Weak central government, strong legal rights: the origins of divergent legal institutions in 18th-century Chinese and Japanese rice markets

The paper addresses why countries at similar development levels provided differing strengths of legal institutions, focusing on 18th-century rice markets in Qing China and Tokugawa Japan. Drawing on the predatory state perspective, the authors argue that formal institutions are created when they serve rulers' interests—particularly stability and longevity—rather than purely maximizing revenue, especially in non-Western contexts without interstate competition. They posit that domestic power threats (conflict and rebellion) and the costs of suppression are key drivers of legal institution-building. The study compares Suzhou (China) and Osaka (Japan), defining legal rights as the ability to trade and dispose of goods freely and to have those rights defended by courts. The authors hypothesize that when the capacity of noncooperation or rebellion to subvert a regime is high and legal protection mitigates this risk, states will enforce rights; when subversive capacity is low and suppression costs are modest, states have little incentive to protect market participants’ rights. The study contributes by offering a central–local perspective through a three-layer model (central government, mid-level agents, market participants) and by highlighting the interplay of formal legal systems and informal organizations.

The paper situates itself within literatures linking institutions to development and divergence (North; Acemoglu and Johnson; Rodrik et al.), as well as predatory state theory (Vahabi; Moselle and Polak; Levi) that emphasizes rulers’ incentives to provide protection to promote their own interests. It notes limitations of applying European interstate competition models to non-Western contexts and highlights the role of internal political structures and central–local relations (Ma and Rubin; Sng and Moriguchi). The authors build on work on the emergence of institutions and credible commitment (North and Weingast; Greif; Alston et al.; Rajan and Zingales) and repressive regime game theory (Ginkel and Smith; Kiss et al.), but extend these by modeling a three-tier interaction. They also engage with Greif and Tabellini’s argument about personal relations and clans in China contributing to weak formal institutions, and with research on decentralization’s benefits for public goods and development (Arze et al.; Faguet; Xu).

The authors develop a formal game-theoretic model—the subversive regime game—with three players: a central government (R), a mid-level agent (M: Chinese local government/Daimyo), and a lower-level participant (P: merchant/farmer). Sequence: (1) Nature draws the central government’s type θ (its ability to retain control if M does not cooperate), observed by R but not M; M holds a belief θ_guess. (2) R sets legal protection level α ∈ [0,1]. (3) M decides whether to cooperate (y=1) or not (y=0). (4) P decides whether to fight embezzlement (δ=0) or not (δ=1), partially reducing embezzlement based on strength z and incurring cost c. Structural variables include: friction cost f (costs to R and M from noncooperation, linked to military size and geographic scale), exposure cost k (risk faced by M when not cooperating), success reward X (benefit to M from toppling R’s market control), cost of fighting c, embezzlement parameter e, P’s strength z, and R’s stability reward ψ. Payoffs are specified for each player under all action profiles, capturing trade-offs among legal protection, embezzlement, suppression, and stability. The authors solve for Perfect Bayesian Nash Equilibria (proofs in Appendix), yielding comparative statics: (1) lower α induces P to fight; (2) higher α induces M to cooperate; (3) higher f and k make R less likely to offer legal protection; (4) stronger R (high θ, f, k) can hold power without strong legal protections. Empirically, the paper conducts a historical-comparative analysis: (a) documents military capacity, geography, and frequency/scale of food riots (Cao; Fukao et al.; Harada; Horiuchi; Kito) to proxy θ, f, k; (b) presents institutional and legal records (e.g., 1730 recognition of Dōjima market; 1761 ban on non-exchangeable rice securities; litigation cases) to measure α; (c) traces merchant organizational investments (clubhouses/huiguan counts) to proxy c and z; (d) compiles market performance indicators (trade volumes in Suzhou, Huai/Yellow river basins; Osaka volumes and participating Daimyos; integration tests) to assess outcomes consistent with the model.

