof the controversy. Following the GreatDepression and World War II, some econo-mists argued that socialized investmentcould offset high private discount rates,relieve bottlenecks, or eliminate the busi-ness cycle. Even Bergson (1948, reprinted1966, p. 230) suggested that “in a highlydynamic economy, a centralist allocation ofinvestment might lead to fewer and smallererrors than a competitive allocation.”Despite the collapse of the Soviet Union andChina’s gradual conversion to a market econ-omy, the debate over the merits of socialismcontinues (Don Lavoie 1985; James Junker1992; Pranab K. Bardhan and John E.Roemer 1993; Shleifer and Vishny 1994;Peter J. Boettke 2001).

5.1 Efficient Choices

The archives illustrate the Hayekian prob-lem: the decisionmakers could not gather or

process the information required to makeefficient decisions. The information anddecision overload came in part from thePolitburo’s inability to distinguish the impor-tant from the trivial. Decision-makingresources were scarce; in the 1930s Stalinand the Politburo made between 2,300 and3,500 recorded decisions per year (Gregory2004). Much time was spent on trivia, fromindividual permits for foreign travel to loca-tions for monuments and vegetable andmetro prices in Moscow. Issues defined asmajor were delegated to ad hoc commis-sions with very high transactions costs: forexample, a special commission made up ofthe chairman of Gosplan, who doubled asdeputy prime minister, the head ofGosplan’s fuel commission, timber ministryofficials, and regional officials had to meet todecide on a requested cut in timber suppliesequal to a fraction of one percent of totaloutput (Rees 1997b).

While Team Stalin labored over detailsthat were often trivial, it delegated majordecisions of project choice by default. Whilethe Politburo required major projects to beapproved on official “title lists,” the produc-ers themselves largely determined how tospend their investment rubles. The title listslacked cost estimates in many cases despitethe efforts of Gosplan, Gosbank, and thefinance ministry to enforce formal rules oncost discipline (Belova and Gregory 2002).

When information was not simply lacking,it was limited by opportunism. Investmentcontractors opportunistically refused to pro-vide information; they complained, forexample, that cost audits would delay urgenttasks or compromise state security (Harrisonand Simonov 2000; Harrison 2004). Moregenerally, those who possessed valuableinformation tried to monopolize it andshared it only when they could extract a rentin exchange.

Information problems led to catastrophicinvestment blunders such as theBaltic–White Sea Canal, which was too shal-low for effective use, and the Baikal–Amur



Mainline which remained unfinished forforty years (Khlevnyuk 2001b; Joyce 2003).Investment projects initiated without feasi-bility studies and budgets proved unexpect-edly difficult to finish. Increases inunfinished construction rose sharply in crisisyears such as 1931–32 and 1936–37 whenthe proportion of gross investment thatmerely added to work in progress rose above15 percent of the total (Harrison 1994).Market economies are not immune toinvestment blunders, but no one has yetcompared unfinished construction in Soviet-type and market systems. One reason, wespeculate, is that statistical offices do notsee the problem of unfinished construc-tion under market arrangements as worthmonitoring.

Not all high-level decisions were ineffi-cient or irrational. Had this been the case,the economy could hardly have grown, sur-vived the onslaught of Germany in WorldWar II, or created modern nuclear andaerospace industries. In some respects,Stalin’s economic decision making did bet-ter than would have been predicted bythose who regarded socialist planning asinfeasible. This mixed evaluation wasalready present in the literature before thearchives. Studies confirmed low rates oftechnological improvement and diffusion(Antony C. Sutton 1968, 1971, 1973;Berliner 1976; Amann, Julian Cooper, andDavies 1977; Bergson 1978; RonaldAmann and Cooper 1982, 1986; Hansonand Keith Pavitt 1987) due to the lack ofincentives for producers to adopt newtechnologies and planning “from theachieved level.”15 But the same systemalso gave rise to some spectacular firstsand near firsts, especially in the technolo-gies of defense and heavy industry. While

15 Russians used this expression to convey a rule-of-thumb that sets the next target equal to the previous levelachieved plus an arbitrary increment. Birman (1978)described the practice and Martin L. Weitzman (1980) andMichael Keren (1982) formalized it in the concept of theratchet effect.

it is true that invention was not typicallymatched by innovation (Amann 1986), therecord of invention was still remarkable fora relatively poor country.

Archives and interviews have unraveledthe decisions that established new atomicand aerospace industries after World War II(David Holloway 1994; Simonov 2000;Harrison 2000). Notably, decisionmakerstended to reserve their scarce attention andavailable funding for military priorities, andthe decisive information that signaledwhere resources should go came from themonitoring of progress abroad.

Previous insider accounts of Soviet inven-tions were highly selective; they concentrat-ed on success stories and gave self-servinginterpretations of what motivated success-ful inventors and designers. The archivesrecord the failures as well as the successes,and Harrison (2000, 2003a, forthcoming)formulates a framework for Soviet inven-tion as an economic activity: the initial offerof funding in a field such as aviation creat-ed a “market for inventions” that attracteddesigners and projects in large numbers. Asone would expect, the designers tended tobe heterogeneous in talent and motivation;on a first pass, it would seem that they weremotivated by a varying mix of intellectualcuriosity, the expected reputational rewardsfor breakthrough projects, and the fundingthat could be consumed before a projectthat failed was terminated. Notably, the dis-tribution of scientific reputation did notrely on the state for enforcement; it wasconferred, at least in part, by the communi-ty of specialists. At the same time, the pro-longed refinancing of unsuccessful projects,for example for steam-powered bombers,suggests that the uncertainty and informa-tion biases surrounding R&D permittedadverse selection and provided fertile soilfor rent seeking.

In contrast, there was no reputation to begained from replicating a technology “notinvented here.” A variety of studies(Holloway 1994; Nataliia Lebina 2000;



Harrison 2000) have shown that, while Stalinregarded the fact that a technology hadalready been developed elsewhere as a posi-tive signal, Soviet nuclear, missile, and air-craft designers disliked being ordered tofollow in others’ footsteps without imposingtheir own stamp; this sometimes resulted inbackward steps. A case study of aeroenginedevelopment (Harrison forthcoming) showsthat, when a breakthrough was still antici-pated, the inventors were willing to work forrelatively modest material rewards that werespread around fairly evenly because themajor prize for which they competed wasthe reputation for priority in an invention.After the breakthrough, when a reputationfor priority could no longer be won, thesame people required monetary rewardsthat were many times higher than beforeand much more regressive to engage in thedevelopment work.

5.2 Planning by Feel

The concept of planning by feel is illustrat-ed by an anecdote: in Moscow in 1981, aSoviet historian told one of the authors thatthe great skill of Gosplan director, NikolaiVoznesenskii, as an economic planner wasthat “he could feel the disproportions in theeconomy through his fingertips.” Thearchives reveal that this was not just a cleverremark. Overwhelmed by unreliable informa-tion and computational constraints, the plan-ners had to turn to intuition andrules-of-thumb based on experience. Some,like Voznesenskii, appeared to have betterintuition than others.

