Abstract
The notion that professional, efficient and non-corrupt bureaucracies foster economic growth is virtually uncontested. In spite of this wide consensus, central questions remain unanswered. Thus, while the harmful effects of dysfunctional administrations are extensively covered in the theoretical literature, little is known about the empirical relevance and the expected costs of insufficient administrative rationalization. And while efficient bureaucracies are considered a key ingredient to institutional performance, the existing research rarely investigates how desirable administrative structures have been implemented in history or which concrete policy measures constitute feasible reform strategies for present-day development countries. The present paper therefore aims at providing empirical evidence to dose this lacuna; to do so, it relies on the case of administrative reforms in the last three decades of the nineteenth century in Meiji Japan. Building on an exceptionally detailed set of official statistics and documentary sources, it constructs a panel of 45 Japanese prefectures and assesses the impact of heterogeneous reform implementation on canonical indicators of economic performance including measures of regional GDP, business activity and financial market development. The central results of the econometric analysis are that delayed administrative rationalization came along with a statistically significant and robust penalty on all development indicators. Moreover, this effect was remarkably persistent over time, as the data show that late-reforming prefectures performed systematically worse than the administrative forerunners until well into the twentieth century.
This is a preview of subscription content, log in via an institution to check access.
Similar content being viewed by others
Notes
Hokkaido and Okinawa were excluded because even before the Restoration, their structural and cultural background differed substantially from the other Japanese regions.
Among the Meiji leaders, there were divisions of opinion whether to put more emphasis on economic or on military development goals. Still, the need to improve administrative performance was less contested as it was beyond doubt that military and economic strength relied on the government’s ability to ensure stable sources of revenue and to assert its monopoly of power. Moreover, there is evidence that transferring competencies to a strong central government headed by representatives from the domains which had supported the anti-shôgunate coalition was conceived as a means to mitigate cleavages between the powerful Restoration domains, which in turn reinforced the case for administrative restructuring. For further sources and references see: Ramseyer and Rosenbluth (1998), McLaren (1916), DeVere Brown/Hirota (ed.), The Diary of KidoTakayoshi (1983), Jansen and Rozman (1986).
Specialization and national market integration where even more directly impeded by shôgunate strategies to prevent alliances of opposing lords (e.g., by deliberately neglecting the maintenance of certain transit and trading routes) and by numerous daimyo who took recourse to measures like the implementation of domain monopolies in the supply of regional products to increase the economic leverage of their fiefs.
This pattern was admittedly thwarted toward the end of the Tokugawa period when the need to counter the increasingly palpable threat of Western aggression created demand for specialist services in the shôgunate’s and the daimyos’ bureaucracies. Yet, advancement in public service careers continued to be a matter of the putative office holder's birth rank and personal ties to the feudal lord rather than his individual ability or achievement (Izumi (2001), p. 15).
Thus, a samurai official with a stipend of 100 hyo and a yakudaka of 150 hyo would be compensated only for the difference between his stipend and the yakudaka and therefore receive an annual remuneration of 50 hyo.
Even if a daikan devoted the entire harvest season (i.e., about 3 months) to the inspection of arable lands and actual revenue, he was left with no more than one to three days per village—the time required to travel not included.
Even in less extreme cases, the fact that peasants routinely paid 2.5–3 % in excesses of the official tax rate to win the favor of local officials caused not even the slightest surprise among contemporary observers (Mizu no Tomenaga 1980/1981: 8–9).
Imperial Rescript on the Prefectural Assemblies, chapter 1, article 5 and chapter 4, article 33–34., in: McLaren, Walter Wallace, Japanese Government Documents, Yokohama (1914, p. 273).
Data from the 1880s for instance show that the median rank and median annual income of civil servants who held an official rank were 14th and 300 Yen, respectively, which was more than twice the wage of a manager at the private Mitsui bank (Kasuya 2005: 224).
Incentives in the civil service were increasingly shaped by inner-administrative competition. This process was initiated by the revoking of samurai privileges which caused birth rank and local family ties to lose their role as criteria of selection to merit-based indicators: Immediately after 1868, mostly activists of the Restoration movement were appointed. From the 1870s with the establishment of the first law schools till the 1887 general examination ordinance, appointment was made a matter of educational attainment and performance in standardized national examinations. As a result, roughly one-third of government positions was occupied by non-samurai newcomers after the first two decades of the Meiji period.
