Social networks are an important lubricant in an economy when market imperfections exist. They provide information (Greif, 1993), promote trust (Karlan et al., 2009), help enforce contracts, and reduce moral hazards (Rauch and Trindade, 2002). However, social networks may also enable agents to collusively divert resources from their best use, which undermines the interests of broader society. Such diversion effects may be especially strong when elite members in a network have substantial power over resource allocation.
A recent line of literature has focused on the dark side of social ties and found the distortions to be large (Haselmann et al., 2018). Because Chinese local governments hold substantial power over local administrative and economic affairs, social ties among local politicians in China could also have significant implications for the spatial allocation of resources and the economic efficiency of such allocation. However, primarily because of data limitation, there is a paucity of empirical evidence in this area. In our recent study, we examine a novel Chinese dataset on firm registrations to study the impacts of social ties among city politicians on the spatial allocation of capital in China and the implications of these impacts.
Local Chinese officials are frequently relocated to different localities following new appointments by the higher government. Depending on the post, local officials are replaced by new appointees every three to five years. Such replacements are often unexpected and therefore provide exogenous changes in the social ties of politicians in different localities (Shi et al., 2021). We consider these relocations as quasi-experiments and examine the causal effects of the social ties of politicians on intercity capital allocation before and after the replacements of the politicians.
We focus on the hometown ties of city party secretaries, who are the top officials at the city level in China. Although social ties can also be formed through other means, such as a common educational background or workplace, hometown ties are the cornerstone of trust and sense of a common identity in Chinese society and therefore a key element of social proximity (Chu et al., 2021). We measure city-dyad investment flows using the capital registration of a new firm in a destination city by its owner from an origin city. This information is obtained from the universe of firm registration data recorded by the State Administration for Industry and Commerce (SAIC) of China. Empirically, we adopt a difference-in-differences strategy to compare the city-dyad investment flows with and without connected city party secretaries while controlling for city-dyad and year fixed effects to remove effects related to time-invariant city-pair heterogeneity or common year shocks.
We find that the total value of city-dyad investment flow increases by 10% on average when city party secretaries share a hometown. The hometown ties also increase the number of firm registrations by 1% and increase the likelihood of any firm registration by 1% between the cities. The effects go in both directions within the city dyad and are stronger when both party secretaries have incentives for political promotion. Event studies verify that these results are not driven by differential pre-trends, and all effects begin only after party secretaries with hometown ties are in position.
Social ties can have both bright-side effects (e.g., reducing transaction costs) and dark-side effects (e.g., inducing rent-seeking behaviors). These two effect types are often observationally equivalent but entail opposite implications for economic efficiency. We focus on three mechanisms through which social ties influence investment – reducing transaction costs, reducing moral hazards, and facilitating rent-seeking – which are relevant to our study and widely discussed in the literature. First, connected officials may exchange information as matchmakers to facilitate bilateral investment. We examine the geographic distance between cities as a proxy for information asymmetry and find that bilateral investment increases more in distantly separated city pairs than in nearby ones. We also find the effects to be stronger among smaller firms that are more likely to be constrained by information frictions and other entry barriers. Second, connected officials may reduce the “grabbing hand” behavior of lower-ranking bureaucrats. If the mechanism holds, then the survival rates of firms would improve. The evidence is consistent with the prediction. Third, connected officials may jointly divert resources to one another for rent-seeking purposes. We examine the heterogeneous effects that vary according to whether party secretaries of the city pairs were convicted of corruption prior to 2021. We find the effects to be stronger in city pairs in which party secretaries were corrupt.
These results, however, are insufficient to establish the mechanisms underlying such phenomena. For instance, hometown ties create trust between officials and lower the barrier to rent-seeking, which explains why distant firms and smaller firms are more affected. Being able to survive for a longer time may also reveal cronyism instead of legitimate protection against extraction (or “grabbing hand” by local bureaucrats). We then examine the effects of hometown ties separately before and after the 2012 Chinese anti-corruption campaign, which marks a watershed in recent Chinese history of corruption deterrence and punishment. The campaign targeted not only corrupt dealings between government officials and private businesses but also coalitions among politicians, which are often based on social ties. Given that the campaign significantly increased the costs of corruption, we expect that the differential results for the pre- and post-campaign periods would reveal the relative importance of the different mechanisms.
We find that the effects of hometown ties on city-dyad investment become nonexistent in the post-campaign period, which is in line with what the rent-seeking mechanism would predict. We further investigate the prices paid by firms in land acquisitions, which have been recognized in the literature as an important rent-seeking behavior (Chen and Kung, 2019) that benefits firms and lowers the firm entry barrier. We find that prior to the anti-corruption campaign, the hometown ties of local politicians reduced land prices paid by firms and increased the amount of land acquisitions. However, these effects no longer occur in the post-campaign period. To further examine the rent-seeking channel, we compute the connected investment share of a city, which is calculated as the share of a city’s investment from or to cities with hometown ties out of the city’s total investment. We find that prior to the anti-corruption campaign, there are strong positive correlations between GDP growth and connected investment share, between the chance of political promotion and connected investment share, and between the chance of corruption conviction and connected investment share. However, these correlations are nonexistent after the campaign started. These outcomes reveal that connected investments may bring career benefits to local politicians in ways that make them susceptible to corruption.
Yu Liu is an Associate Professor of Economics at Fudan University.
Xiangyu Shi is a PhD candidate in Economics at Yale University.
This post was adapted from their paper, “Reciprocity in Political Network: Hometown Ties and Intercity Investment in China,” available on SSRN.