Planners were supposed to distributematerials according to engineering norms,but the first allocations took place beforenorms were compiled (Gregory andMarkevich 2002; Gregory 2004). Supplyagencies used intuition, trial and error, and“historical experience.” According to onesupply official: “We give 100 units to onebranch administration, 90 to another. In thenext quarter we’ll do the reverse and seewhat happens. You see, we do this on the

basis of feel; there is no explanation”(Gregory and Markevich 2002, pp. 805–06).According to another: “Our problem is thatwe can’t really check orders and are notable to check them. . . . We operate partial-ly on the basis of historical material—weare supposed to give so and so much in thisquarter, and at the same time you are sup-posed to give us this much.” (cited byGregory 2004, p. 172). Workers’ piece-ratenorms were also set from the achieved levelor just “by eye” (Davies and Khlevnyuk2002, p. 877). Mises (1949, reprinted 1998,p. 696) predicted that in the absence ofeconomic calculation planning would bereduced to “groping in the dark.” This, thenis how they groped.

Ministry and supply officials understoodthat enterprises, being “greedy oppor-tunists,” demanded “too much.” When theminister of heavy industry asked: “Tell usplease how our enterprises received 50 per-cent of supplies they requested and fulfilledtheir production programs 100 percent?”the response was: “In July, I told my peopleto prepare a report about the fulfillment ofplans for individual branches. I then beganto edit these reports and saw that the pro-duction programs had been fulfilled 102percent but only 40 percent of supplies hadbeen received. I believed there was some-thing left over from the previous year. Ilooked into the report for the previousyear—again 103 percent and supplies only40 percent. I couldn’t look at the year previ-ous to that because I could not find it” (citedby Gregory 2004, p. 172). When the firstvehicles began rolling off assembly lines in1933, the producer demanded the entireyear’s output for its own use (Lazarev andGregory 2002, pp. 329–30); Gosplan retali-ated by ordering customers to “liquidatesloppy and unjustified requests.”

Without norms and with “sloppy” materialrequests, initial allocations of materials andequipment were arbitrary but reached equi-librium with surprising speed. As the distri-bution of domestically produced vehicles



16 From a starting point of violent disagreement overthe distribution of vehicles, the correlation coefficientsamong the distribution plans prepared by the three agen-cies involved—producer, Gosplan, and the responsiblePolitburo commission—rose to above 95 percent by thethird quarter of 1933 despite their different agendas(Lazarev and Gregory 2002, p. 336). The initial endow-ment was essentially random; planning from the achievedlevel then locked users into an “endowment” that driftedincrementally in subsequent periods.

began in 1932, only fifteen percent of orderswere filled, rising to fifty percent by thefourth quarter of 1934 and to almost ninetypercent by the first quarter of 1937 (Lazarevand Gregory 2002). This rapid convergenceis not explained by the increase in supplyalone. Enterprises, government agencies,and individuals had a virtually unlimitednotional demand. Their effective demand,however, was limited by the expectation thatthis year’s allocation would start from lastyear’s; this illustrates planning from theachieved level. Once initial allocations werefixed, a basic distributional consensus quick-ly emerged.16 Given the difficulty of balanc-ing supplies with notional demands forcommodities in short supply, transactioncosts were limited by adapting effectivedemands to the existing distribution. Whilelimiting transaction costs, therefore, planningfrom the achieved level also limited structur-al adaptation. Even worse, and contrary towhat has been argued (Manove 1971), it didnot necessarily eliminate large initial errorssince it induced expectations to converge oninitial allocations however inefficient.

5.3 Obstacles to Reform

In the early 1960s, Soviet officials began todiscuss openly whether it was possible to nestthe informational and incentive advantagesof markets within the hierarchical structuresof the command economy under the bannerof socialist economic “reforms.” The archivesshow, however, that they had been pursuingthis quest behind closed doors from the early1930s. The need to reform became evidentto insiders from the start, refuting the text-book stereotype of a planning system that

worked well while the Soviet economy wasless developed, and needed reform only afterindustrial modernization.

The story that Davies (1996) tells is rough-ly as follows. The difficulties of supply plan-ning in 1929 and 1930 quickly convincedStalin’s industry chief, Ordzhonikidze, thatdetailed interplant transactions should bedecentralized. By 1931 he had become akeen advocate of cost accounting and theidea that, given harder budget constraints,enterprises could subcontract for supplies ina decentralized way without planners. Hiseconomic mechanism to enforce budgetconstraints and make managers automatical-ly accountable for their own choices wasclearly enunciated: “If you supply [thebuyer] everything on time in accordancewith the contract, you will get the appropri-ate payment . . . If you don’t meet the obli-gations that you have taken on, [the buyer]won’t pay you, the bank won’t pay you on[the buyer’s] behalf any more, and you, dearcomrade, will have to have a very hard thinkabout how to pay your wages, how to carryon the work at your factory” (cited by Davies1996, p. 12: emphasis added). Some ofOrzhonikidze’s officials went still further,advocating the hardening of investmentbudget constraints by financing it on thebasis of repayable loans, and a considerabledegree of price liberalization.

Three barriers to reform reinforced eachother. First, Stalin and Molotov regardedthe retention of money and the return tocost accounting after the chaos of 1929–30as essentially temporary expedients. Theycould have no intellectual sympathy for hardbudget constraints, particularly for invest-ment. Second, these reservations appearedjustified by events. At the end of 1932,Ordzhonikidze unexpectedly cancelled cen-tralized equipment supply plans for the iron,steel, coal, and oil industries for 1933, andtold producers and users to sort it out them-selves. “This sudden freedom caused panic”(Davies 1996, p. 269). On the one side,equipment suppliers had no instructions on



how to ration supplies; in law they werecriminally liable if they refused an order. Onthe other hand, the equipment purchasers,freed from all restraint, tried to place ordersthat were vastly inflated, but succeeded inplacing only a small fraction. Still committedto the reform, Ordzhonikidze turned thebalancing of supply and demand over to aconference of industry representatives inMoscow. But, in the spring of 1933, the Politburo instigated the dismissal of the most radical reformists, leavingOrdzhonikidze frustrated and humiliated.

A third reason for reform failure lay under-neath the other two: “With the high level ofinvestment which obtained in 1932, such areformed system was quite impracticable,and it may have been entirely incompatiblewith Soviet economic objectives” (Davies1996, p. 267). If so, then Ordzhonikidze wasat fault in failing to understand this.

The archives reveal that the reform stale-mate of the post-Stalin years dates to thebeginning of the Stalin era. Agents could nottrust principals not to make mistakes, point-ed to stupid, contradictory, changing ordersand excessive meddling, and demandedgreater autonomy. But when they wonautonomy they behaved opportunisticallyand sought rents rather than profits.Principals could not trust agents to behavealtruistically and eventually had to interveneto curb rent-seeking and restore order.Agents behaved the way they did becausethey rationally expected principals to inter-vene. Living in a nested dictatorship, theysaw no role for the dictators at every levelother than to dictate. As long as they expect-ed this there was little point in looking forfinancial savings that were could be confis-cated at any time. Better to go on hoarding,concealing, and bargaining.

6. What’s New?

What do recent studies of the Sovietarchives offer to the political economy ofcommand systems and dictatorship?

First, the archives show a powerful but fal-lible dictatorship comprised of Stalin and hisPolitburo immersed in the detail of econom-ic decisions. They guided the general direc-tion of the economy by infrequent majordecisions that set aggregate investment,which they poorly controlled, while reservingthe right to make unlimited detailed inter-ventions in current operations. The com-mand system that resulted featured extremecentralization and the priority of vertical overhorizontal relationships. Minor decisionswere delegated from top to bottom through ahierarchy of “nested” dictatorship. Formalrules were avoided in favor of ad hoc decisionmaking. A compliant planning board was dis-engaged from responsibility for detailed allo-cation to guarantee its loyalty; the dictatorwas particularly loathe to delegate economicdecisions to politically unreliable tech-nocrats. As a result, delegation did not workwell: subordinates funneled even trivial deci-sions upwards to limit their own exposure,placing a “dictator’s curse” of excessiveadministrative burdens on their superiors ateach levels, most heavily on Stalin himself.