Law on the Duties of prefectural officials, issued July 27, 1876, article II,2, in: McLaren, Walter Wallace, Japanese Government Documents, Yokohama (1914, p. 268).
The corresponding article can be found in: Nishimura, Shigeki, ‘On Change’, Meiroku-Zasshi vol. 43 (Nov. 1875).
Corresponding records were kept on the prefectural and local levels and listed the dates, location, the reason and the number of persons involved in a protest. Thanks to the work of Aoki these data have been cataloged both for the Tokugawa and Meiji periods. In the present graph, I classified an uprising as caused by malfeasance if “officials’ corruption,” “despotic administration” and the like were reported as the causes of the disorder. For my data see: Aoki, Kouji, Hyakusho Ikki no Nenjiteki Kenkyu, Tokyo (1974), statistical appendix; Aoki, Kouji, Meiji Noumin Soujo no Nenteki Kenkyu, Tokyo (1967).
One particularly valuable source to investigate the experiences of the different prefectures is the Meiji 15nen/16nen—Chihô-Junsatsu-shi Fukumeishô (CJF), a collection of reports submitted to the Meiji government in 1882/1883 which gives detailed accounts of the economic, political and administrative state of affairs in every of the then existing Japanese prefectures.
For instance, gôshi were granted the right to wear arms and could be appointed to a limited number of public offices. Yet, their hereditary stipends were typically so small that they had to rely on additional sources of income—mostly agricultural activities.
The “Meiji Kan-in Roku” were published from the early years of Meiji onwards and contain comprehensive monthly lists of all administrative posts in the different prefectures together with the names and native provinces of the corresponding officeholders.
An important advantage of “Meiji Kan-in Roku” is the fact that it proves relatively robust against changes of prefectural boundaries in the early Meiji period. In fact, a central government act known as the Haihan-Chiken first transformed the feudal territories of the Tokugawa period into 72 prefectures, a number which was subsequently reduced to 47. Consequently, the variable "Native" might misstate the fraction of native to non-native officials if officeholders remained in their ancestral fief but were no longer recorded as natives when the territory in question was amalgamated into another prefecture. Fortunately, though, the historical circumstances and certain characteristics of the Kan-In Roku mitigate potential bias from this source. For one, there is no evidence that changes in prefectural boundaries were more frequent in regions which had performed particularly well or badly before the Meiji Restoration. The direct effect of territorial modifications was therefore probably small. As the indirect effect of misstating the true number of natives in newly formed prefectures is concerned, it suffices to note the following characteristics of the Kan-In Roku: Officeholders were listed with their names and family names. Hence, by following backwards the career path of an administrator, I could verify his native province even if prefectural boundaries changed in the course of his professional life. The problem then came down to dealing with new entrants to the civil service. Kan-In Roku treats this staff always in a way that—to the extent old prefectures ceased to exist—natives of these territories were recorded as born in the prefecture to which the old province was annexed. Changes in prefectural boundaries therefore tended to slightly overstate the true fraction of native to non-native officials. Bias from this source is, however, likely to distort the expected statistical estimates downwards rather than upwards. The reason is as follows: Since changes in prefectural boundaries were frequently undertaken for efficiency reasons, they ought to have improved administrative performance after a certain time. Distortions from the misclassification of newly appointed officials also becoming sizeable with a time lag, prefectures with overstated values of “Native” were arguably already reaping the benefits from the efficiency gains after the geographical restructuring. Hence, if high realizations of “Native” indicated institutional backwardness, the positive impact of improved administrative performance after boundary changes will at least partially offset this effect in the econometric estimation. This implies that if the statistical analysis confirms the anecdotical historical evidence—i.e., if the coefficients of “Native” are negative—these estimates would constitute a lower bound for the true relationship between administrative backwardness and economic development.
As mentioned before, the chôkunin rank was exclusively awarded by the Emperor and was restricted to high-ranking officials.
The reason for relying on relative rather than their absolute tax yields stems from the fact that—while taxation regulations were centrally determined—the set of taxed items and the corresponding rates changed over time. As a result, absolute tax revenue is not directly comparable for different years which rendered its use unsuitable for the present analysis.
Shuyo kigyo no keifuzu arguably constitutes the most comprehensive record of business activity in the early Meiji period and is the subject of a recent in-depth analysis by (Tang 2011). His work features a prefectural-level database with the annual number of company start-ups from the first year of Meiji onwards.
