nanshin: budget-maximizing behavior, the imperial japanese navy and the origins of the pacific war

Public Organization Review: A Global Journal 4: 135–155 (2004)# 2004 Kluwer Academic Publishers. Manufactured in The Netherlands.

Nanshin: Budget-Maximizing Behavior,the Imperial Japanese Navy and the Originsof the Pacific War

BRIAN DOLLERY [email protected]

School of Economics, University of New England, Armidale, NSW 2351, Australia

ZANE SPINDLERDepartment of Economics, Simon Fraser University, Vancouver, BC, Canada

CRAIG PARSONS

Faculty of Economics, Yokohama National University, Yokohama, Japan

Key words: budget maximization, Japanese Imperial Army, Japanese Imperial Navy, Nanshin,Pacific War, public choice

Abstract

Inter-service rivalry over budget allocations between the Japanese Imperial Navy and the ImperialArmy played a crucial role in the genesis of World War Two in the Pacific. The adoption of a nanshin

(‘‘southward advance’’) strategy by the Navy may be explained as an attempt to maximize its budget

leading directly to the attack on Pearl Harbor in 1941. To date, this argument has been presented in

the form of historical narrative without any explanatory theoretical framework. The present paperseeks to place inter-service budgetary rivalry within the context of public choice theory to enhance

understanding of this historical perspective.

Introduction

The origins of World War II in the Pacific are complex, multi-faceted and still notwell understood by scholars. Indeed, Dockrill (1994) has observed recently that‘‘historians could not even agree on when the war started or what it should becalled’’. The analysis of the causes of the Pacific War began in earnest with thejudicial hearings of the International Military Tribunal for the Far East, morecommonly known as the Tokyo Trials, over the period 1946 to 1948, andcontinues to the present time. Early efforts at explaining the origins of the conflictwere severely hampered by the classification of a good deal of the officialJapanese and allied documentation, including crucial material gathered throughthe magic code-breaking system (Department of Defense, 1978). Contemporarywork remains handicapped by translation difficulties and significant errors in thetranscription of much of the decoded magic documentation (Komatsu, 1999).

With some exceptions (Moriyama, 1990), historians have produced avoluminous narrative literature largely devoid of theoretical frameworks asorganizing and explanatory tools. This literature presently has two main streams.On the one hand, a number of scholars have argued that the outbreak of thePacific War was not simply an inevitable consequence of the expansion of the‘‘China Incident’’, but rather the result of complex power relations betweenAmerica, Great Britain, Italy, Japan and the Third Reich in a global context,including the war in Europe (Hatano, 1991). On the other hand, another (mostlyJapanese) school of thought has contended that the pre-war decision-makinginstitutional structure in Japan played a significant role in the events leading up tothe attack on Pearl Harbor on December 7, 1941. In particular, ‘‘the increasinginfluence of the military’s power resulted in a lack of national control, rather thanthe establishment of a totalitarian control system, because the military’s poweritself had eventually fallen apart into various divisions’’ (Komatsu, 1999). Thispaper falls squarely within the latter camp.

The significance of venomous inter-service rivalry between the ImperialJapanese Navy and the Imperial Japanese Army over the size of their annualbudgetary allocations from the central government has been recognized as a keycausal factor in the Pacific War by a number of historians.1 Moreover, thisappears to have been explicitly acknowledged by Imperial Navy officers at thetime. For example, in 1934 when asked whether the Navy contemplated war withthe United States Navy, Admiral Suetsugu Nobumasa replied: ‘‘Certainly, eventhis is acceptable if it will get us a budget’’ (Barnhart, 1987). Similarly, shortly aftera key ministerial conference on January 19, 1939, Navy Captain Takagi Sokichiassured delegates that they should not erroneously presume that ‘‘the Navy,although prepared to use Britain and the United States as pretexts for a budget,actually did not want to confront them’’ (Asada, 1973).

The implications of inter-service rivalry are perhaps at their most stark in thecontext of the competing strategies developed by the Imperial Navy and theImperial Army prior to the Pacific War. Nanshu hokushin (defense in the south andadvance in the north) became the official doctrine of the Army, known ashokushin-ron (school of thought for northward advance). Similarly, hokushunanshin (defense in the north and advance in the south), termed nanshin-ron(school of thought for southward advance) became the standard naval strategy(Frei, 1991). In essence, through its nanshin-ron policy, the Navy hoped to seizethe resource-rich British Burma, Malaya and Singapore, the Dutch East Indies,French Indochina and the American Philippines, thereby easing the drasticshortages of essential materials in Japan contingent upon the Americaneconomic boycott. It understood that this meant war with both the BritishEmpire and the United States. However, it was hoped that this conflict would beshort lived. The adoption of a hokushu nanshin strategy meant that the Navycould lay legitimate claim to the vast budgetary resources required to build amodern ‘‘blue-water’’ fleet of sufficient strength to defeat the Royal Navy andU.S. Fleet.

136 B. DOLLERY ET AL.

By contrast, the Army sought a ‘‘holding strategy’’ in the Pacific, with conquestlimited to the Netherlands East Indies (and possibly British Malaya), thus avertingwar with America, providing the necessary access to strategic resources, andallowing for a nanshu hokushin campaign aimed at the defeat of China and aneventual attack on the Soviet Union. This would ensure that the preponderance ofbudgetary funds would flow to the Army since the Navy would not need a largerfleet. Needless to add, the Imperial Navy hotly contested this plan ‘‘because theAnglo-Dutch naval presence was only token, the Navy saw no role for itself in anyadvance limited to those countries’ possessions’’ and thus ‘‘the fleet would notbe positioned to demand the materials allocations for the completion of currentbuilding plans, much less the initiation of new ones’’ (Barnhart, 1987).

Budgetary competition between the Imperial Army and the Imperial Navy wasfurther complicated by the existence of other public bureaucracies and importantinterest groups in Japanese industry. Notwithstanding the steadily increasingplethora of regulation, ‘‘Japan’s economy was by no means governmentcontrolled’’ (Barnhart, 1987), despite the existence of a national Planning Board.Industrial cartels represented by ‘‘control associations’’ created under the aegisof the 1930 Major Industries Control Law exerted significant influence on publicdecision making. However, the fact that the armed services could appoint theirown ministerial representatives to the Imperial Cabinet and jeopardize thesurvival of a government by threatening to withdraw their respective ministersgave them far more political power than their counterparts in other publicbureaucracies (Akira, 1987). Moreover, since steel production was overwhel-mingly nationalized, the critical annual question of steel quotas was usuallyresolved in favor of the armed services to the cost of private industry.Accordingly, competition for scarce resources from civilian public bureaucraciesand organized industry weakened during the 1930s, particularly after theescalation of armed conflict following the China Incident in 1937. This is reflectedin dwindling budgets and falling steel allocations (Barnhart, 1987).

