Principal investigator: Omar Barroso Khodr
Authors: Junxue Jia, Xuan Liang, Guangrong Ma
Original title: Political hierarchy and regional economic development: Evidence from a
spatial discontinuity in China
Location of the Intervention: China (Chongqing and Sichuan counties)
Sample Size: 200 companies in 26 counties of Chongqing and 500 companies in 54 counties of Sichuan.
Primary Variable of Interest: The result of interest in city i (that is, the
rate of increase in nighttime light intensity over that explored city).
Type of Intervention: Economic growth (GDP) policies
Methodology: Spatial Discontinuity Regression (RD)
Summary
This study investigates the impact of political hierarchy on regional economic development, based on Chongqing's promotion to provincial-level municipality status in China. According to the authors, in 1997, Chongqing, then a prefecture-level city, was elevated to the category of municipality directly subordinate to the central government, separating from Sichuan province. This change resulted in a substantial increase in the city's decision-making autonomy in administrative, fiscal, and social matters. The border areas between Chongqing and Sichuan presented similar characteristics before the intervention, allowing us to adopt a spatial discontinuity regression approach. By analyzing the growth of light intensity at the municipal level between 1992 and 2013, the authors observe that economic activities in the cities of Chongqing followed trajectories parallel to those of Sichuan before the promotion, but began to grow significantly soon after the institutional change—implying an increase of 1,8 percentage points in the annual GDP growth rate. Furthermore, the authors explain that public officials in Chongqing demonstrate a higher degree of accountability and implement more growth-oriented policies than their counterparts in Sichuan. The results offer new evidence that strengthening the political hierarchy of local governments can boost regional development.
- Policy Problem
The central question lies within the long-standing theoretical debate on decentralization. The authors directly examine whether strengthening local governments by granting them greater political autonomy leads to better economic outcomes or, conversely, fosters corruption and the capture of power by elites. The authors' aim is to move beyond abstract theory and conflicting evidence from other contexts, offering a clear causal answer based on the analysis of a concrete case of comprehensive political decentralization.
A second, closely related issue concerns the ideal design of the political hierarchy and the size of administrative units. The study investigates the relationship between the benefits of smaller, more responsive governments—capable of adapting policies to local needs—and the costs arising from the loss of economies of scale and increased administrative complexity. By analyzing the rise of Chongqing, it presents empirical evidence on when and how the strategic fragmentation of a large province can constitute an effective political tool to stimulate regional growth.
Thirdly, the study emphasizes the crucial role of political institutions in determining the outcomes of decentralization, especially in authoritarian regimes. A key issue addressed is how to ensure local accountability in the absence of democratic elections. The results suggest that a specific institutional mechanism—a centralized performance evaluation system that rewards local authorities for economic growth—can act as an effective safeguard. This implies that the success of decentralization is not automatic, but depends on the broader system of incentives and controls in which it is embedded.
Finally, the research addresses a methodological gap present in the literature. Previous studies on decentralization have yielded heterogeneous results, partly because the term "decentralization" can encompass distinct dimensions (fiscal, administrative, political). By focusing on the case of Chongqing—which involved a comprehensive set of administrative, fiscal, and personnel powers—the study offers a clearer test of the impact of a significant and unequivocal change in the political hierarchy. In this way, it provides more robust insights for policymakers considering similar comprehensive reforms.
- Policy Implementation Context
This study is situated within the classic public policy debate on decentralization and the ideal size of governments. The elevation of Chongqing is used as an experiment to test traditional theoretical trade-offs. The first of these concerns the balance between suitability to local preferences and administrative efficiency. According to the authors, the creation of a smaller, more homogeneous jurisdiction, such as Chongqing, allows the local government to formulate policies more tailored to the specific needs of the population and to have superior information about the territory. On the other hand, fragmentation can increase costs, reducing economies of scale in the provision of public goods and hindering the management of externalities between jurisdictions.