Second, in the context of the great increasein investment in the 1930s, the record ofdecision making in the Politburo supportsthe hypothesis of a dictator interested ininvestment, growth, and efficiency, while bal-ancing these against other objectives such asretaining loyalty and avoiding revolts. Hismanagement of investment required unflag-ging attention to the effect of consumptionon worker incentives. The fear of consump-tion falling to a point where worker unrestwould spread contributes to explaining thecyclical behavior of investment.

Third, the archival literature sheds consid-erable light on limits on the power of thedictator to extract a surplus through large-scale coercion. Frightened and intimidatedpeople still did not do as they were told;rather, they invested all the more in horizon-tal transactions that protected them fromrepression and diverted effort from plannedgoals. One of the most surprising results of



research on the Gulag is that its own admin-istrators regarded it as an economic andsocial failure, and that its supreme adminis-trator was the major force pushing for its liq-uidation shortly after Stalin’s death.Coercion was scaled down not because new,more principled leaders came to power butbecause the old unprincipled leaders knew itdid not work. This insight raises an interest-ing problem: if, as many have argued, theSoviet-type system cannot function withoutcoercion, what should it do when coercionalso does not work?

Fourth, having created a surplus, Stalinhad to prevent vigorous rent seeking frompoliticizing its distribution. Special branchand regional interests formed immediatelywithin the dictator’s own circle, leavingStalin and a few associates to battle for whatthey viewed as encompassing interests. Evenwith access to virtually all official docu-ments, we cannot measure the degree towhich the top leadership allocated invest-ment to purchase loyalty rather than for eco-nomic rationality. Stalin may have beenmore tolerant of rent seeking in earlier peri-ods when his regime was still fragile.Political motives may help explain the alloca-tion of regional investment and motor vehi-cles in the early 1930s, but Stalin seems tohave stamped on rent seekers when hedetected them in military matters.

Fifth, the industrial archives contribute todebate about whether there existed anythingthat we should call “planning.” All so-calledoperational plans of the Stalin era were pro-visional and subject to change by any superi-or. Faced with provisional plans and theprospect of endless interventions, enterprisesand ministries sometimes refused to committo plans; some enterprises operated without“plans” for years. Resource allocation byintervention rather than by plan is consistentwith the dictator’s aversion to formal rules. A“final” plan represents a formal rule thatcould prevent superiors from exercising“resource mobility.” The chaos of ad hocinterventions then explains the attraction of

planning from the achieved level, a custom-ary rule which introduced order at theexpense of freezing initial allocations andinhibiting adaptation.

Sixth, the dictator struck a deal with func-tional agencies, such as the State PlanningCommission, various control commissions,and, above all, the secret police that hewould not hold them responsible for finalresults if they provided honest information.But the information available to thePolitburo, although far greater in quantity,suffered from the same flaws as that whichwas published. Most information was pro-vided by the producers themselves, who hadincentives to exaggerate inputs needs and tounderstate capacity. Even the secret policehad an incentive to exaggerate dangers toenhance their claim on resources. The dicta-tor’s honest information brokers could atbest perform infrequent audits. This infor-mation flowed with difficulty because pro-cessing capacity was limited, and becauseinformation was monopolized and traded forprivate gain rather than freely shared. Goodinformation was so hard to get that planningrelied more on intuition than on the “scien-tific” methodology of Soviet handbooks.Huge blunders resulted. Where decisionshave appeared wise in retrospect, for exam-ple, in defense technology, they often reliedon information from abroad and mecha-nisms such as scientific reputation that didnot depend on state enforcement.

Seventh, the soft budget constraint originat-ed historically in the Stalinist political commit-ment to mobilization and was perpetuated bycentralized institutions that prevented ex antecommitments to financial discipline frombeing implemented after the event. Notably,subsidies were not extended automatically toloss makers without lobbying investments,which helps explain the unexpected interest ofproducers in higher prices and money, includ-ing traceable bank money. The archives pro-vide no support for the proposition that Sovietenterprises sought lower prices to maximizebribes; rather enterprises took advantage of



the seller’s markets to extract higher priceseven from influential buyers such as thedefense ministry. Corruption existed but thearchives leave the impression that most ille-gal siphoning was done to fulfill the plan, notfor private gain.

Eighth, the pervasiveness of markets inwhich even high-level administrators (suchas ministers) participated reflects the dilem-ma of the dictator who requires subordi-nates to solve their own problems “at anyprice,” including by recourse to informalmarkets, while the networks that formed asa result had the potential to undermine dic-tatorial power. These markets did not workwell, were severely circumscribed, andrelied heavily on relational enforcement;private contracts were not legally enforce-able and reputations could spread onlywithin narrow circles.

Ninth and finally, the flaws of the plannedeconomy became immediately apparent toits leaders, prompting reform proposals toincrease enterprise independence. But theresults of partial reforms were if anythingworse than those of the unreformed system.Enterprises might prefer to circumvent theapparently arbitrary, even stupid interven-tions of superiors; but the same enterprises,unconstrained by the discipline of marketforces, could not be trusted with independ-ence. The resulting stalemate continued tothe last days of the Soviet Union.

REFERENCES

Acemoglu, Daron, and James A. Robinson. 2000.“Political Losers as a Barrier to EconomicDevelopment.” American Economic Review, 90(2):126–30.

Akerlof, George A. 1984. “Gift Exchange andEfficiency-Wage Theory: Four Views.” AmericanEconomic Review, 74(2): 79–83.

Akerlof, George A., and Janet L. Yellen. 1990. “TheFair Wage–Effort Hypothesis and Unemployment.”Quarterly Journal of Economics, 105(2): 255–83.

Amann, Ronald. 1986. “Technical Progress and SovietEconomic Development: Setting the Scene,” inTechnical Progress and Soviet EconomicDevelopment. Ronald Amann and Julian Cooper,eds. Oxford: Blackwell, 5–30.

Amann, Ronald, and Julian Cooper, eds. 1982.Industrial Innovation in the Soviet Union. New

Haven: Yale University Press.Amann, Ronald, and Julian Cooper, eds. 1986.

Technical Progress and Soviet EconomicDevelopment. Oxford: Blackwell.

Amann, Ronald, Julian Cooper, and R. W. Davies, eds.1977. The Technological Level of Soviet Industry.New Haven: Yale University Press.

Ames, Edward. 1965. Soviet Economic Processes.Homewood, Ill.: Irwin.

Bardhan, Pranab K., and John E. Roemer. 1993.Market Socialism: The Current Debate. New York:Oxford University Press.

Barone, Enrico. 1908/1935. “Il Ministro dellaProduzione nello Stato Collettivista,” translated as“The Ministry of Production in the CollectivisticState,” in Collectivist Economic Planning: CriticalStudies on the Possibility of Socialism. F. A. Hayek,ed. London: Routledge, 245–90.

Becker, Gary S. 1968. “Crime and Punishment: AnEconomic Approach.” Journal of Political Economy,76(2): 169–217.

Belova, Eugenia. 2001. “Economic Crime andPunishment,” in Behind the Façade of Stalin’sCommand Economy: Evidence from the State andParty Archives. Paul R. Gregory, ed. Stanford:Hoover Institution Press, 131–58.