Despite the discovery of the importance of budgetary rivalry in Japan as acritical factor in precipitating and shaping the Pacific War, economists and othersocial scientists have ignored this dimension of the Second World War. This isunfortunate since some of the theoretical approaches developed by economistsappear eminently suited to an analysis of budget-maximizing behavior, not leastthe literature derived from the seminal work of Niskanen (1971).2 Moreover, anexamination of budgetary rivalry between the two Japanese armed services is ofinterest to public choice economists in its own right. Analyses of budget-maximizing behavior typically focus on the allocative inefficiencies deriving fromthis conduct and the social costs they impose on the country in question (Mueller,1989). However, in the context of the Pacific War, these costs were not onlyincalculably higher for Japan, but were also catastrophic for many othercountries.

The paper itself is divided into four main parts. Section 2 provides a briefsynoptic review of the historical importance of inter-service budgetary rivalry in

NANSHIN: BUDGET-MAXIMIZING BEHAVIOR 137

Japan’s fatal decision to attack the United States and Great Britain. Section 3presents a theoretical perspective by first introducing Niskanen’s bureaucratictheory and then developing alternative (and unique) extensions to his theoreticalapparatus, which can serve as a conceptual basis for the assessment andanalysis of Imperial Army and Navy behavior. Section 4 then applies ourtheoretical perspective to the historical evidence of Japan’s government andbureaucratic behavior and the consequent Pacific War. The paper ends withsome brief concluding remarks in Section 5.

Historical background

Although Japanese seafarers and settlers had a long history of engagement inthe south-west Pacific, modern interest in this region was ignited through thepublicity generated by naval training cruises through the area that began in 1875(Frei, 1991). In particular, the populist geographical writings of Hattori Toru,Suganuma Teifu and Shiga Shigetaka seem to have aroused a tremendousinterest in an almost mystical nan’yo, or South Seas, abundant in majesticislands and unlimited natural resources. In the popular imagination ‘‘it was awarm tropical paradise, a territory in which to gain personal achievements andfulfill a sense of adventure’’ and to nascent expansionists the nan’yo represented‘‘the one area untouched by Western imperialists and thus the optimal place forthe new nation of Japan to acquire territories’’ (Schencking, 1999). ShrewdJapanese naval propagandists were to turn this popular current to theiradvantage.

The Imperial Japanese Navy was established as a separate and independentforce in 1872 with the founding of its own service ministry. However, with thecreation of an Army General Staff in 1878, and no naval equivalent, the Navy‘‘began to slip into a subordinate role, a position it occupied until the beginning ofthe 1890s’’ (Evans and Peattie, 1997). An analogous Naval General Staff was onlyfounded in 1893. Nevertheless, the pivotal role played by the Imperial Navy in theSino-Japanese War (1894–1895) and the Russo-Japanese War (1904–1905),where it destroyed both enemy fleets in decisive battles, brought home topoliticians and the public alike the importance of the Navy for the defense of thehome islands and the development of a Japanese empire. Moreover, ‘‘warproved immensely profitable for the navy’’ (Schencking, 1998): whereas in 1890the total naval budget represented only 64% of the Army’s expenditure, by 1905this had leapt to 127% of the Army’s outlays (Ono, 1922).

The 1907 Imperial Defense Conference represented another significantmilestone in the Navy’s rivalry with the Army. Although this conference soughtto remedy bitter inter-service antagonism and develop a unified imperial defensestrategy, it generated the ironical outcome that each service could define its ownpotential opponents. Inter-service rivalry thus emerged as the major beneficiary.It also served to formalize the earlier rather inchoate and nebulous positions of


the two services. From henceforth nanshu hokushin, also known as hokushinron,became the official doctrine of the Army. Similarly, hokushu nanshin, sometimestermed nanshin-ron became the standard naval strategy. It is also evident thatnanshin-ron had become ‘‘a bureaucratic tool through which the navy hoped togain a larger share of military appropriations’’ (Schencking, 1998).

The budgetary impetus afforded the Navy by the 1907 Conference wascomparatively short lived. Longstanding suspicions about corruption in navalarmament acquisitions exploded in the so-called ‘‘Siemens Navy scandal’’, withdamaging revelations of bribery, theft and blackmail involving naval officers(Evans, 1978). This lead directly to the fall of the pro-Navy Yamamoto Cabinetand the rescinding of the record naval budgetary allocation intended for the 1914fiscal year. It was thus most fortuitous for naval leaders that the First World Warintervened.

Under the 1902 Anglo-Japanese Alliance, Japan declared war on Germany.Even before the onset of hostilities the Imperial Navy had already benefited whenthe Diet approved an extraordinary allocation to build ten destroyers for useagainst the German East Asiatic Squadron. Despite explicit orders from NavyMinister Yashiro Rokuro against aggressive conduct towards German interests inthe Pacific, in favor of the acquisition of German territories in China, opportunisticnaval commanders succeeded in capturing German Micronesia north of theequator by October 1914. Separate garrisons were rapidly consolidated under aunified and Navy-controlled Provisional South Seas Defence Force that enabledthe Imperial Navy to garner further public funds to administer the newpossessions. But the most important outcome for the Navy from World War Ilay in the fact that it could now provide concrete justification for the need for alarge and powerful ‘‘blue-water’’ fleet. Moreover, operational plans were revisedto make Micronesian islands an integral ‘‘component of its strategy to defeat itschief hypothetical enemy, the U.S. Navy’’ (Schencking, 1998). Nanshin had thusbecome the cornerstone of the Navy’s budgetary struggle with its Armyadversaries for public funds. Indeed, so successful had been the Imperial Navystrategy during the war that by 1921 it had a budget almost twice the size of theArmy.

In the aftermath of World War I, all three great naval powers, Great Britain,Japan and the United States, faced ambitious, and potentially ruinous, navalconstruction programs. The Washington Treaty (1921–1922) sought to thwart anarms race and set tonnage and other limitations on its eight signatory states, witha 6 to 10 ratio of capital ships between Japan and the two Anglo-Saxon powers.The immediate effect of this much-hated Treaty for the Imperial Navy was theimmediate cancellation of its grandiose plans for expansion to an eight-battleshipfleet. Instead it had to be satisfied with six capital ships and then focussedinstead on the development of smaller vessels, marine aviation and submarines.It is possible that the great Tokyo earthquake and fire of 1923 and the severeeconomic depression later in the same decade might have spiked theexpansionary plans of the Navy even in the absence of the treaty.


The 1930 disarmament treaty signed in London had an even more dramaticeffect on the fortunes of the Navy. This agreement dealt with lighter auxiliarynaval vessels, a matter not covered in the Washington Treaty, and set the ratio at6.975 for Japan and 10 for the other two great naval powers. A divisive andemotional debate ensued in Japan between the ‘‘treaty faction’’, whichsupported the treaty, and the ‘‘fleet faction’’, that opposed the Londonagreement. Incensed by the perceived ‘‘internationalism’’ of the government,extremists assassinated Prime Minister Hamaguchi thereby launching anongoing (and often violent) campaign by military hardliners against civilian andservice moderates (Akira, 1987).