The second critical commitment concerns political accountability. In theory, decentralization can make the government more accountable to the population, bringing it closer to citizens, strengthening social monitoring, and stimulating interjurisdictional competition for growth. However, it also carries the risk of intensifying local corruption and the capture of power by elites, especially when these elites are more cohesive and media scrutiny is limited (Hayek, 1945; Tiebout, 1956; Oates, 1972; Qian and Roland, 1998). This study explicitly frames its analysis within the context of an authoritarian state, where accountability mechanisms differ from those observed in democracies. The hypothesis is that China's specific institution—a centralized performance appraisal system that rewards local officials for economic growth—may act as a safeguard, allowing the benefits of decentralization (pro-growth policies and accountability to superiors) to be achieved while mitigating its typical negative effects (corruption and capture).
Thus, from a political perspective, the Chongqing case is not simply about creating a new administrative unit. It is a practical test to assess whether a central government can strategically use decentralization as an instrument for targeted economic development. The implementation is based on the belief that, for a large and populous region like Chongqing, the benefits of tailored policymaking, greater fiscal autonomy, and robust career incentives for officials would outweigh the costs of lost economies of scale and the greater coordination challenges faced by the central government. This study seeks to provide empirical evidence on whether this calculated policy gamble was successful.
- Evaluation Details
The empirical analysis of this study is based on satellite-derived nighttime light intensity data from the DMSP-OLS program, used as an indicator of local economic activity between 1992 and 2013. This measure is crucial since there are no official time series of economic data at the municipal level. A key methodological step involves combining these satellite images with the 2013 municipal administrative boundaries of China using ArcGIS. This procedure ensures a consistent measure of economic growth over time, neutralizing the confounding effect caused by changes in administrative boundaries during the period, by applying a single, stable geographic framework.
In this context, to complement and validate the conclusions derived from the luminosity data, the authors gathered a diverse set of cross-sectional and population census information. This includes official economic data at the municipal level (such as industrial production and urbanization rate) from 2013, information on public goods (roads, schools, hospitals) digitized from Google Maps, and comprehensive census data at the business level from 2008, allowing for detailed analyses at the municipal scale. In addition, the dataset includes specialized research and records, such as the World Bank Business Survey (2005), used to assess the business environment; census data at the individual level (2000 and 2005), used to study migration and ethnic composition; and fiscal and audit records at the municipal level, which allow for examination of government spending and the misuse of public resources.
Finally, the study incorporates topographic data (elevation and slope) from the Space Shuttle Radar Topography Mission. This geographic information is essential for the spatial discontinuous regression model, as it allows controlling for physical factors of the territory and testing the pre-treatment equilibrium on the border between Sichuan and Chongqing.
- Method
Initially, the authors considered a simple model. Difference-in-Differences (DID)However, they rejected it due to significant imbalances between the cities of Chongqing and Sichuan prior to the treatment, related to factors such as altitude, minority population proportion, and nighttime illumination in 1996. An event study confirmed that the assumption of parallel trends—necessary for the validity of the DID model—was violated, since the trajectories of illumination growth were not parallel before the reform.
In this way, to circumvent these problems, the study employs a Spatial Discontinuity Regression (SDR) model, using cities near the border between Sichuan and Chongqing. The main identifying premise of this approach is the "smoothness assumption," according to which all relevant characteristics before treatment vary continuously across the border. The basic econometric specification consists of a local linear regression that models economic growth (measured by the logarithmic variation in nighttime light intensity between 1996 and 2013) as a function of a variable. dummy treatment for location in Chongqing, simultaneously controlling a smooth polynomial function of geographic position (latitude and longitude). The coefficient of the variable dummy The treatment captures the causal effect of the Chongqing promotion, and the standard errors are grouped at the county level.
Fundamentally, the empirical strategy is validated by pre-treatment equilibrium tests. Within a 30 km range of the border, the study finds no statistically significant differences in light intensity in 1996, altitude, slope, distance to major cities, or proportion of minority population. This reinforces the claim that cities on both sides of the border were similar before treatment. Furthermore, a dynamic analysis shows that these cities exhibited parallel trends in light intensity growth in the years preceding the promotion (1997), with significant divergence only after treatment. The robustness of the main results is confirmed through alternative bandwidths and a global polynomial specification.