Belova, Eugenia. 2005. “Legal Contract Enforcementin the Soviet Economy.” Comparative EconomicStudies, 47(2): 387–401.

Belova, Eugenia, and Paul Gregory. 2002. “Dictator,Loyal, and Opportunistic Agents: The SovietArchives on Creating the Soviet Economic System.”Public Choice, 113(3–4): 265–86.

Berelovich, A., ed. 2000. Sovetskaia derevnia glazamiVChK-OGPU-NKVD, 2: 1923–1929. Moscow:ROSSPEN.

Berelovich, A., and V. P. Danilov, eds. 1998. Sovetskaiaderevnia glazami VChK-OGPU-NKVD, 1:1918–1922. Moscow: ROSSPEN.

Bergson, Abram. 1947. “A Problem in Soviet Statistics.”Review of Economics and Statistics, 29(4): 234–42.

Bergson, Abram. 1953. “Reliability and Usability ofSoviet Statistics: A Summary Appraisal.” AmericanStatistician, 7(3): 13–16.

Bergson, Abram. 1961. The Real National Income ofSoviet Russia Since 1928. Cambridge: HarvardUniversity Press.

Bergson, Abram. 1964. The Economics of SovietPlanning. New Haven: Yale University Press.

Bergson, Abram. 1966. Essays in NormativeEconomics. Cambridge: Harvard University Press.

Bergson, Abram. 1978. Productivity and the SocialSystem: The USSR and the West. Cambridge:Harvard University Press.

Bergson, Abram. 1987. “On Soviet Real InvestmentGrowth.” Soviet Studies, 39(3): 406–24.

Berliner, Joseph S. 1952. “The Informal Organizationof the Soviet Firm.” Quarterly Journal of Economics,66(3): 342–65.

Berliner, Joseph S. 1957. Factory and Manager in theUSSR. Cambridge: Harvard University Press.

Berliner, Joseph S. 1976. The Innovation Decision inSoviet Industry. Cambridge: MIT Press.



Birman, Igor. 1978. “From the Achieved Level.” SovietStudies, 30(2): 153–72.

Birman, Igor. 1981. Secret Incomes of the Soviet StateBudget. London: Nijhoff.

Blanchard, Olivier, and Michael Kremer. 1997.“Disorganization.” Quarterly Journal of Economics,112(4): 1091–1126.

Boettke, Peter J. 2001. Calculation and Coordination:Essays on Socialism and Transitional PoliticalEconomy. London: Routledge.

Bornstein, Morris. 1985. “Improving the SovietEconomic Mechanism.” Soviet Studies, 37(1): 1–30.

Borodkin, Leonid, and Simon Ertz. 2003. “Coercionversus Motivation: Forced Labor in Norilsk,” in TheEconomics of Forced Labor: The Soviet Gulag. PaulR. Gregory and Valery Lazarev, eds. Stanford:Hoover Institution Press, 75–104.

Borodkin, Leonid, and Simon Ertz. 2005. “ForcedLabor and the Need for Motivation: Wages andBonuses in the Stalinist Camp System.” ComparativeEconomic Studies, 47(2): 418–36.

Carr, Edward H., and R. W. Davies. 1969. A History ofSoviet Russia 3: Foundations of a Planned Economy,1926–1929, Part 1. London: Macmillan.

Chuev, F. I. 1991. Sto sorok besed s Molotovym.Moscow: Terra.

Chuev, F. I. 1992. Tak govoril Kaganovich. Moscow:Otechestvo.

Clark, Colin. 1939. A Critique of Russian Statistics.London: Macmillan.

Conquest, Robert. 1967. Industrial Workers in theUSSR. London: The Bodley Head.

Conquest, Robert. 1987. The Harvest of Sorrow: SovietCollectivization and the Terror-Famine. Oxford:Oxford University Press.

Danilov, Viktor P., Roberta T. Manning, and LynneViola, eds. 1999. Tragediia sovetskoi derevni.Kollektivizatsiia i raskulachivanie. 1927–1939.Dokumenty i materialy. Moscow: ROSSPEN.

Davies, R. W. 1980a. The Industrialization of SovietRussia, 1: The Socialist Offensive: TheCollectivization of Soviet Agriculture, 1929–1930.Basingstoke: Macmillan.

Davies, R. W. 1980b. The Industrialization of SovietRussia, 2: The Soviet Collective Farm, 1929–1930.Basingstoke: Macmillan.

Davies, R. W. 1989a. The Industrialization of SovietRussia, 3: The Soviet Economy in Turmoil,1929–1930. Basingstoke: Macmillan.

Davies, R. W. 1989b. Soviet History in the GorbachevRevolution. Basingstoke: Macmillan.

Davies, R. W. 1993. “Soviet Military Expenditure andthe Armaments Industry, 1929–33: AReconsideration.” Europe–Asia Studies, 45(4):577–608.

Davies, R. W. 1996. The Industrialization of SovietRussia, 4: Crisis and Progress in the Soviet Economy,1931–1933. Basingstoke: Macmillan.

Davies, R. W. 1997. Soviet History in the Yeltsin Era.Basingstoke: Macmillan.

Davies, R. W. 2001a. “Making Economic Policy,” inBehind the Façade of Stalin’s Command Economy:Evidence from the State and Party Archives. Paul R.

Gregory, ed. Stanford: Hoover Institution Press,61–80.

Davies, R. W. 2001b. “Why Was There a SovietInvestment Cycle in 1933–37?” University ofWarwick. PERSA Working Paper No. 16.

Davies, R. W. 2003. “The Archives and the StalinistEconomy.” University of Warwick. PERSA WorkingPaper No. 27.

Davies, R. W. 2004. “The Soviet Economy and theLaunching of the Great Terror.” University ofWarwick. PERSA Working Paper No. 29.

Davies, R. W., and Mark Harrison. 1997. “The SovietMilitary–Economic Effort under the Second Five-Year Plan 1933–1937.” Europe–Asia Studies, 49(3):369–406.

Davies, R. W., Melanie Ilic, and Oleg Khlevnyuk. 2004.“The Politburo and Economic Policy-Making,” inThe Nature of Stalin’s Dictatorship: The Politburo,1924–1953. E. A. Rees, ed. Basingstoke: Palgrave,108–34.

Davies, R. W., and Oleg Khlevnyuk. 1997. “Gosplan,”in Decision-Making in the Stalinist CommandEconomy, 1932–37. E. A. Rees, ed. Basingstoke andLondon: Macmillan, 32–66.

Davies, R. W., and Oleg Khlevnyuk. 1999. “The End ofRationing in the Soviet Union, 1934–1935.”Europe–Asia Studies, 51(4): 557–610.

Davies, R. W., and Oleg Khlevnyuk. 2002.“Stakhanovism and the Soviet Economy.”Europe–Asia Studies, 54(6): 867–903.

Davies, R. W., Oleg V. Khlevnyuk, E. A. Rees,Liudmilla P. Kosheleva, and Larisa A. Rogovaia, eds.2003. The Stalin–Kaganovich Correspondence,1931–36. New Haven: Yale University Press.

Davies, R. W., Mark B. Tauger, and Stephen G.Wheatcroft. 1995. “Stalin, Grain Stocks and theFamine of 1932–1933.” Slavic Review, 54(3): 642–57.

Davies, R. W. and Stephen G. Wheatcroft. 2004. TheIndustrialization of Soviet Russia, 5: The Years ofHunger: Soviet Agriculture, 1931–1933. Basingstoke:Palgrave.