The ascendancy of the ‘‘fleet faction’’ from the mid-1930s meant that a nanshinpolicy now dominated the naval agenda. The successful conquest of Manchuriaby the Imperial Army in 1932, and the attendant public acclaim it enjoyed,provided a spur to advocates of nanshin-ron in the Navy. Moreover, abandon-ment of cooperation with the Anglo-American naval powers and the subsequentuncertainties surrounding a reliable source of fuel oil for the Imperial Navy addedfurther urgency to the need for a ‘‘southward advance’’ towards the oil-richDutch East Indies. Similarly, the outbreak of the China War in the summer of1937, which soon spread from north to central China, not only plunged Japan intofull-scale hostilities, but also meant ‘‘the navy was loathe to let the armymonopolize the government funds and public support generated by the conflict’’(Peattie, 1996). As a first step towards the realization of its nanshin plans, theImperial Navy thus autonomously seized a number of islands off the coast ofChina, including Hainan, which had strategic potential in any move southwards.In sum, ‘‘by 1939, therefore, the ‘southward advance’ and the expansion of theJapanese fleet (and the resources needed for that expansion) had attained analmost symbiotic relationship in the minds of the navy’s aggressive middleechelon’’ (Peattie, 1996).

Fundamental differences still existed between the Army and the Navy on thenature of any nanshin strategy. In line with its hokushin-ron doctrine, the ImperialArmy envisaged that any military operations towards the south would be directedagainst the Dutch East Indies, or in the worst-case scenario, against British andDutch territories. Given the paucity of Anglo-Dutch naval forces in the Pacificfrom 1940, it believed that this would enable Japan to secure access to oil andother vital resources without recourse to war with the United States and itspowerful Pacific Fleet. The Army would thus be assured of sufficient budgetaryallocations to be able to pursue its ongoing campaign to defeat China and thentackle the Soviet threat north of Manchuria. But this conception of Imperialstrategy held little promise for the Navy. If Japan could conquer the NetherlandsEast Indies without the intervention of the British Empire and the American Fleet,then the navy would become an under-funded observer of great land battles incontinental Asia without legitimate claims on national budgetary allocations.Accordingly, after simulated staff exercises involving an attack on the Dutch EastIndies, the Naval General Staff concluded that since the United States was


already providing substantial military assistance to Britain in its war withGermany, it would probably intervene alongside Britain in the Pacific. The‘‘indivisibility’’ of Britain and the United States thus became a cornerstone of any‘‘southward advance’’ on the part of naval strategists and formed the basis of thedecision to launch pre-emptive attacks on the American Philippines and PearlHarbor.

These differences were finally settled by compromise. The Navy agreed tosupport the Army’s plans for the war in China and Tripartite Pact with Germanyand Italy that would minimize the threat of Soviet intervention in Manchuria. Theprice it exacted was Army support for its nanshin strategy and an immediate anddrastic increase in its budget. The German invasion of the Soviet Unionheightened concerns in the Navy that the Imperial Army might use this as apretext for invading Russia. Rapid American military rearmament and theimposition of further U.S. economic boycotts complicated naval calculations.Barnhart (1987) has described the Navy’s dilemma as follows: ‘‘The longer the(southward) advance could be delayed, the more warships could be constructedfor the imperial fleet and the more steel could be requisitioned to build thosewarships’’, but ‘‘too much delay, however, would be disastrous, because it wouldgive the Americans time to accomplish their own colossal naval expansionprogram’’. Moreover, in the light of its successful budgetary claims, it couldhardly argue that it would not be successful against Anglo-American naval powerin the Pacific. One method of resolving this dilemma was to embark on a surpriseattack upon the U.S. Pacific Fleet in Hawaii and thereby even up the odds in thesubsequent naval war. The scene was thus set for the fateful attack on PearlHarbor and the resulting conflagration in the Pacific.

A theoretical perspective

The public choice approach to the budgetary conflict between the Imperial Armyand the Imperial Navy involves modeling each as an optimizing entity. Theinnovator for this form of bureaucratic analysis was Niskanen (1971). WhileNiskanen’s work has been both extended and criticized,3 his initial modelingoffers an interesting starting point for understanding how bureaucratic budgetmaximization might drive the competition and cooperation between the ImperialArmy and Navy, and ultimately affect the nature of the Pacific War. Niskanen alsoprovided both a formal mathematical model (with arithmetic examples in hisoriginal book) and a graphical analysis, which opened bureaucratic analysis tothe interest and understanding of economists, who previously had largely ignoredthis field of study.

In this paper, we will apply only the less formal, graphical analysis, which drawson Niskanen’s (1971) methodology to explore his insights on competing bureausas well as perspectives of others economic theorists, such as Congleton (1980)and Faith (1980), on cooperating bureaus. Niskanen defined bureau’s as non-


profit organizations providing an output (Q) to a sponsor in return for a periodicgrant or budget (B). Bureaucrats are defined as employees who cannot directlyappropriate any bureaucratic surplus (an excess of budget B over cost C),although their perquisites of office (salary, power, prestige, etc.) are generally amonotonic function of budget size, giving them an inherent interest in, and anidentification with, budget size. Those bureaucrats, who positively value suchperquisites, will have an obvious incentive to maximize budget size rather thanbudget surplus. (Indeed, standard government budgeting procedures, whichinvolve the loss of an unspent budget at the fiscal year’s end, tend to discouragebudget surplus maximization.)

Curiously, in the construction of his formal model, Niskanen adopted a ratherpeculiar assumption from the standpoint of many bureaucratic observers, thoughnot necessarily economists, in that he presumed a ‘‘hard’’ budget constraint:4

That is, the bureau’s total budget must at least cover its total costs (B > or¼C).Furthermore, Niskanen assumed that both the budget (B) obtainable from thesponsor and the cost (C) of bureaucratic production are direct, but different,functions of output (Q); i.e., B ¼ B(Q) and C ¼ C(Q). In what follows, it will beconvenient to display these functions in a manner familiar to economists andothers who employ demand and supply analysis; namely, in their average or per-unit form. That is, we will graph per unit curves derived by dividing the functionsfor total budget and total cost by quantity (i.e., B/Q ¼ B(Q)/Q; C/Q ¼ C(Q)/Q).

Figure 1 illustrates the various outcomes of budget maximization. First, asponsor who values the output of a bureau sufficiently to provide that bureau witha budget for a given output would be willing to pay a different per unit ‘‘price’’ fordifferent levels of bureau output. This is shown by curve B, which is the sponsor’smarginal budget curve, the bureaucratic analog of a product demand curve.Given the objective of budget maximization (rather than surplus maximization), abureau would supply output along its average cost curve C (rather than along itsmarginal cost curve), the bureaucratic analog of a product supply curve. The waythis curve is drawn in Figure 1, at every per-unit budget-price, a budget-maximizing bureau would supply twice the output that a profit or surplusmaximizing firm or bureau would supply.

In contrast, a profit-maximizing firm’s (or surplus-maximizing bureau’s) supplycurve would be given by its marginal (not average) cost curve. (For graphicalsimplicity, the relevant marginal cost curve is not drawn in Figure 1, but, if drawn,would bisect the distance between the vertical axis and the average cost curveC.) This unique modeling of bureaucratic overproduction, and, hence, allocativeinefficiency, was Niskanen’s key insight and analytical contribution whichlaunched an extensive literature on the economics and political economy ofbureaucracy. Even the severest critics of Niskanen’s modeling concede that theircounter-modeling suggests bureau’s will tend to be technically inefficient even ifand when they are allocatively efficient (since any surplus maximized must bewasted by turning it into an excess budgetary cost). Put differently, whetherbureaucrats are utility maximizers, surplus maximizers or budget maximizers,


they nevertheless tend to overcharge and/or overproduce. It is this keycharacteristic that makes the political economy of bureaucracy relevant forunderstanding the processes of bureau rivalry, including expansion and evenpossible collapse.