Considering this indicator, critical spirits might be concerned with the fact that Startup jt reflects only a part of overall economic activity as businesses which failed or were not listed at the Tokyo stock exchange in 1984 are not included in the database. However, issues of survivor bias are mitigated by the inclusion of defunct firms whose assets were transferred to direct successors (Tang 2011). Moreover, since favorable institutional conditions like administrative rationality should have similar effects both on the establishment and on the long-term success of businesses, this shortcoming does not seem to jeopardize the overall validity of the development indicator Startup jt .
In fact, there is a substantial literature (see p.e. Francks 2002; Howell 1992; Smith 1988) showing that agricultural and commercial development in rural areas relied strongly on better-off merchants and land owning families (sometimes referred to as "Meibôku") who combined traditional cultivation and landowning activities with money lending, trading or the establishment and operation of putting-out systems. If those agents were better able to assess the prospects of investments and the solvency of their local clientele, they might have picked the "plums" from the prospective borrowers while leaving more risky enterprises to other financial intermediaries. Hence, ordinary banks might have increased their risk premia which in turn would cause Lendingrate to be higher in the corresponding prefectures.
“AgriShare” arguably also reflects agricultural productivity: As shown in the seminal work of Nakamura (1966), agricultural innovation tended to be labor-intensive throughout the early Meiji period such that prefectures that derived most of their per capita income from agriculture should exhibit high realizations of “AgriShare.” Hence, “AgriShare” is unlikely to underestimate the relative importance of agriculture in highly productive regions as it would be the case if innovation had been primarily labor-saving.
Note that the necessity for modifications in the number of primary schools stems from the fact that historical records contain only data of private primary schools (shingakkô) and temple schools (terakoya). These schools were mainly frequented by commoners while most samurai had their children privately educated. Hence, measuring human capital formation by the unadjusted number of schools per capita is likely to bias the estimates in favor of regions with relatively large commoner populations. In order to correct for this problem, I proceeded as follows: Schooling rates are estimated to have reached about 40 % for commoners and 100 % for samurai. Using 1868/1869 census data, I identified regional samurai population p samurai at the time of the Meiji Restoration. Based on these material, I multiplied the number of schools in each prefecture by a correction factor of (1 + p samurai = p total × (1 = 0/4)). For an exhaustive account of measuring school attendance in pre-modern Japan see: Dore, R.P., Education in Tokugawa Japan, London 1984, S. 317–323.
The shôgunate and the daimyo collected lists of their fiefs' taxable kokudaka output. These data can be found in Hansei Ichiran (A Summary of the Domain System) edited by the governmental department for historical research as well as in the official Kokudaka and Population Tables of the Domains and Prefectures issued in 1869 and hold by the Okuma Shigenobu online library of Waseda University.
This phenomenon was widespread enough to enter contemporary Japanese under a special expression: “amakudari”—literally translated “descending from the heavens”.
Another piece of corroborative evidence can be derived from a direct measure of corruption. In its account of the late Tokugawa and early Meiji bureaucracies, Sect. 2 has outlined that administrative failure and questionable if not illegal activities of local officials were particularly frequent in the tax assessment of cultivated land. To reduce the tax losses from the common practice of undermeasuring the true size of cultivated fields, the imperial government conducted two large-scale cadastral surveys in 1872 and 1899. Unfortunately, the accuracy of these surveys is unknown. It is, however, known that they were mainly undertaken by local officials whose work was subjected to on-the-spot controls from centrally appointed surveyors. Hence, if undermeasurement survived into the twentieth century, it most likely did so in prefectures where traditional local influences were strong and administrative rationalization retarded. The first reliable quantitative assessment of undermeasurement was undertaken in the course of a comprehensive countrywide sampling survey in the 1950s (survey data are reported in Nakamura (1966) statistical annex). The corresponding reports include an Undermeasurement Index (UI) which denotes the ratio of the true size of a cultivated area relative to its reported size (i.e., if there is underreporting, the UI is larger than 1). Under the admittedly strong assumption that cross-prefectural differences were perpetuated for a long time or vanished very slowly, a high prefectural UI in the twentieth century would suggest a lack of institutional quality in the nineteenth century. I investigated this possibility by, respectively, correlating the two reported UIs (one for paddy and one for upland fields) with the institutional indicator “Native.” The resulting correlation coefficients were both positive and amounted to 12 % for paddy and 29 % for upland fields. With all necessary precaution regarding the long time span between the observed institutional data and the undermeasurement indices, the estimates obtained in section 3 are therefore also supported by direct indicators of corruption. (For corresponding data see report in Nakamura (1966) statistical annex).