Dewatripont, Mathias, and Eric Maskin. 1995. “Creditand Efficiency in Centralized and DecentralizedEconomies.” Review of Economic Studies, 62(4):541–55.

Djankov, Simeon, Edward Glaeser, Rafael la Porta,Florencio Lopez-de-Silanes, and Andrei Shleifer.2003. “The New Comparative Economics.” Journalof Comparative Economics, 31(4): 595–619.

Duhamel, Luc. 2004. “The Last Campaign AgainstCorruption in Soviet Moscow.” Europe–Asia Studies,56(2): 187–212.

Easterly, William, and Stanley Fischer. 1995. “TheSoviet Economic Decline.” World Bank EconomicReview, 9(3): 341–71.

Ellman, Michael. 1972. Soviet Planning Today.Cambridge: Cambridge University Press.

Ellman, Michael. 1975. “Did the Agricultural SurplusProvide the Resources for the Increase inInvestment in the USSR during the First Five YearPlan?” Economic Journal, 85(340): 844–63.

Ellman, Michael. 1979. Socialist Planning. Cambridge:Cambridge University Press.



Ellman, Michael. 2000. “The 1947 Soviet Famine andthe Entitlement Approach to Famines.” CambridgeJournal of Economics, 24(5): 603–30.

Ericson, Richard E. 1983. “On an Allocative Role ofthe Soviet Second Economy,” in Marxism, CentralPlanning, and the Soviet Economy: Economic Essaysin Honor of Alexander Erlich. Padma Desai, ed.Cambridge: MIT Press, 110–32.

Ericson, Richard E. 1984. “The ‘Second Economy’ andResource Allocation under Central Planning.”Journal of Comparative Economics, 8(1): 1–24.

Erlich, Alexander. 1960. The Soviet IndustrializationDebate, 1924–1928. Cambridge: Harvard UniversityPress.

Ertz, Simon, 2003. “Building Norilsk,” in TheEconomics of Forced Labor: The Soviet Gulag. PaulR. Gregory and Valary Lazarev, eds. Stanford:Hoover Institution Press, 127–50.

Ertz, Simon. 2005. “Trading Effort for Freedom:Workday Credits in the Soviet Camp System.”Comparative Economic Studies, 47(2): 476–91.

Filtzer, Don. 2002. Soviet Workers and Late Stalinism:Labour and the Restoration of the Stalinist Systemafter World War II. Cambridge: CambridgeUniversity Press.

Fitzpatrick, Sheila. 1999. Everyday Stalinism:Ordinary Life in Extraordinary Times: Soviet Russiain the 1930s. Oxford: Oxford University Press.

Ganelin, R. S., ed. 2004. Istoriia sozdaniia i razvitiiaoboronnopromyshlennogo kompleksa Rossii i SSSR.1901–1963 gg. Dokumenty i materialy, 1: Voennaiapromyshlennost’ Rossii v nachale XX veka(1901–1918 gg.). Moscow: Novyi Khronograf.

Gerschenkron, Alexander. 1947. “The Soviet Indices ofIndustrial Production.” Review of Economics andStatistics, 29(4): 217–26.

Getty, J. Arch, and Oleg Naumov. 1999. The Road toTerror: Stalin and the Self-Destruction of theBolsheviks, 1932–1939. New Haven: Yale UniversityPress.

Glaeser, Edward L., Rafael La Porta, Florencio Lopez-de-Silanes, and Andrei Shleifer. 2004. “DoInstitutions Cause Growth?” NBER Working Paper10568.

Goldfeld, Stephen M., and Richard E. Quandt. 1988.“Budget Constraints, Bailouts, and the Firm underCentral Planning.” Journal of ComparativeEconomics, 12(4): 502–20.

Goldfeld, Stephen M., and Richard E. Quandt. 1990.“Output Targets, the Soft Budget Constraint and theFirm under Central Planning.” Journal of EconomicBehavior and Organization, 14(2): 205–22.

Gomulka, Stanislaw. 1985. “Kornai’s Soft BudgetConstraint and the Shortage Phenemenon: ACriticism and Restatement.” Economics of Planning,19(1): 1–11.

Gorlizki, Yoram. 2001a. “Stalin’s Cabinet: ThePolitburo and Decision-Making in the Post-WarYears.” Europe–Asia Studies, 53(2): 291–312.

Gorlizki, Yoram. 2001b. “Theft Under Stalin: AProperty Rights Analysis.” University of Warwick.PERSA Working Paper No. 10.

Gorlizki, Yoram. 2002. “Ordinary Stalinism: The

Council of Ministers and the Soviet NeopatrimonialState, 1946–1953.” Journal of Modern History, 74(4):699–736.

Gorlizki, Yoram, and Oleg Khlevnyuk. 2004. ColdPeace: Stalin and the Soviet Ruling Circle,1945–1953. New York: Oxford University Press.

Granick, David. 1954. Management of the IndustrialFirm in the USSR. New York: Columbia UniversityPress.

Granick, David. 1987. Job Rights in the Soviet Union:Their Consequences. New York: CambridgeUniversity Press.

Gregory, Paul R. 2001. “The Dictator’s Orders,” inBehind the Façade of Stalin’s Command Economy:Evidence from the State and Party Archives. PaulGregory, ed. Stanford: Hoover Institution Press,11–33.

Gregory, Paul R. 2003a. “An Introduction to theEconomics of the Gulag,” in The Economics ofForced Labor: The Soviet Gulag. Paul R. Gregoryand Valery Lazarev, eds. Stanford: HooverInstitution Press, 1–21.

Gregory, Paul R. 2003b. “Soviet Defence Puzzles:Archives, Strategy, and Underfulfillment.” Europe–Asia Studies, 55(6): 923–38.

Gregory, Paul R. 2004. The Political Economy ofStalinism: Evidence from the Soviet Secret Archives.New York: Cambridge.

Gregory, Paul R., and Andrei Markevich. 2002.“Creating Soviet Industry: The House That StalinBuilt.” Slavic Review, 61(4): 787–814.

Gregory, Paul R., and Robert C. Stuart. 1974. SovietEconomic Structure and Performance. New York:Harper and Row.

Gregory, Paul R., and Aleksei Tikhonov. 2000. “CentralPlanning and Unintended Consequences: Creatingthe Soviet Financial System, 1930–1939.” Journal ofEconomic History, 60(4): 1017–40.

Greif, Avner. 1993. “Contract Enforceability andEconomic Institutions in Early Trade: The MaghribiTraders’ Coalition.” American Economic Review,83(3): 525–48.

Greif, Avner. 1994. “Cultural Beliefs and theOrganization of Society: A Historical and TheoreticalReflection on Collectivist and Individualist Societies.”Journal of Political Economy, 102(5): 912–50.

Greif, Avner. 2000. “The Fundamental Problem ofExchange: A Research Agenda in HistoricalInstitutional Analysis.” European Review ofEconomic History, 4(3): 251–84.

Greif, Avner, Paul Milgrom, and Barry R. Weingast.1995. “Coordination, Commitment, and Enforce-ment: The Case of the Merchant Guild.” Journal ofPolitical Economy, 102(4): 745–76.

Grossman, Gregory. 1953. “Scarce Capital and SovietDoctrine.” Quarterly Journal of Economics, 67(3):311–43.