Given these essential elements of budget-maximizing bureau analysis, we cannow explore the outcome under various circumstances of the bureaucraticsetting, such as whether a bureaucratic situation is one of monopoly, competingduopoly, or cooperative duopoly and whether the sponsor’s review processresults in bureaux being ‘‘budget-takers’’ or ‘‘budget-makers’’.5 This 362

elementary classification scheme might yield six distinct cases, except that thecase of a budget-making, competing duopoly is analytically contradictory; to bejointly budget-making, requires cooperation, not competition, at least notcompetition vis-a-vis the sponsor. Thus, there are only five distinct cases toanalyze here: (1) Budget-taking monopoly, (2) budget-making monopoly,(3) budget-taking, competing duopoly, (4) budget-taking, cooperative duopoly,and (5) budget-making, cooperating duopoly. Consideration of these five casesforms the basis for a comparison of bureaucratic forms that is novel to this paper.

Figure 1. Comparative bureau budget maximization.


1. Budget-taking monopoly bureau: In this case, the sponsor sets the per-unitbudget: This might occur if sponsor officials cannot directly appropriate anysurplus arising from socially optimal bureaucratic production, but onlyindirectly benefit from their constituents generally enjoying some ‘‘consumersurplus’’ from government provision, and if sponsor officials have somealternative source of information about bureaucratic production costs. Inthese circumstances, it may be likely that sponsor officials have an incentiveto get the best ‘‘deal’’ for their constituents. That would involve not onlyclosely monitoring the bureau to insure supply at minimum cost, but alsochoosing an output level that leaves their constituents with some ‘‘consumersurplus’’ from government provision. Given a per-unit budget allocation, abureau will offer an output determined by its average cost curve C. Thesponsor and bureau coordinate their budget and supply offers at the point yin Figure 1, where curves B and C intersect. Output is at 0b and averagebudget is at 0k; Total budget equals total cost as shown by area 0byk,which measures both. While budget-maximizing bureau output is twice ashigh (and the total budget is higher) as would occur with a profit-maximizingfirm or a surplus-maximizing bureau, this case still represents one of the moreefficient extremes of the bureaucratic form and provides a benchmark forwhat follows.

2. Budget-making monopoly bureau: This case represents the prototypicalNiskanen bureau, which is able to control the supply of information about itsproduction costs, and/or which has a sponsor whose officials (and/or theirspecial-interest constituents) benefit indirectly (directly) from the level ofbureaucratic output (rather than from the surplus generated by bureaucraticproduction), such that the bureau can make an ‘‘all or none’’ offer to thesponsor (a total budget for a total output), thereby extracting all sponsor and/or constituent surplus. Now the bureau can regard the sponsor’s averagebudget curve (B0) as showing the maximum amount that the sponsor will payfor any level of output when it is offered on an ‘‘all or none’’ or ‘‘take it or leaveit’’ basis. The budget-making, budget-maximizing bureau maximizes itsbudget where its minimum average cost curve C intersects its maximumaverage budget curve B at point z in Figure 1. Output is at 0d and averagebudget is at 0j; Total budget and total cost are equal and both measured byarea 0bzj. Figure 1 illustrates that this bureau, as in Case 1, is ‘‘technicallyefficient’’ because it is producing at minimum cost,6 but it is more ‘‘allocativelyinefficient’’ because it is producing too much output—again, approximatelydouble the amount produced by a private firm facing similar demand and costconditions. Sponsor (or constituent) surplus is captured and expended onbureaucratic expansion far beyond social optimality. The main net bene-ficiaries of bureaucratic supply may be the bureaucrats themselves, andwhatever other of the bureau’s factor suppliers who may be compensated inexcess of their minimum supply price. That is, there may only be a factorsurplus.7


3. Budget-taking, competing duopoly bureau: Actual competitive bureaucraticsupply may involve any number of bureaus offering identical bureaucraticproducts. However, here we assume, for simplicity of exposition and fordevelopment of an appropriate theoretical analog, that there are only twobureaus with identical costs of production C, which, when summed, yieldcurve C0. Competition would likely reveal to the sponsor the true cost ofbureaucratic supply, so, as in Case (1), the sponsor could choose to purchasefrom either bureau on a per unit basis. Thus, the intersection of sponsor’smarginal budget curve B with the bureau’s joint supply curve C0 at point x0

would give the equilibrium total output 0d, equally divided between eachbureau as 0a and ac,8 and total budget (equals total cost) area 0cx0m, equallydivided between each bureau as 0axm and 0acm, respectively. Again, bothbureau’s are technically efficient individually and jointly, but allocativelyinefficient jointly, although less so than in the previous case of the budget-making monopoly bureau because both total output and total budget arelower in this case.9

4. Budget-taking, cooperative duopoly bureau: Instead of competing directly bysupplying identical bureaucratic ‘‘products’’ that are essentially perfectsubstitutes, bureau’s may compete indirectly for budget allocations by tacitlyor explicitly supplying bureaucratic products that are imperfect substitutes oreven complements to the sponsor and/or the sponsor’s constituents. If thereis only tacit or implicit cooperation where bureaus spontaneously developtheir individual product niches, the sponsor may still have independentsources of information that allow budget allocation along the sponsor’smarginal, rather than average, budget curve. However, with imperfectsubstitutes or complements, the sponsor’s marginal budget curve for bothbureaucratic ‘‘products’’ would shift out. We can show that in Figure 1 byassuming that both bureau’s produce complementary products that areequally valued by the sponsor. Now each bureau’s output is valued the samealong the same marginal budget curve B. When we combine these marginalbudget curves by adding them horizontally to get the sponsor’s totaldemand, B0 is now the relevant curve whose intersection with the bureaus’joint supply curve at point y0 determines the equilibrium output 0e, averagebudget ‘‘price’’ 0k, and total budget (¼ total cost) 0ey0k. The two bureaus’individual outputs and budget shares are 0b and be, and 0byk and bey0y,respectively. The efficiency of this outcome is similar in nature if not extent tothat of Case (1); these bureaux are technically efficient but allocativelyinefficient relative to profit maximizing private duopolies without price-discriminating power.

5. Budget-making, cooperating duopoly: Finally, consider two bureaus that notonly consciously offer complementary bureaucratic ‘‘products’’, but also(perhaps with the sponsor’s tacit or explicit approval) collude to extract jointlythe maximum possible budget subject only to the total budget equals totalcost constraint. Again, this strategy involves an equilibrium position


determined by the intersection of the bureaux joint supply function C0 with thesponsor’s average budget curve for the bureaus joint, complementary,outputs, shown in Figure 1 as B00, at point z0. The two bureau’s individualoutputs and budget shares are 0d and dg, and 0dzj and dgz0z, respectively.Total output 0g, average budget 0j, and total budget (¼ total cost) 0gz0j are allhigher than when bureaux are unable to pursue an ‘‘all or none’’ budget-maximization strategy with their sponsor. The budget-constrained outcomehere is again technically efficient but allocatively inefficient (and to a muchgreater extent than previous cases) relative to the social optimum and relativeto profit maximizing private duopolies with price-discriminating power.10

Again, as in Case 2, only the bureaucrats, other factor suppliers, and,possibly, high-demand constituents11 are net beneficiaries.