- Main results
According to the authors, the main conclusion of the study, derived from a Spatial Regression with Discontinuity (SR) model, is that the province of Chongqing had a statistically and economically significant positive effect on economic growth, as measured by nighttime light intensity. The baseline specification—a local linear regression with a bandwidth of 30 km—estimates that the cities of Chongqing showed an annual growth in light intensity 6,3% higher than that of the cities of Sichuan, which corresponds to an increase of approximately 1,8 percentage points in the annual GDP growth rate. This result is visually corroborated by a clear discontinuity in the growth of luminosity at the border.
The robustness of this conclusion is rigorously evaluated through several methodological checks. First, the result proves stable across different model specifications, including local linear and quadratic regressions with bandwidths of 30 km and 50 km, as well as global polynomial regressions applied to the entire sample. Second, the estimates remain remarkably consistent even after changes in key parameters: they are insensitive to a wide range of bandwidths (from 20 km to 100 km), the use of higher-order polynomials, the inclusion of additional geographic controls and fixed effects of boundary segments, as well as the application of triangular kernel weighting, which assigns greater weight to observations closer to the boundary. Finally, the result holds when the unit of analysis is changed from cities to uniform-sized grid cells, mitigating concerns related to heterogeneous spatial units.
To reinforce the validity of the empirical strategy, the authors conduct convincing placebo tests. First, they artificially shift the treatment boundary by 30 kilometers, both east and west; in both falsified scenarios, no significant discontinuity in the increase in luminosity is observed, confirming that the effect is exclusive to the real administrative boundary. Next, they implement a more comprehensive placebo test, drawing 1.000 random, non-rectilinear boundaries within the region. The cumulative distribution of the coefficients from these tests shows that the estimated true effect is greater than 95% of the placebo effects, strongly indicating that the result does not stem from chance or unobserved spatial confounding factors, but rather from a genuine consequence of the political intervention.
- Lessons in Public Policy
Based on the study's results, several relevant suggestions for public policies emerge, focusing primarily on the benefits and strategic considerations of decentralizing power to local governments.
The main political implication is that strengthening local governments—through raising their political status and expanding their decision-making autonomy—can act as a powerful catalyst for regional economic development, as highlighted by Bolton and Dewatripont (1994), Qian (1994), and Garicano (2000). The study demonstrates that the promotion of Chongqing resulted in greater accountability of local officials and the adoption of pro-growth policies, generating significant economic gains.
Thus, it is suggested that, in large and heterogeneous countries, a certain degree of political and fiscal decentralization can improve governance and boost growth by allowing local leaders—more familiar with the specific conditions of their regions—to formulate and implement more effective policies. Furthermore, the results indicate that, in authoritarian systems, a well-structured hierarchy with clear reward and punishment mechanisms can enhance the benefits of decentralization while mitigating its risks, such as corruption and capture by local interests.
However, the authors also warn against the indiscriminate adoption of regional promotion policies. They highlight two critical constraints that policymakers should consider. First, the economic costs: increased administrative expenses resulting from the loss of economies of scale, greater difficulties in managing cross-border impacts (such as environmental or infrastructure issues), and the information overload on the central government, which hinders coordination. Second, the political costs: such reforms tend to face resistance from incumbent provincial governors, who would lose authority and control over the personnel in the promoted regions. Thus, the decision to raise the status of a region must be strategic, carefully weighing the proven economic benefits against the concrete administrative and political obstacles. This implies that this policy tool is more suitable for specific and strategically relevant regions than for universal application.
References
Bolton, P., Dewatripont, M., 1994. The firm as a communication network. Quart. J.
Econ. 4, 809–839.
Garicano, L., 2000. Hierarchies and the organization of knowledge in production. J. Polit. Economy. 108(5), 874–904.
Hayek, FA, 1945. The use of knowledge in society. Am. Econ. Rev. 35 (4), 519–530.
Oates, W., 1972. Fiscal Federalism. Harcourt, New York.
Qian, Y., 1994. Incentives and loss of control in an optimal hierarchy. Rev. Econ. Stud. 61(3), 527–544.
Qian, Y., Roland, G., 1998. Federalism and the Soft Budget Constraint. Am. Econ. Rev. 88 (5), 1143–1162.
Tiebout, CM, 1956. A pure theory of local expenditures. J. Polit. Economy. 64(5), 416–424.