Hanson, Philip. 1983. “Success Indicators Revisited: TheJuly 1979 Decree on Planning and Management.”Soviet Studies, 35(1): 1–13.

Hanson, Philip. 1984. “The CIA, the TsSU, and theReal Growth of Soviet Investment.” Soviet Studies,86(4): 571–81.



Hanson, Philip, and Keith Pavitt. 1987. TheComparative Economics of Research, Developmentand Innovation in East and West: A Survey. Chur,Switzerland: Harwood.

Harris, James R. 1999. The Great Urals: Regionalismand the Evolution of the Soviet System. Ithaca:Cornell University Press.

Harrison, Mark. 1994. “National Income,” in TheEconomic Transformation of the Soviet Union,1913–45. R. W. Davies, Mark Harrison, and S. G.Wheatcroft, eds. Cambridge: Cambridge UniversityPress, 38–56.

Harrison, Mark. 1998. “Prices, Planners, and Producers:An Agency Problem in Soviet Industry, 1928–1950.”Journal of Economic History, 58(4): 1032–62.

Harrison, Mark. 2000. “New Postwar Branches (1):Rocketry,” in The Soviet Defence-Industry Complexfrom Stalin to Khruschev. John Barber and MarkHarrison, eds. Basingstoke: Macmillan, 118–49.

Harrison, Mark. 2002. “Coercion, Compliance, and theCollapse of the Soviet Command Economy.”Economic History Review, 55(3): 397–433.

Harrison, Mark. 2003a. “The Political Economy of aSoviet Military R&D Failure: Steam Power forAviation, 1932 to 1939.” Journal of EconomicHistory, 63(1): 178–212.

Harrison, Mark. 2003b. “Soviet Industry and the RedArmy under Stalin: A Military–Industrial Complex?”Les Cahiers du Monde Russe, 44(2–3): 323–42.

Harrison, Mark. 2004. “Why Secrets? The Uses ofSecrecy in Stalin’s Command Economy.” Universityof Warwick. PERSA Working Paper No. 34.

Harrison, Mark. 2005. “The Fundamental Problem ofCommand: Plan and Compliance in a PartiallyCentralized Economy.” Comparative EconomicStudies, 47(2): 296–314.

Harrison, Mark. Forthcoming. “A Soviet Quasi-Marketfor Inventions: Jet Propulsion, 1932 to 1946.”Research in Economic History.

Harrison, Mark, and Byung-Yeon Kim. Forthcoming.“Plans, Prices, and Corruption: The Soviet FirmUnder Partial Centralization, 1930 to 1990.” Journalof Economic History.

Harrison, Mark, and Nikolai Simonov. 2000.“Voenpriemka: Prices, Costs, and Quality Assurancein Interwar Defence Industry,” in The SovietDefence-Industry Complex From Stalin toKhrushchev. John Barber and Mark Harrison, eds.Basingstoke: Macmillan, 223–45.

Hayek, F. A. 1937. “Economics and Knowledge.”Economica, 4(13): 33–54.

Hayek, F. A. 1940. “Socialist Calculation: TheCompetitive ‘Solution’.” Economica, 7(26): 125–49.

Hayek, F. A. 1944. The Road to Serfdom. Chicago:University of Chicago Press.

Hayek, F. A. 1945. “The Use of Knowledge in Society,”American Economic Review, 35(4): 519–30.

Heinzen, James W. 2004. Inventing a SovietCountryside: State Power and the Transformation ofRural Russia, 1917–1929. Pittsburgh: University ofPittsburgh Press.

Heinzen, James W. 2005. “Corruption in the Gulag:Dilemmas of Officials and Prisoners.” Comparative

Economic Studies, 47(2): 456–75.Hessler, Julie. 2004. A Social History of Soviet Trade:

Trade Policy, Retail Practices, and Consumption,1917–1953. Princeton: Princeton University Press.

Holloway, David. 1994. Stalin and the Bomb: TheSoviet Union and Atomic Energy, 1939–1956. NewHaven: Yale University Press.

Jasny, Naum. 1949. The Socialized Agriculture of theUSSR: Plans and Performance. Stanford: StanfordUniversity Press.

Joyce, Christopher. 2003. “The Gulag in Karelia:1929–1941,” in The Economics of Forced Labor: TheSoviet Gulag. Paul R. Gregory and Valery Lazarev,eds. Stanford: Hoover Institution Press, 163–88.

Junker, James. 1992. Socialism Revised andModernized: The Case for Pragmatic MarketSocialism. New York: Praeger.

Katz, Barbara G. 1975. “Purges and Production: SovietEconomic Growth, 1928–1940.” Journal ofEconomic History, 35(3): 567–90.

Keren, Michael. 1982. “The Ministry, Plan Changes,and the Ratchet in Planning.” Journal ofComparative Economics, 6(4): 327–42.

Khlevnyuk, Oleg V. 1993. Stalin i Ordzhonikidze.Konflikty v Politburo v 30-e gody. Moscow: Rossiiamolodaia.

Khlevnyuk, Oleg V. 1995. “The Objectives of the GreatTerror, 1937–38,” in Soviet History, 1917–1953:Essays in Honour of R. W. Davies. J. M. Cooper,Maureen Perrie, and E. A. Rees, eds. New York: St.Martin’s Press, 158–76.

Khlevnyuk, Oleg V. 1996. Politburo. Mekhanizmypolitchecheskoi vlasti v 1930-e gody. Moscow:ROSSPEN.

Khlevnyuk, Oleg V. 2000. “Sovetskaia ekonomicheskaiapolitika na rubezhe 40–50 godov i delo Gosplana.”Paper presented at a conference on “Stalin’sPolitburo,” Florence, Italy, March 29–31, 2000.

Khlevnyuk, Oleg V. 2001a. “Stalinism and the StalinPeriod after the ‘Archival Revolution’.” Kritika, 2(2):319–28.

Khlevnyuk, Oleg V. 2001b. “The Economy of theGulag,” in Behind the Façade of Stalin’s CommandEconomy: Evidence from the State and PartyArchives. Paul R. Gregory, ed. Stanford: HooverInstitution Press, 111–30.

Khlevnyuk, Oleg V. 2003a. “The Economy of theOGPU, NKVD, and MVD of the USSR, 1930–1953:The Scale, Structure, and Trends of Development,”in The Economics of Forced Labor: The SovietGulag. Paul R. Gregory and Valery Lazarev, eds.Stanford: Hoover Institution Press, 43–66.

Khlevnyuk, Oleg V. 2003b. “Sistema tsentr-regiony v1930–1950-e gody. Predposylki politizatsii nomen-klatury.” Les Cahiers du Monde Russe, 44(2–3):253–68.

Khlevnyuk, Oleg V. 2004. “Vzaimootnesheniia tsentra iregionov v 30-e gg.” Vestnik MoskovskogoUniversiteta, 21(3): 79–98.

Khlevnyuk, Oleg V., R. W. Davies, L. P. Kosheleva, E.A. Rees, and L. A. Rogovaia, eds. 2001. Stalin iKaganovich. Perepiska. 1931–1936 gg. Moscow:ROSSPEN.



Khlevnyuk, Oleg V., Yoram Gorlizki, L. P. Kosheleva, A.I. Miniuk, M. I. Prozumenshchikov, L. A. Rogovaia,and S. V. Somonova, eds. 2002. Politburo TsK VPK(b)i Sovet Ministrov SSSR. 1945–1953. Moscow:ROSSPEN.

