Transportation Revolution and Banking Development：Evidence from Prewar China

doi:10.21203/rs.3.rs-2480760/v1

Download PDF

Research Article

Transportation Revolution and Banking Development：Evidence from Prewar China

https://doi.org/10.21203/rs.3.rs-2480760/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

This study investigates the effects of railroad placement on the development of modern banking in China from the late nineteenth century to the Pacific War. We construct datasets on railroad stations and bank locations across 250 prefectures between 1895 to 1937. The results show that construction of railroad stations significantly promoted the development of banking. More specifically, a new railroad station in a prefecture brought about 3.063, or 24%, more banks. We explore two potential mechanisms for this, finding that railroad placements stimulated modern banking both through 1) helping develop modern industries along rail routes and 2) through the cooperation of traditional local institutions (qianzhuang). Furthermore, we find that railroad placement allowed areas far from treaty ports to gain connection with the international market. The development of foreign trade also enabled modern banks to develop new business based on cargo mortgages, thus spurring geographic expansion of bank branches.

JEL Classification: N25, N75, O16

Railroad

Modern Banks

Financial Development

Semicolonial country

The modern banking industry is essential to promoting economic growth (Levine, 1997; Rajan and Zingales, 1998). It has been argued that the development of the modern banking industry was associated with legal origins (La Porta et al. 1998), cultural characteristics (Pascali, 2016; D'Acunto et al., 2018), and information technology changes (Petersen and Rajan, 2002; Berger and DeYoung, 2006). The transportation revolution of the 19th century has also been postulated as stimulating the expansion of banking networks. For example, Atack and Rousseau (2014) found that railroad construction in Alabama, in the United States, increased the trade potential of areas along its routes and attracted more banks to enter. Such transportation revolution was introduced to Asia by western colonial forces in late 19th century and had a profound impact on colonized or semi-colonized countries’ economy. However, empirical evidence on the role of railroads in the development of the modern banking industry in these regions is still lacking. Different from the studies on the economic impacts of western railroad construction in Asian colonies,¹ this study focuses on the functions of railroads in a semi-colonized country rather than a fully colonized one.

In China (circa 1840–1936), modern banks were introduced by Western colonial forces and were highly competitive with China’s traditional financial institutions.² As one of the institutional changes sweeping China in the late nineteenth century, the development of modern banking was associated with railroad construction in the process of early-stage industrialization. While many studies have found that railroads generated economic gains and promoted economic growth (Fogel, 1964; Donaldson, 2016), the economic contribution of railroads in early industrial era China is different from that in the United States, where railroads created new urban settlements as population shifted along the lines (Atack et al., 2010). In contrast, the most important benefits of railroads in modern China were the establishment of linkages between the agricultural hinterlands and regional economic centers (Köll, 2019). Railroads integrated regions into the domestic market with their reductions in transportation costs, increasing the volume of local trade, in turn stimulating an enormous demand for trade financing along railroad lines (Rawski, 1989). Another aspect that makes the Chinese case unique was its monetary system, dominated by precious metals. During the period we studied (1895–1937), China was the last and largest country on the silver standard, with the most important currencies in the market silver tael and copper coins. The weight standard of the coins and the accounting units of currency varied widely among regions in China, however. It thus resulted in major extra costs to not only domestic but international trade settlements. In this context the link between railroad construction and banking development is still an unsolved problem in China's economic history. Specifically, China’s regional diversity of economic development and very different levels of railroad penetration, give us a chance to investigate the economic effects of railroads more deeply. Despite railroads boosting regions' trade potentials and thus banking businesses, banks still had to overcome the challenges of complex monetary systems to master the emerging trade opportunities. The traditional institutions,³ qianzhuang, have their advantage in their cross-regional trade settlement networks, and cooperation with qianzhuang has become one of the most important means for modern banks to expand their businesses in China.