References
Printed sources and statistical compendia
Bank of Japan First Hundred Years’ Editing Committee [ed.], Bank of Japan: The First Hundred Years: Materials, Tokyo 1986
Gabe, Masao [ed.], Meiji-15nen-Meiji-16nen-chiho-junsatsushi-fukumeisho (Collected Reports of Local Police Officers 1882–1883), reprint of original source, Tokyo 1980
MacLaren, Walter W. [ed.], Japanese government documents, Tokyo, 1914
Meiji-6-society (Meirokusha) [ed.], Meiroku Zasshi (Journal of the Meiji-6-society), Tokyo 1874–1876
National Statistical Office (Naikaku-Tôkei-Kyoku) [ed.], Nihon Teikoku Tôkei Nenkan (Statistical Annuals of the Japanese Empire), Tokyo various years
National Print Office at the Ministry of Finance [ed.], Kan-in Roku (later renamed Shoku-in Roku) (List of Government Officials), Tokyo various years
Statistical Offices of the Japanese Prefectures (Fu-Ken Tôkei-Kyaku) [ed.], Fu-Ken Tôkeisho (Statistical Annuals of the Prefectures), various years 1881–1902
Ministry of Education [ed.], Nihon Kyouiku-shi Shiryou (Documents on the History of Education in Japan), Tokyo 1890
Ôkawa, Kazushi, Shinohara Miyohei [ed.], Shôhi Shijutsu (Private Consumption), Tokyo 1967
Ôkawa, Kazushi [ed.], Chôki-Keizai-Tôkei—Bukka (Long Term Economic Statistics of Japan, Prices), Tokyo 1992
Ôkawa, Kazushi [ed.], Chôki-Keizai-Tôkei—Suikei to Bunseki (Long Term Economic Statistics of Japan, Regional Statistics), Tokyo 1992
Special Government Committee for Historical Research [ed.], Hansei Ichiran (A Summary of the Domain System), reprint Tokyo 1967
Literature
Acemoglu D, Johnson S, Robinson JA (2008) Persistence of power, elites, and institutions. Am Econ Rev 98(1):267–293
Aoki K (1967) Meiji Nômin Sôjô no Nenjiteki Kenkyû (chronological research on peasant uprisings in the Meiji period), Tokyo Shinsensha
Aoki K (1974) Hyakushô Ikki no Nenjiteki Kenkyû (chronological research on peasant uprisings), Tokyo Shinsensha
Baxter JC (1994) The Meiji unification through the lens of Ishikawa prefecture. Harvard University Asia Center, Cambridge
Besley TJ, Ghatak M (2009) Property rights and economic development. CEPR Discussion Papers 7243
Beasley WG (1972) The Meiji Restoration, 1st edn. Stanford University Press, Stanford
Befu H (1966) Duty, reward, sanction, and power. The four-cornered office of the Tokugawa Village Headman. In: Silberman BS, Haroothunian HD (eds) Modern Japanese leadership, 1st edn. University of Arizona, Tucson, pp 25–50
Befu H (1968) Village autonomy and articulation with the state. In: Hall JW, Hansen MB (eds) Studies in the institutional history of early modern Japan, 1st edn. Princeton University Press, Princeton
Buchheim C (2006) What causes late development? Insights from history. S Afr J Econ Hist 21(1):52–83
De Long JB, Shleifer A (1993) Princes and merchants: European city growth before the industrial revolution. J Law Econ 36(2):671–702
DeVere B (1966) Sidney, Okubo Toshimichi and the first home ministry bureaucracy 1873–1878. In: Harootunain HD, Silberman B (eds) Modern Japanese leadership. Transition and change, 1st ed. University of Arizona, Tucson, pp 195–232
Epstein L (2000) Freedom and growth, markets and states in Europe, 1300–1750, 1st edn. Routledge, London
Francks P (2002) Rural industry, growth linkages, and economic development in nineteenth-century Japan. J Asian Stud 61(1):33–55
Fukao K et al (2009) Zensôki Nihon no Ken-nai Sôseisan to Sangyô Kôzô, Hitotsubashi University Global COE Hi-Stat Discussion Paper Series 071
Harootunian HD (1970) Toward restoration, 1st edn. University of California Press, Berkeley
Honmi Y (1898) Satsuma Kenbun-Roku (Memoirs of Satsuma), Tokyo
Howell DL (1992) Proto-industrial origins of Japanese capitalism. J Asian Stud 51(2):269–286