The perspective applied

The theoretical perspective developed above can be used to analyze the inter-service rivalry over budget allocations between the Imperial Navy and theImperial Army. Either the Navy or the Army might have offered the typical‘‘product’’ of any military organization, which, when arrayed on a continuum,could be conceived as stretching from zero (no defense) through defense toconquest of varying levels, where the measurement metric might be area ofterritory defended (conquered). In the case of an island nation like Japan, andgiven the nature of army and navy production technicalities being either land orwater based, it is clear that it would not be rational for either of these bureaus tooffer identical products. That is, there is a natural product stratificationdetermined by both demand-side and supply-side realities.12

Thus, while there might be a rivalry, even intense, venomous, and deadlyrivalry, over budget allocations, the Imperial Army and Navy would not likely becharacterized as ‘‘competing bureaus’’ as defined and analyzed above. Rather,they would more naturally be ‘‘cooperating bureaus’’ in terms of their productsbeing imperfect substitutes or direct complements. (Indeed, as observed inSection 2, there was an Imperial Army and Imperial Navy compromise overmutually supporting both nanshin and hokushin.) That symbiosis would alsomake it likely for their bureaucrats to realize jointly the greater potential forbudget maximization in not only designing their products to maximizecomplementarities, but also to collude to extract the maximum budget allocationfrom the Japanese government.13 Indeed, with their own ministerial representa-tives within the Imperial Cabinet, it would be possible for the Navy and/or Armyto collude directly with their ‘‘sponsor’’, from which, for practical purposes, theymay have become indistinguishable as their bureaucratic growth, and the PacificWar, progressed. For each bureau, it is as simple as x, y, and z. That is, assimple as designing bureau products and cooperating to move individually (and


totally) from outcomes like those at point x(x)0 to outcomes at point z(z0) inFigure 1.

Of course, the analysis demonstrated with Figure 1 presumes equal-sizedbureaus whereas, in reality, relative size would still be a variable determined byinter-service competition over the explicit nature of each service’s products.Thus, while the Imperial Navy would complement the role of the Imperial Army inits campaign on the Asian mainland (hokushin), it would gain in relative size bydeveloping a plan and competency in pursuing a Pacific campaign (nanshin)where the Imperial Army would necessarily play a more subsidiary role. In termsof Figure 1, this can be interpreted as attempting to change the ratio of 0d/dg.The Imperial Army had an incentive to cooperate to some extent in the ImperialNavy’s ‘‘product development’’ (nanshin) if the size of the budget available toboth services was expanded thereby. In terms of Figure 1, cooperation would beforthcoming it was expected that B00 would shift out so that 0g would increase, asa result.

These ideas can be shown in Figure 2, where the Navy’s ‘‘product innovation’’,nanshin, shifts out the Navy’s demand and cost curves, both absolutely andrelative to the Army’s budget and cost curves, which are now relabeled,

Figure 2. The expected nanshin effect.


respectively, BNavy, CNavy, BArmy, and CArmy.14 Now BNavy and BArmy sum to BN

00

and CNavy and CArmy sum to C0n, while the new equilibrium point shifts to z00, with a

higher ‘‘output’’ (a larger empire conquered and defended) g0, a larger totalbudget for the armed services 0g0z00j, and a larger (absolutely and relatively) Navybudget 0gz0j.15

The joint effects of ‘‘cooperative competition’’ between nanshu hokushinand hokushu nanshin ideologies can be illustrated by arranging available data(given in the Appendix) to show in Figure 3 the rapid growth of both the ImperialNavy’s and the Imperial Army’s budgets throughout the mid-1930s to early1940s.

We might note from the ‘‘Military as a % of Overall Budget’’ column in theAppendix data that these strategies enabled both military bureaucracies to outcompete other government bureaucracies for a share of the total budget.Meanwhile, the government sector was dominating the private sector in its shareof the total economy.16 Imperial wars obviously well serve the interests of thosefactor owners employed in the projection of Imperial force; at least up to the pointwhere that force is met by a superior counterforce and/or otherwise absolutelydiminishing returns, and then collapse ensues.

The relative effects of ‘‘rivalrous competition’’ between nanshu hokushin andhokushu nanshin ideologies can be illustrated by arranging the data in Figure 3 toshow the Imperial Navy’s budget relative to the Imperial Army’s budget aspresented in Figure 4.

Here we can see that the Navy’s budget share, while growing in comparison toother non-military bureaucracies and the economy, was falling relative to theArmy for a few years before the Pearl Harbor attack at the end of 1941,continuing into the 1942 budget year. However, note that these figures do notinclude temporary wartime allocations, which may have reversed the Navy’sshare much sooner, indeed, even to finance the Pearl Harbor attack. Only in the1943 budget year, with the oncoming U.S. response to Pearl Harbor, was thenanshin strategy successful in earning the Navy a larger relative share. Thisbudgetary success continued into 1944 before being ended by Japan’ssurrender in 1945.

Figure 3. Army and navy budgets, 1890–1944.


In the short run, Navy bureaucrats may have considered nanshin a veryeffective bureaucratic strategy. Within this strategy, the attack on Pearl Harborwas a tactical move to reverse the prior relative dominance of hokushin and toenhance and extend nanshin’s short-run effect. Had the United States’ ownmilitary bureaucracies not responded in a similar budget-maximizing fashion,Pearl Harbor (and nanshin) might also have been a successful long-run tactic(and strategy). Whether nanshin would have had long-run viability without thePearl Harbor tactic is, no doubt, open to endless argument. However, from apublic choice perspective, Pearl Harbor may have provided a key ingredient forspecial interest budget competition by the U.S. military, in general, and the U.S.Navy, in particular—namely, a coalescing, general interest ideology for Pacificexpansion and domination.

In summary, with the resulting Pacific War, the share of both Imperial armedservices expanded not only in total, but also as a share of the total governmentbudget and of the Japanese economy. Thus, nanshin would have been a verysuccessful bureaucratic strategy, though not necessarily the most successfulnational strategy, if only it had been sustainable. Unfortunately for the long-runviability of both Japanese bureaucracies, another nation’s military bureaucracieswere pursuing similar strategies in a world-wide bureaucratic competitionprovoked by Japan’s inter-service rivalry within its own unique institutionalenvironment.

Figure 4. Navy budget as share of army budget.