In this paper, we analyze the causal effect of railroads on banking development in late 19th and early 20th century China.We find that railroad construction significantly promoted the development of modern banks. This impact remains after solving endogeneity issues by using the least-cost paths (LCP) of railroad networks as the instrumental variable. Specifically, a prefecture has 3.063 more banks if it has a railway station, and a prefecture has 24% more banks if it has one more such station. We conducted placebo tests, using the planned but unconstructed railroad lines proposed in Dr. Sun Yat-sen’s ‘International Development of China.’ The analysis shows that the impact of railroads on the modern banking industry is quite robust. Moreover, we explore two potential mechanisms for this: The railroads promoted industrialization in the areas along their lines and stimulated enormous financing demands, which in turn promoted the development of the banking industry. Meanwhile, railroad placement also increased trade credit demands, and modern banks have been able to expand their branches to wider geographical areas by collaborating with traditional financial institutions (i.e., qianzhuang).

This study thus contributes to existing literatures in three ways. First, it contributes to the literature on the effect of railroads in the early stage of industrialization (Bogart and Chaudhury, 2013; Berger and Enflo, 2017; Donaldson and Hornbeck, 2016; Jedwab et al., 2017). In a recent study, Hornbeck and Rotemberg (2019) find that the expansion of American railroads in the late 19th century improved the efficiency of resource allocation, allowing greater capital and labor inputs to flow to manufacturing and service industries. We take early 20th-century China as another example due to its unique monetary system and the relevance of modern banking to the development of its overall economy. The results illustrate that in the presence of currency inconsistencies among regions, modern banks can additionally benefit from the business opportunities brought about by railroad expansion, while their success still depends on collaboration with traditional financial institutions. To the best of our knowledge, this study is the first to provide new evidence on the relationship between railroad construction and the development of modern banking in early 20th-century China.

Second, this study enriches the literature on the determinants of the development of the banking industry in modern China, such as banking entrepreneurship (Cheng, 2003), telegraph technology (Lin et al., 2021), and Confucian culture (Chen et al., 2021). We emphasize that in addition to these, railroads support the business innovation of China's modern banking system in the early 20th century. For example, documentary billing service – a new credit financing business closely tied to trade – emerged after railroad construction. This new business required the storage of guaranteed freights of merchants, and modern banks usually built warehouses near railroad stations to facilitate storage and transportation. The high profits from new business prompted bankers to expand their branch networks across wider geographic areas. The potential profits from such business innovation thus contributed to the rapid development of the modern banking industry.

Third, this paper is also related to studies on how early Western institutions influenced China's economy in the 20th century, especially on the impact of colonial institutions on financial development (Ma, 2019; Keller and Shiue, 2020; Levine et al., 2021). Previous studies found that the legal system regarding concessions provided better protection of property rights than was usually the case in colonies, which in turn contributed to the development of the financial industry. Ma (2019), for example, illustrates that extraterritorial jurisdiction under the treaty system was conducive to the protection of property rights in times of political instability and promoted the development of China's banking industry. Despite this, little progress has been made in the explanation of how bankers expanded their business far from concession areas, and this study provides new evidence. For prefectures that have no concessions, railroad construction did not contribute to the establishment of headquarters, but it significantly stimulated branch openings even though the effect was smaller than that of prefectures having concessions. This study thus enriches existing body of research by addressing the role of railroads in connection of headquarters and branches. Railroad placement allowed areas far from coastal cities to be integrated into the international trade network, in turn spurring the expansion of the banking network and the establishment of branches in prefectures far from concessions.

The remainder of the paper is structured as follows: Section 2 describes the historical background of railway construction and the rise of modern banks in the early 20th century. Section 3 introduces data and models used. In section 4, we discuss the main empirical results and assess the robustness of our findings, the validity of our instruments, and the heterogeneity of the results. We then examine the role of modern (early 20th century) industry and traditional financial institutions (qianzhuang) on banking expansion. Section 5 offers some concluding remarks.