Izumi S (2001) Bakufu no Chiki Shihai to Daikan, 1st edn. Koeibunkasha, Tokyo
Jansen MB, Roszman G (1986) Japan in transition, 1st edn. Princeton University Press, Princeton
Kasuya M (2005) Continuity and change in the employment and promotion of Japanese white-collar employees: the case of the house of Mitsui. Enterp Soc 6(2):224–253
Koh BC (1991) Japan’s administrative elite, 1st edn. University of California Press, Berkeley
La Porta R, Lopez-de-Silanes F, Shleifer A, Vishny R (1999) The quality of government. J Law Econ Organ 15(1):222–279
Machlup F, Taber M (1960) Bilateral monopoly, successive monopoly, and vertical integration. Econ New Ser 27(106):101–119
McLaren WW (1916) A political history of Japan during the Meiji era 1867–1912, 1st edn. Frank Cass & Co., London
Matsushita S (2006) Kagoshima Han no Minshu to Seikatsu, 1st edn. Nanpoushinsha, Kagoshima
Mizu no Tomenaga (1980/1981) Yoshi no Sasshi, vol 8–9. Choukaironsha, reprint of historical documents, Tokyo
Mosk C (2011) Traps embraced or escaped: elites in the economic development of modern Japan and China, 1st edn. World Scientific, Singapore
Motoyama Y (1997) Proliferating talent: essays on politics, thought, and education, in the Meiji Era, 1st edn. University of Hawai Press, Honolulu
Murakami T (1997) Edo Bakufu no Daikan, 1st edn. Dohsei, Tokyo
Nakamura JI (1966) Agricultural production and the economic development of Japan, 1873–1922. Princeton University Press, Princeton
Nishizawa A (2004) Daikan no Ichijou Seikatsu, 1st edn. Kodansha, Tokyo
North D (2009) Violence and social orders: a conceptual framework for interpreting recorded human history, 1st edn. Cambridge University Press, Cambridge
Ohkawa K (1957) The growth rate of the Japanese economy since 1878, Tokyo
Ohshima M (1994) Meiji Kokka to Chiiki Shakai, 1st edn. Iwanami Shoten, Tokyo
Ono T (1932) Kinsei Chihou Keizai Shiryou, Tokyo
Ôtsu J (1881) Nihon Kanri Ninyôron, Tokyo
Paik C, Steele A, Tanaka S (2012) Rebellion and taxation: evidence from early modern Japan (working paper)
Ramseyer JM, Rosenbluth FM (1998) The politics of oligarchy: institutional choice in imperial Japan, 1st edn. Cambridge University Press, Cambridge
Ravina M (1998) Land and lordship in early modern Japan, 1st edn. Stanford University Press, Stanford
Shleifer A, Vishny RW (1993) Corruption. Q J Econ 108(3):599–617
Smith TC (1988) Native sources of Japanese industrialization, 1750–1920, 1st edn. University of California Press, Berkeley
Soseki N (1914) Kokoro, Maddison Books (1992, Reprint of 1914); New York
Spaulding RM (1967) Imperial Japan's higher civil service examinations, 1st edn. Princeton University Press, Princeton
Steenstrup C (1996) A history of law in Japan until 1868, 1st edn. Brill, Leiden
Steiner K (1965) Local government in Japan, 1st edn. Stanford University Press, Stanford
Tacke V (1975) Die Entwicklung der japanischen Landwirtschaft in der Tokugawa- und Meji-Periode. Dissertation University of Basel, Basel
Tang J (2011) Technological leadership and late development: evidence from Meiji Japan, 1868–1912. Econ Hist Rev 64(1):99–116
Westney ED (1987) Imitation and innovation, 1st edn. Cambridge University Press, Cambridge
Wilson GM (1992) George, patriots and redeemers in Japan: motives in the Meiji Restoration, 1st edn. University of Chicago Press, Chicago
Yamamura K (1974) A study of samurai income and entrepreneurship, 1st edn. Cambridge University Press, Cambridge
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
About this article
Cite this article
Muehlhoff, K. The economic costs of sleaze or how replacing samurai with bureaucrats boosted regional growth in Meiji Japan. Cliometrica 8, 201–239 (2014). https://doi.org/10.1007/s11698-013-0099-x
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11698-013-0099-x