Concluding remarks

This paper does not contend that inter-service budgetary rivalry between theJapanese Imperial Navy and the Imperial Army represented the only, or even themost important, cause of World War Two in the Pacific. Many factorsundoubtedly played a contributory role (Akira, 1987). Nevertheless, MichaelBarnhart (1987) and many other historians have demonstrated that there issubstantial evidence to support the proposition that budget-maximizing behavioron the part of the Army and Navy not only helped precipitate the onset of thePacific War, but also decisively shaped its course. For instance, if the Navy hadnot adopted a nanshin strategic philosophy as its chief means of contesting theArmy’s claims to budget allocations, then in counter-factual terms it isconceivable that the pre-war oil and resource crisis could have been resolvedthrough the conquest of the Dutch East Indies alone. This would not havenecessitated an attack on Pearl Harbor and thus may not have involved conflictwith the United States. The nature and outcome of World War Two couldtherefore have been very different.

Given the significance of budgetary competition between the two branches ofthe Japanese armed forces, and given that previously this competition has beenexamined predominantly by scholars using historical narrative, it seems that theanalytical apparatus of public choice theory, specifically bureaucratic theory, canprovide additional explanatory power in understanding this dimension of theorigins of the Pacific War.


Appendix

Tab

le1.

Vario

us

bud

gets

,allo

catio

ns,

and

overa

llo

fJap

an

(1890–1944).

Fis

cal

year

Arm

y(A

)N

avy

(N)

Aþ

NT

em

po

rary

mili

tary

allo

catio

ns

(T)

Genera

l

overa

ll

bud

get

(G)

Oth

er

transfe

rs(O

)*N

as

%

of

A

Mili

tary

as

%

of

overa

ll

bud

get**

GN

E**

*G

enera

l

bud

get

as

%o

fG

NE

1890

15,5

33

10,1

59

25,6

92

082,1

25

0152.9

031.3

1,0

56,0

00

7.7

81891

14,1

80

9502

23,6

82

083,5

56

0149.2

328.3

1,1

39,0

00

7.3

4

1892

14,6

35

9133

23,7

68

076,7

35

0160.2

431

1,1

25,0

00

6.8

2

1893

14,7

21

8101

22,8

22

084,5

82

0181.7

227

1,1

97,0

00

7.0

71894

10,4

09

10,2

53

20,6

62

107,1

70

185,2

99

0101.5

269

1,3

38,0

00

13.8

5

1895

10,0

16

13,5

20

23,5

37

93,3

05

178,6

22

074.0

865.4

1,5

52,0

00

11.5

1

1896

53,2

43

20,0

06

73,2

48

0168,8

57

0266.1

443.4

1,6

66,0

00

10.1

4

1897

60,1

48

50,3

95

110,5

43

0223,6

79

0119.3

549.4

1,9

57,0

00

11.4

31898

53,8

98

58,5

30

112,4

28

0219,7

58

092.0

951.2

2,1

94,0

00

10.0

2

1899

52,5

51

61,6

62

114,2

13

0254,1

66

085.2

244.9

2,3

14,0

00

10.9

8

1900

74,8

38

58,2

75

133,1

13

0292,7

50

0128.4

245.5

2,4

14,0

00

12.1

3

1901

58,3

82

43,9

79

102,3

61

0266,8

57

0132.7

538.4

2,4

84,0

00

10.7

41902

49,4

42

36,3

26

85,7

68

0289,2

27

0136.1

129.7

2,5

37,0

00

11.4

0

1903

46,8

85

36,1

18

83,0

02

66,3

73

315,9

69

0129.8

147.3

2,6

96,0

00

11.7

2

1904

12,0

88

20,6

13

32,7

01

608,5

93

822,2

18

63,4

30

58.6

478

3,0

28,0

00

27.1

51905

11,1

09

23,4

12

34,5

21

586,1

96

887,9

37

119,0

00

47.4

569.9

3,0

84,0

00

28.7

9

1906

67,8

70

61,8

77

129,7

46

247,3

11

711,5

87

0109.6

953

3,3

02,0

00

21.5

5

1907

126,0

44

72,2

72

198,3

16

0602,4

01

0174.4

032.9

3,7

43,0

00

16.0

9

1908

141,8

05

71,5

78

213,3

84

0636,3

61

0198.1

133.5

3,7

66,0

00

16.9

01909

106,1

66

71,0

46

177,2

12

0532,8

94

0149.4

333.3

3,7

80,0

00

14.1

0

1910

101,3

24

83,8

41

185,1

64

0569,1

54

0120.8

532.5

3,9

25,0

00

14.5

0

1911

105,0

00

100,4

64

205,4

64

0585,3

74

0104.5

235.1

4,4

63,0

00

13.1

2

1912

104,1

25

95,4

85

199,6

11

0593,5

96

0109.0

533.6

4,7

74,0

00

12.4

31913

95,4

40

96,4

46

191,8

86

0573,6

34

098.9

633.5

5,0

13,0

00

11.4

4

1914

87,7

00

83,2

60

170,9

60

45,9

36

617,9

94

76,3

63

105.3

335.1

4,7

38,0

00

13.0

4

1915

97,7

91

84,3

77

182,1

68

30,5

27

595,4

50

18,3

48

115.9

035.7

4,9

91,0

00

11.9

31916

94,8

13

116,6

25

211,4

38

23,6

97

598,5

25

15,9

67

81.3

039.3

6,1

48,0

00

9.7

4

1917

123,4

37

162,4

35

285,8

72

34,8

00

731,0

07

38,8

17

75.9

943.9

8,5

92,0

00

8.5

1

1918

152,0

82

215,9

03

367,9

85

125,7

71

1,0

27,3

82

115,4

24

70.4

448.1

11,8

39,0

00

8.6

8

1919

220,2

68

316,4

19

536,6

87

149,0

30

1,3

19,3

58

069.6

151.8

15,4

53,0

00

8.5

4


Tab

le1.

Co

ntinued

.

Fis

cal

year

Arm

y(A

)N

avy

(N)

Aþ

NT

em

po

rary

mili

tary

allo

catio

ns

(T)

Genera

l

overa

ll

bud

get

(G)