Khlevnyuk, Oleg V., A. V. Kvashonkin, L. P. Kosheleva,and L. A. Rogovaia, eds. 1995. Stalinskoe Politburo v30-e gody. Sbornik dokumentov. Moscow: AIRO-XX.

Kim, Byung-Yeon. 1997. “Soviet Household SavingFunction.” Economics of Planning, 30(2–3): 181–203.

Kim, Byung-Yeon. 1999. “The Income, Savings, andMonetary Overhang of Soviet Households.” Journalof Comparative Economics, 27(4): 644–68.

Kim, Byung-Yeon. 2002. “Causes of Repressed Inflationin the Soviet Consumer Market, 1965–1989: RetailPrice Subsidies, the Siphoning Effect, and the BudgetDeficit.” Economic History Review, 55(1): 105–27.

Kontorovich, Vladimir. 1986. “Discipline and Growthin the Soviet Economy.” Problems of Communism,34(6): 18–31.

Kontorovich, Vladimir. 1988. “Lessons of the 1965Soviet Economic Reform.” Soviet Studies, 40(2):308–16.

Kornai, János. 1980. The Economics of Shortage.Amsterdam: North-Holland.

Kornai, János. 1992. The Socialist System: The PoliticalEconomy of Communism. Princeton: PrincetonUniversity Press.

Kornai, János, Eric Maskin, and Gerard Roland. 2003.“Understanding the Soft Budget Constraint.”Journal of Economic Literature, 41(4): 1095–1136.

Kozlov, V. P., ed. 2004. Istoriia Stalinskogo Gulaga.Moscow: ROSSPEN.

Kroll, Heidi. 1986. “Decentralization and PrecontractDisputes in Soviet Industry.” Soviet Economy, 2(1):51–71.

Kroll, Heidi. 1988. “The Role of Contracts in the SovietEconomy.” Soviet Studies, 40(3): 349–66.

Kuromiya, Hiroaki. 1988. Stalin’s IndustrialRevolution: Politics and Workers, 1928–1932.Cambridge: Cambridge University Press.

Kuznets, Simon. 1963. “A Comparative Appraisal,” inEconomic Trends in the Soviet Union. AbramBergson and Simon Kuznets, eds. Cambridge:Harvard University Press, 333–82.

Lange, Oskar. 1964. On the Economic Theory ofSocialism. New York: McGraw-Hill.

Lavoie, Don. 1985. Rivalry and Central Planning: TheSocialist Calculation Debate Reconsidered.Cambridge: Cambridge University Press.

Lazarev, Valery. 2005. “Promotion Incentives andSupport for the Soviet Regime.” ComparativeEconomic Studies, 47(2): 346–63.

Lazarev, Valery, and Paul R. Gregory. 2002. “TheWheels of a Command Economy: Allocating SovietVehicles.” Economic History Review, 55(2): 324–48.

Lazarev, Valery, and Paul R. Gregory. 2003.“Commissars and Cars: A Case Study in the PoliticalEconomy of Dictatorship.” Journal of ComparativeEconomics, 31(1): 1–19.

Lebina, Nataliia. 2000. “The Defence-IndustryComplex in Leningrad (2): The Postwar UraniumIndustry,” in The Soviet Defence-Industry Complex

from Stalin to Khruschev. John Barber and MarkHarrison, eds. Basingstoke: Macmillan, 184–94.

Lewin, Moshe. 1968. Russian Peasants and SovietPower: A Study of Collectivization. London: Allenand Unwin.

Lewin, Moshe. 2003. “Rebuilding the SovietNomenklaturas, 1945–1948.” Les Cahiers du MondeRusse, 44(2–3): 219–52.

Lih, Lars, Oleg Naumov, and Oleg Khlevnyuk. 1995.Stalin’s Letters to Molotov, 1925–1936. New Haven:Yale University Press.

Manning, Roberta T. 1993. “The Soviet EconomicCrisis of 1936–1940 and the Great Purges,” inStalinist Terror: New Perspectives. J. Arch Getty andRoberta T. Manning, eds. Cambridge: CambridgeUniversity Press, 116–41.

Manove, Michael. 1971. “A Model of Soviet-TypeEconomic Planning.” American Economic Review,61(3): 390–406.

Markevich, Andrei. 2003. “Was the Soviet EconomyPlanned? Planning in the People’s Commissariats inthe 1930s.” University of Warwick. PERSA WorkingPaper No. 25.

Markevich, Andrei. 2005. “Soviet Planning Archives:The Files that Bergson Could Not See.”Comparative Economic Studies, 47(2): 364–86.

Markevich, Andrei, and Mark Harrison. Forthcoming.“Quality, Experience, and Monopoly: The SovietMarket for Weapons under Stalin.” EconomicHistory Review.

Maskin, Eric S. 1996. “Theories of the Soft Budget-Constraint.” Japan and the World Economy, 8(2):125–33.

Maskin, Eric S. 1999. “Recent Theoretical Work on theSoft Budget Constraint.” American EconomicReview, 89(2): 421–25.

Millar, James R. 1974. “Mass Collectivization and theContribution of Soviet Agriculture to the First Five-Year Plan: A Review Article.” Slavic Review, 33(4):750–66.

von Mises, Ludwig. 1920/1935. “Die Wirtschaftsrechnungim sozialistischen Gemeinwesen,” translated inCollectivist Economic Planning: Critical Studies on thePossibility of Socialism. F. A. Hayek, ed. London:Routledge, 87–130.

von Mises, Ludwig. 1949/1998. Human Action: ATreatise on Economics. Auburn, Ala.: Ludwig vonMises Institute.

Montefiore, Simon Sebag. 2003. Stalin: The Court ofthe Red Tsar. London: Weidenfeld and Nicolson.

Morukov, Mikhail. 2003. “The White Sea–BalticCanal,” in The Economics of Forced Labor: TheSoviet Gulag. Paul R. Gregory and Valery Lazarev,eds. Stanford: Hoover Institution Press, 151–62.

Nordlander, David. 2003. “Magadan and the EconomicHistory of Dalstroi in the 1930s,” in The Economicsof Forced Labor: The Soviet Gulag. Paul R. Gregoryand Valery Lazarev, eds. Stanford: HooverInstitution Press, 105–25.

Nove, Alec. 1958. “The Problem of ‘Success Indicators’in Soviet Industry.” Economica, 25(97): 1–13.

Nove, Alec. 1961. The Soviet Economy. London: Allenand Unwin.



Nove, Alec. 1969. An Economic History of the USSR.London: Allen and Unwin.

Nove, Alec. 1977. The Soviet Economic System.London: Allen and Unwin.

Ofer, Gur. 1987. “Soviet Economic Growth:1928–1985.” Journal of Economic Literature 25(4):1767–1833.

Olson, Mancur. 1993. “Dictatorship, Democracy, andDevelopment.” American Political Science Review,87(3): 567–76.

Osokina, Elena. 2001. Our Daily Bread: SocialistDistribution and the Art of Survival in Stalin’sRussia, 1927–1941. Armonk, NY: M. E. Sharpe.

Podkaminer, Leon. 1989. “Theory of Investment Cyclesin the Centrally Planned Economies: An EconomicInterpretation.” Forschungberichte no. 158. WienerInstitut für Internationaler Wirtschaftsvergleiche.

Rees, E. A., ed. 1997a. Decision-Making in the StalinistEconomy, 1932–37. Basingstoke: Macmillan.

Rees, E. A. 1997b. “The People’s Commissariat of theTimber Industry,” in Decision-Making in theStalinist Economy, 1932–37. E. A. Rees, ed.Basingstoke: Macmillan, 124–49.