• Model results: (1) Lower legal protection α increases P’s propensity to fight; (2) Higher α raises M’s incentive to cooperate; (3) Higher friction f and exposure k lead R to avoid offering legal protections; (4) Stronger R (high θ, f, k) is more likely to maintain control even without strong legal institutions.
• Mapping to cases: Qing China (strong central government with large military and high suppression capacity; fewer, smaller food riots; strong dominance over local officials/Yaren) tended to provide weak legal protections to rice market participants. Edo Japan (weaker central monopoly over force, powerful Daimyos, more frequent and larger food riots) provided stronger legal protections to reduce rebellion risk and stabilize Daimyo finances.
• Historical evidence: • Military and scale: Qing military peaked near 800,000 men and consumed about 60% of annual expenditure; Tokugawa Bakufu directly controlled ~30,000 troops (~5% of total), with some Daimyos able to mobilize ~20,000 each. • Geography increased Qing suppression costs and friction (vast territory; long distances), while Japan’s smaller size and coastal proximity reduced costs and increased the potential threat from mid-level agents. • Food riots per million population were over twice as frequent in Japan as in China in the mid–late 18th century; Japanese riots were larger and more destructive (e.g., 1787 Edo paralysis), implying higher domestic threat to the Bakufu. • Legal changes in Japan: shogunate recognized rice markets (e.g., Osaka Dōjima) after 1730; banned non-exchangeable rice securities in 1761; late-18th-century litigation upheld rights of holders of rice securities. • Weak enforcement in Qing: Yaren (Lgb) often embezzled; magistrates favored negotiated settlements and seldom punished Lgb; merchants’ rights commonly went unprotected.
• Merchant behavior: In China, merchants invested heavily in informal protections (huiguan/clubhouses), collective action, and private ties with officials; in Japan, legal protections enabled more individualistic strategies and business model investments (e.g., speculative trading by Honma Munehisa).
• Market performance: • China: Suzhou’s rice trade peaked mid-18th century and halved by early 19th century; broader declines in Huai and Yellow River basins and nationally thereafter. • Japan: Osaka’s rice trade volume rose steadily from early 18th century and more than doubled by early 19th century; number of participating Daimyos increased nearly fourfold. By the 19th century, Osaka’s commodity rice value was ~38% of total commodities; Suzhou handled about 20% of China’s rice trade; Osaka handled ~20–40% of Japan’s. • Integration: Late-18th/19th-century rice markets in Japan exhibited stronger national integration than China (ADF cointegration results).

The findings support the hypothesis that a state’s relative strength vis-à-vis domestic challengers shapes its incentives to provide legal protections. In Tokugawa Japan, the Bakufu’s limited monopoly over force and substantial Daimyo autonomy generated significant domestic threats. Providing credible legal rights in rice markets stabilized Daimyo finances (heavily dependent on rice sales), mitigated rebellion risks, and improved market performance and integration. Conversely, the highly centralized Qing state—with strong military control, limited local autonomy, and fewer/smaller food riots—faced lower risks of subversion and lower relative costs of repression, reducing incentives to enforce private legal rights. This institutional divergence shaped merchant strategies: weak formal enforcement in China encouraged collectivist, informal mechanisms (guilds/huiguan and official ties) to protect rights, diverting resources from trade expansion; strong Japanese protections facilitated individual participation and innovation, supporting market deepening (e.g., Dōjima market, securities/futures). The results underscore the importance of central–local relations for state-building beyond European interstate competition narratives and show how legal institutions feed back into economic stability, participation, and growth.

The study explains why pre-modern China and Japan adopted divergent legal institutions in core commodity markets despite similar overall stability. The subversive regime model shows that where mid-level agents pose credible threats and suppression is costly (Edo Japan), rulers are incentivized to supply strong legal protections; where the center is dominant and suppression is cheaper (Qing China), rulers have little incentive to protect private rights. This institutional choice influenced merchant behavior (individualistic vs collectivist strategies) and market trajectories (Osaka’s growth vs Suzhou’s decline). The paper contributes a central–local, three-layer perspective to state-building and institutional development in non-Western contexts and highlights how legal protections can stabilize regimes by aligning incentives across levels of governance. Future research could extend the framework across regions and periods, incorporate signaling about regime strength, and assemble richer quantitative evidence to test mechanisms more directly.

• Scope and parameters: The model focuses on a limited set of parameters (e.g., θ, f, k, e, c, z) and abstracts from other potentially important factors (culture, ideology, elite conflicts, international pressures).
• Data constraints: Archival litigation records are sparse, limiting comprehensive assessment of dispute outcomes in Qing markets; several quantitative proxies (e.g., huiguan counts for c and z, riot counts per capita) involve interpolation and may be noisy.
• External validity: The framework is tailored to pre-modern China and Japan; generalization to other contexts and time periods requires caution.
• Measurement: Some variables (e.g., central strength, exposure risk, embezzlement propensity) are inferred from qualitative and indirect indicators; causal identification is not claimed.
• Theoretical emphasis: Authors stress that elements of the empirical section are illustrative; more systematic econometric validation is left for future work.