Oth

er

transfe

rs(O

)*N

as

%

of

A

Mili

tary

as

%

of

overa

ll

bud

get**

GN

E**

*G

enera

l

bud

get

as

%o

fG

NE

1920

246,5

57

403,2

02

649,7

59

219,8

77

1,5

49,1

67

30,6

87

61.1

556.1

15,8

96,0

00

9.7

5

1921

246,9

79

483,5

90

730,5

68

108,7

47

1,5

98,6

03

051.0

752.5

14,8

86,0

00

10.7

4

1922

230,9

09

373,8

92

604,8

01

85,4

94

1,5

15,1

83

061.7

645.6

15,5

73,0

00

9.7

31923

223,9

27

275,1

44

499,0

71

28,4

62

1,5

49,5

13

081.3

934

14,9

24,0

00

10.3

8

1924

206,7

35

248,4

58

455,1

93

29,4

90

1,6

44,5

14

10,0

00

83.2

129.5

15,5

76,0

00

10.5

6

1925

214,8

05

229,0

03

443,8

08

1831

1,5

26,8

19

093.8

029.2

16,2

65,0

00

9.3

91926

196,9

41

237,3

08

434,2

49

01,5

78,8

26

082.9

927.5

15,9

75,0

00

9.8

8

1927

218,1

04

273,5

36

491,6

40

01,7

65,7

23

079.7

427.8

16,2

93,0

00

10.8

4

1928

249,1

06

268,1

31

517,2

38

01,8

14,8

55

092.9

028.5

16,5

06,0

00

11.0

0

1929

227,2

55

267,6

65

494,9

20

01,7

36,3

17

084.9

028.5

16,2

86,0

00

10.6

61930

200,8

24

242,0

35

442,8

59

01,5

57,8

64

082.9

728.4

14,6

71,0

00

10.6

2

1931

227,4

88

227,1

29

454,6

17

01,4

76,8

75

0100.1

630.8

13,3

09,0

00

11.1

0

1932

373,5

75

312,8

09

686,3

85

01,9

50,1

41

0119.4

335.2

13,6

60,0

00

14.2

8

1933

462,6

45

409,9

75

872,6

20

02,2

54,6

62

0112.8

538.7

15,3

47,0

00

14.6

91934

458,5

29

483,3

53

941,8

82

02,1

63,0

04

094.8

643.5

16,9

66,0

00

12.7

5

1935

496,5

59

536,3

78

1,0

32,9

37

02,2

06,4

78

092.5

846.8

18,2

98,0

00

12.0

6

1936

510,7

19

567,4

51

1,0

78,1

70

02,2

82,1

76

090.0

047.2

19,3

24,0

00

11.8

11937

591,4

75

645,3

65

1,2

36,8

40

2,0

34,2

98

4,7

42,3

20

1136

91.6

569

22,8

23,0

00

20.7

8

1938

487,5

00

679,2

46

1,1

66,7

46

4,7

95,3

95

7,7

66,2

59

317,1

65

71.7

776.8

26,3

94,0

00

29.4

2

1939

825,0

76

803,5

35

1,6

28,6

10

4,8

44,2

96

8,8

02,9

43

535,1

87

102.6

873.5

31,2

30,0

00

28.1

9

1940

1,1

92,4

70

1,0

33,7

12

2,2

26,1

82

5,7

22,5

42

10,9

82,7

55

600,0

00

115.3

672.4

36,8

51,0

00

29.8

01941

1,5

15,2

50

1,4

97,3

75

3,0

12,6

25

9,4

87,0

23

16,5

42,8

32

1,0

78,0

83

101.1

975.6

n/a

n/a

1942

56,4

54

22,6

17

79,0

71

18,7

53,1

50

24,4

06,3

82

2,6

23,2

44

249.6

177.2

n/a

n/a

1943

678

1139

1816

29,8

18,4

52

38,0

01,0

15

4,3

69,2

50

59.5

378.5

n/a

n/a

1944

728

1145

1874

73,4

93,5

54

86,1

59,8

61

7,2

05,6

42

63.5

885.3

n/a

n/a

So

urc

e:

Okura

sho

(Min

istr

yo

fF

inance),

‘‘O

kura

sho

-shi,

Vo

lum

e2’’

for

all

series

excep

tG

NE

estim

ate

s.

*T

ransfe

rsfr

om

Genera

lto

oth

er

(no

n-m

ilita

ry)

Tem

po

rary

acco

unts

.

**(A

þN

)d

ivid

ed

by

(G�

Oþ

T)

and

then

co

nvert

ed

into

ap

erc

enta

ge

as

isd

one

inth

eo

rig

inalso

urc

e.

***

Gro

ss

Natio

nalE

xp

end

iture

estim

ate

sb

yO

hkaw

a,

Takam

ats

u,

Yam

am

oto

in‘‘H

isto

ricalS

tatistics

of

Jap

an,’’

Vo

l.3,

Jap

an

Sta

tisticalA

sso

cia

tio

n.

Oth

er

no

tes:

units

are

in1000s

yen.

Genera

lm

ilita

ryb

ud

get

sho

uld

be

the

sum

of

‘‘A

rmy’’

and

‘‘N

avy’’

but

so

metim

es

diffe

rsd

ue

toro

und

ing

.


Notes

1. See, especially, Barnhart (1987), Frei (1991), and Schencking (1998, 1999).

2. Specifically, see Migue and Belanger (1974), Niskanen (1975, 1994), Orzechowski (1977), and

Jackson (1982) for variations on the budget-maximizing theme.3. Migue and Belanger (1974), Niskanen (1975, 1994), and Orzechowski (1977) advanced well-cited

critical extensions, Clark (1997), Duncombe et al. (1997), Fedeli (1999), and Wyckoff (1990)

provided insightful modeling of bureaucratic micro decision-making, while Wintrobe (1998) also

modeled the macro aspects of bureaucracies in authoritarian regimes.4. That is there is no possibility of a ‘‘soft budget constraint’’ made famous by Janos Kornai’s

observations and analysis of Soviet style budgeting.

5. The terms budget-taking and budget-making are directly analogous to the terms price-taking and

price-making which some economists use to designate the absence or presence, respectively, ofthe power to determine all of the terms of exchange in the private market place.

6. The case illustrated is called ‘‘budget constrained’’ by Niskanen because the bureau’s binding

constraint is actually the B ¼ C condition. The case not illustrated here is the one Niskanen called

‘‘demand constrained’’ which occurs if C intersects B0 at a point beyond where the marginalbudget curve B intersects the horizontal axis (point f ). That is the point where the sponsor’s

budget outlay, and, hence, the bureau’s budget, is maximized and beyond which a bureau would

expend costs while losing budget. At point f total budget exceeds minimum total cost (B>C).Such a demand-constrained bureau would have a ‘‘surplus’’—or ‘‘fat’’ as Niskanen calls it—

which it can use to pursue bureaucrats’ own objectives unrelated to the output desired by the

sponsor or by the sponsor’s constituents. Its reported ‘‘costs’’ would equal its budget but would

be higher than the minimum cost necessary to produce the promised output. Hence, a demand-constrained bureau would be both technically and allocatively inefficient. We do not need to

explore this case here as a way to expose the nature of imperial inter-service rivalry. Indeed, it

might be argued that such rivalry might constrain each service to a ‘‘budget-constrained’’, rather

than a ‘‘demand-constrained’’ outcome.7. In majoritarian governments, a high-demand constituency may prevail over a low-demand

constituency to remain a net positive beneficiary while the latter becomes a net negative

‘‘beneficiary’’ (or sufferer) from bureaucratic production (Niskanen 1971, ch. 14).8. If the bureaux do not have identical cost functions, the lower (higher) cost bureau would produce a

larger (smaller) proportion of the total output.

9. Niskanen (1971) concluded for his constant cost example that competition would not solve the

problem of bureaucratic oversupply, although it might improve technical efficiency in the demand-constrained case. In our increasing cost example, there is a slight improvement in allocative

efficiency from reduced oversupply. Niskanen did not consider the subsequent cases of

cooperative bureaux; their graphical exposition is unique to this paper.

10. See Fedeli (1999) for a more complex analysis of competing cooperating bureaucracies.11. See footnote 6.

12. After some point, defending an island requires a navy as well as army; conquest of other

landmasses also requires both services. However, given the relative cost advantage of sea versusland (rail and road) transport during the 1930s and of conquering/defending geographical areas

with lesser land densities, it is conceivable that nanshin would have been seen to dominate

hokushin on the basis of standard benefit/cost criteria (given static assumptions about a British

and U.S. response). We will leave this speculation about relative costs of alternative productiontechnologies for other investigators.