Rees, E. A. 2001. “Leaders and their Institutions,” inBehind the Façade of Stalin’s Command Economy:Evidence from the State and Party Archives. Paul R.Gregory, ed. Stanford: Hoover Institution Press,35–60.

Rees, E. A., ed. 2004. The Nature of Stalin’sDictatorship: The Politburo, 1924–1953. Basing-stoke: Palgrave.

Rees, E. A., and D. H. Watson. 1997. “Politburo andSovnarkom,” in Decision-Making in the StalinistCommand Economy, 1932–37. E. A. Rees, ed.Basingstoke: Macmillan, 9–31.

Rittersporn, Gábor T. 2001. “New Horizons:Conceptualizing the Soviet 1930s.” Kritika, 2(2):307–18.

Rutland, Peter. 1985. The Myth of the Plan. LaSalle,IL: Open Court.

Samuelson, Lennart. 2000. Plans for Stalin’s WarMachine: Tukhachevskii and Military–EconomicPlanning, 1925–41. London: Macmillan.

Schaffer, Mark E. 1989. “The Credible-CommitmentProblem in the Center–Enterprise Relationship.”Journal of Comparative Economics, 13(3): 359–82.

Schelling, Thomas C. 1960. The Strategy of Conflict.Oxford: Oxford University Press.

Schrettl, Wolfram. 1982. Consumption, Effort, andGrowth in Soviet-Type Economies: A TheoreticalAnalysis. Ann Arbor: University MicrofilmsInternational.

Schrettl, Wolfram. 1984. “Anspruchdenken,Leistungbereitschaft und Wirtschaftszyklen,” inWachstumsverlangsamung und Konjunkturzyklen inUnterschiedlichen Wirtschaftssystemen. A. Schuller,ed. Berlin: Duncker and Humblot, 153–91.

Schroeder, Gertrude E. 1973. “Recent Developmentsin Soviet Planning and Incentives,” in SovietEconomic Prospects for the Seventies. U.S. CongressJoint Economic Committee. Washington, D.C.: U.S.Government Printing Office, 11–38.

Schroeder, Gertrude E. 1979. “The Soviet Economy on

a Treadmill of ‘Reforms’,” in Soviet Economy in aTime of Change. U.S. Congress Joint EconomicCommittee. Washington, D.C.: U.S. GovernmentPrinting Office, 312–40.

Schroeder, Gertrude E. 1982. “Soviet Economic‘Reform’ Decrees: More Steps on the Treadmill,” inSoviet Economy in the 1980s: Problems andProspects. U.S. Congress Joint EconomicCommittee. Washington, D.C.: U.S. GovernmentPrinting Office, 65–88.

Schroeder, Gertrude E. 1985. “The Slowdown in SovietIndustry, 1976–1982.” Soviet Economy, 1(1): 42–74.

Sen, Amartya K. 1981. Poverty and Famines: An Essayon Entitlement and Deprivation. Oxford: OxfordUniversity Press.

Shleifer, Andrei, and Robert W. Vishny. 1992.“Pervasive Shortages under Socialism.” RANDJournal of Economics, 23(2): 237–46.

Shleifer, Andrei, and Robert W. Vishny. 1994. “ThePolitics of Market Socialism.” Journal of EconomicPerspectives, 8(2): 165–76.

Simonov, Nikolai S. 2000. “New Postwar Branches (2):The Nuclear Industry,” in The Soviet Defence-Industry Complex from Stalin to Khrushchev. JohnBarber and Mark Harrison eds. Basingstoke:Macmillan, 150–72.

Sokolov, Andrei K. 2003. “Forced Labor in SovietIndustry: The End of the 1930s to the Mid-1950s: AnOverview,” in The Economics of Forced Labor: TheSoviet Gulag. Paul R. Gregory and Valery Lazarev,eds. Stanford: Hoover Institution Press, 23–42.

Sokolov, Andrei K. 2005. “Before Stalinism: TheDefense Industry of Soviet Russia in the 1920s.”Comparative Economic Studies, 47(2): 437–55.

Spulber, Nicolas. 1964. Soviet Strategy for EconomicGrowth. Bloomington, IN: Indiana University Press.

Stalin, I. V. 1937. Voprosy Leninizma. Moscow:Gospolitizdat.

Strauss, Erich. 1969. Soviet Agriculture in Perspective:A Study of its Successes and Failures. London: Allenand Unwin.

Sutton, Antony C. 1968. Western Technology andSoviet Economic Development, 1: 1917–1930.Stanford: Hoover Institution Press.



Swianiewicz, S. 1965. Forced Labour and EconomicDevelopment: An Enquiry into the Experience ofSoviet Industrialization. London: Oxford UniversityPress.

Tauger, Mark B. 1991. “The 1932 Harvest and the SovietFamine of 1932–1933.” Slavic Review, 50(1): 70–89.

Tikhonov, Aleksei. 2003. “The End of the Gulag,” inThe Economics of Forced Labor: The Soviet Gulag.Paul R. Gregory and Valery Lazarev, eds. Stanford:Hoover Institution Press, 67–74.

United States Department of Commerce. 1960.Statistical Abstract of the United States. Washington,D.C.: U.S. Government Printing Office.



Vatlin, A. I. 2004. Terror raionnogo masshtaba.“Massovye operatsii” NKVD v Kuntsevskom raioneMoskovskoi oblasti. 1937–1938 gg. Moscow:ROSSPEN.

Weitzman, Martin L. 1980. “The ‘Ratchet Principle’and Performance Incentives.” Bell Journal ofEconomics, 11(1):302–08.

Wheatcroft, Stephen G. 2004. “From Team Stalin toDegenerate Tyranny,” in The Nature of Stalin’sDictatorship: The Politburo, 1924–1953. E. A. Rees,ed. Basingstoke: Palgrave, 79–107.

Wheatcroft, Stephen G., and R. W. Davies, eds. 1985.Materials for a Balance of the Soviet NationalEconomy, 1928–1930. Cambridge: CambridgeUniversity Press.

Wheatcroft, Stephen G., and R. W. Davies. 2002. “TheSoviet Famine of 1932–3 and the Crisis inAgriculture,” in Challenging Traditional Views of

Russian History. S. G. Wheatcroft, ed. Basingstoke:Palgrave, 112–46.

Wiles, Peter J. D. 1982. “The Soviet InvestmentDeflator Again.” Soviet Studies, 34(4): 616–17.

Wintrobe, Ronald. 1998. The Political Economy ofDictatorship. Cambridge: Cambridge UniversityPress.

Zaleski, Eugène. 1971. Planning for Economic Growthin the Soviet Union, 1918–1932. Chapel Hill:University of North Carolina Press.

Zaleski, Eugène. 1980. Stalinist Planning for EconomicGrowth, 1933–1952. Chapel Hill: University ofNorth Carolina Press.

Zolotarev, V. A., ed. Forthcoming. Istoriia sozdaniia irazvitiia oboronno-promyshlennogo kompleksaRossii i SSSR. 1901–1963 gg. Dokumenty i materi-aly, 2: Sovetskoe voenno-promyshlennoe proizvodst-vo (1918–1926). Moscow: Novyi Khronograt.

Documents

Allocation under Dictatorship: Research in Stalin's Archives · Allocation under Dictatorship: Research in Stalin’s Archives Paul Gregory and Mark Harrison∗ We survey recent research