13. Budget-maximizing bureaus can further expand what is available by interacting directly with their

sponsor’s constituents to increase the latter’s demand for specific bureaucratic services as well

as for an overall increase in the size of the government’s budget. Military bureaus can pursue suchbudget expansion by forming allegiances with their suppliers (‘‘The Military-Industrial Complex’’

lobby) and their customers (patriots, nationalists, and imperialists).


14. For graphical simplicity and ease of comparison, we assume that BNavy and CNavy have simplyshifted to be identical with the former B0 and C0, respectively.

15. The Army’s budget could have been larger (smaller) as well if the Navy’s cost curve had shifted

out by less (more) than its budget curve. Since hokushin was also pursued, as agreed between theNavy and Army, the Army’s budget actually increased as well (not shown in Figure 2), though not

to the same extent as the Navy’s budget given that nanshin was more effectively pursued than

hokushin.

16. Unfortunately, we do not have complete data for GNE after 1940, but the government’s share ofthe economy was increasing up to those war dates, so it may be a safe supposition that trend

would continue until final defeat.

References

Akira, I. (1987). The Origins of the Second World War. New York: Longman.

Asada, S. (1973). ‘‘The Role of the Japanese Navy.’’ In D. Borg and O. Shumpei (eds.), Pearl Harbor as

History. New York: Columbia University Press.

Barnhart, M.A. (1987). Japan Prepares for Total War: The Search for Economic Security, 1919–1941.Ithica: Cornell University Press.

Breton, A., and R. Wintrobe. (1986). ‘‘The Bureaucracy of Murder Revisited.’’ Journal of Political

Economy 94, 905–926.Casas-Pardo, J., and M. Puchades-Navarro. (2001). ‘‘A Critical Comment on Niskanen’s Model.’’

Public Choice 107, 147–167.

Clark, D.J. (1997). ‘‘Pressure and the Division of the Public Budget.’’ Public Choice 93, 179–195.

Congleton, R. (1980). ‘‘Competitive Process, Competitive Waste, and Institutions.’’ In J.M. Buchananet al. (eds.), Toward a Theory of the Rent-Seeking Society. College Station: Texas A&M University

Press.

Department of Defense. (1978). The ‘‘Magic’’ Background of Pearl Harbor. Washington: United States

Department of Defense.Dockrill, S. (1994). From Pearl Harbor to Hiroshima: The Second World War in Asia and the Pacific,

1941–45. Basingstoke: Macmillan.

Duncombe, W., J. Miner, and J. Ruggiero. (1997). ‘‘Empirical Evaluation of Bureaucratic Models ofInefficiency.’’ Public Choice 93, 1–18.

Evans, D.C. (1978). The Satsuma Faction and Professionalism in the Japanese Naval Officer Corps of

the Meiji Period, 1868–1912. Unpublished Ph.D. Thesis, Standford University.

Evans, D.C., and M.R. Peattie. (1997). Kaigun: Strategy, Tactics and Technology in the ImperialJapanese Navy, 1887–1941. Annapolis: Naval Press Institute.

Faith, R.L. (1980). ‘‘Rent-Seeking Aspects of Bureaucratic Competition.’’ In J.M. Buchanan et al.

(eds.), Toward a Theory of the Rent-Seeking Society. College Station: Texas A&M University Press.

Fedeli, S. (1999). ‘‘Competing Bureaus and Politicians: A Compliance Approach to the Diversion ofPublic Funds.’’ Public Choice 100, 253–270.

Frei, H.P. (1991). Japan’s Southward Advance and Australia: From the Sixteenth Century to World

War II. Honolulu: University of Hawaii Press.

Hatano, S. (1989). ‘‘From the Sino-Japanese War to the Japanese-Anglo-American War.’’International Politics 91, 1–18.

Jackson, P.M. (1998). The Political Economy of Bureaucracy. Oxford: Philip Allan.

Katz, E., S. Nitzan, and J. Rosenberg. (1990). ‘‘Rent-Seeking for Pure Public Goods.’’ Public Choice65, 49–60.

Komatsu, K. (1999). The Origins of the Pacific War and the Importance of Magic. Richmond: Origi

Japan Library.

Migue, J.L., and G. Belanger. (1974). ‘‘Toward a General Theory of Management Discretion.’’ PublicChoice 17, 27–43.


Moriyama, A. (1990). ‘‘Evasion of Decision-Making: The Diplomatic Policy of the Second and ThirdKonoe Cabinets.’’ In World War II: Toward Pearl Harbor. Tokyo: Kinsei-sha.

Mueller, D.C. (1989). Public Choice II. Cambridge: Cambridge University Press.

Niskanen, W.A. (1971). Bureaucracy and Representative Government. Chicago: Aldine Atherton.Niskanen, W.A. (1975). ‘‘Bureaucrats and Politicians.’’ Journal of Law and Economics 18, 617–643.

Niskanen, W.A. (1994). Bureaucracy and Public Economics. Brookfield: Edward Elgar.

Ono, G. (1922). War and Armament Expenditures of Japan. New York: Oxford University Press.

Orzechowski, W. (1977). ‘‘Economic Models of Bureaucracy: Survey and Evidence.’’ In T.E.Borcherding (ed.), Budgets and Bureaucrats: The Sources of Government Growth. Durham: Duke

University Press.

Peattie, M.R. (1996). ‘‘Nanshin: The Southward Advance, 1931–1941, as a Prelude to the Japanese

Occupation of Southeast Asia.’’ In P. Duus, R.H. Myers and M.R. Peattie (eds.), The JapaneseWartime Empire, 1931–1941. Princeton: Princeton University Press.

Schencking, J.C. (1998). ‘‘Bureaucratic Politics, Military Budgets and Japan’s Southern Advance: The

Imperial Navy’s Seizure of German Micronesia in the First World War.’’ War in History 5, 308–326.Schencking, J.C. (1999). ‘‘The Imperial Japanese Navy and the Constructed Consciousness of a

South Seas Destiny, 1872–1921.’’ Modern Asian Studies 33, 769–796.

Wintrobe, R. (1998). The Political Economy of Dictatorship. Cambridge: Cambridge University Press.

Wyckoff, P.G. (1990). ‘‘The Simple Model of Slack-Maximizing Bureaucracy.’’ Public Choice 67, 35–47.

Brian Dollery is Professor of Economics at the University of New England in Armidale, New South

Wales, Australia. His research interests fall broadly in the area of the economics of government andmore specifically the economics of local government.

Zane Spindler is Professor of Economics in the Department of Economics at Simon Fraser University

in Burnaby, British Columbia, Canada. His research endeavors focus on public choice theorygenerally, and constitutional political economy in particular.

Craig Parsons is an Associate Professor in the Faculty of Economics at Yokohama NationalUniversity in Yokohama, Japan. His research agenda deals with international trade and the political

economy of international trade.


nanshin: budget-maximizing behavior, the imperial japanese navy and the origins of the pacific war

Documents