[1] For example, Donaldson (2018) studies the impacts of British railroad construction in India.

[2] Three main types of financial intermediaries dominated China’s financial market at the outset of the period discussed, in the late 19^th century: piaohao, qianzhuang and foreign banks. In this paper, we mainly focus on the development of indigenous modern banks, and our use of the term ‘modern banks’ refers to domestic commercial banks. Though this paper deals primarily with developments in early 20th century (and, to some extent, late 19th century) China, it uses the term ‘modern’ in a qualified way to characterize the newly emergent banking institutions of that period, in contradistinction to ‘traditional’ lending and financial operations that go back many centuries and, in some cases, to ancient times.

[3] Three main types of financial intermediaries dominated China's financial market at the outset of the period discussed, in the late 19th century: piaohao, qianzhuang and foreign banks. In this paper, we mainly focus on the development of indigenous modern banks, and our use of the term ‘modern banks’ refers to domestic commercial banks. Though this paper deals primarily with developments in early 20th century (and, to some extent, late 19th century) China, it uses the term ‘modern’ in a qualified way to characterize the newly emergent banking institutions of that period, in contradistinction to ‘traditional’ lending and financial operations that go back many centuries and, in some cases, to ancient times.

Since the first Opium War (1839–1842), the Qing Empire (1644–1911) was continuously confronted with the invasion of Western (and Japanese) allied forces that successively demanded that China sign so-called ‘unequal treaties.’ The Qing government was forced to open a series of ‘treaty ports’ in specific coastal locations that had previously been closed to Western traders, in which tariffs on imports into China were quite low (usually 5%). Meanwhile, western powers established a group of concessions in China. Each concession had its own police force and different legal jurisdictions with their own separate laws. In addition, to broaden Western powers’ political influence in China, a new institution, the Chinese Maritime Customs (CMC), was founded in 1854 by the foreign powers to collect maritime trade taxes. In this semi-colonial context, the Qing government started the ‘Self-Strengthening Movement,’ whose main purpose was to learn modern science and technology from the West. Railroad construction was also one of the many industrialization attempts by imperial authorities during this period.

2.1 Transportation Revolution from Railroad Construction

Railroads and steam engine technology originated in the United Kingdom in the early 19th century. Due to the seclusion policy of the Qing Empire, these modern transportation technologies were not adopted in China until the late 19th century. In June 1881, the Qing government began to construct the Kaiping Colliery Tramway, its first such project, which opened in November of that year. It was approximately 9.2 km long, connecting the Tangshan coal mine in Hebei, a northern province of China, with Xugezhuang, a town near the capital Beijing. In 1888, the Kaiping Colliery Tramway was further extended to connect Tangshan with Tianjin, the largest treaty port in North China.

Two main north-south railroad lines were built by the end of the Qing dynasty. One was the 1,214-kilometer Jinghan railway that connected Beijing with Hankou, the economic center of the middle reaches of the Yangtze River. The other was the 1,009-kilometer Jinpu railway, which connected Tianjin with Pukou in Jiangsu Province in the Yangtze River Delta region. Panel (B) of Fig. 1 shows the railroad network in 1912.

In the two decades after 1912, the railroad network was further expanded by 4,265 kilometers, not only on the main lines but also through the extension of branch lines. From 1927 to the outbreak of the Second Sino-Japanese War in 1937, the railroad network was further expanded in the southwestern region of China. This ever-expanding network allowed more convenient transportation of products from China's vast hinterland to coastal cities. Panels (C) and (D) of Fig. 1 show the national railroad network in 1920 and 1935, respectively.

2.2 China's Financial System: from Traditional to Modern

Before the emergence of modern banks, three main types of financial intermediaries dominated China's financial market: piaohao, qianzhuang and foreign banks (Cheng, 2003). Piaohao was a national financial institution, mainly providing cross-regional exchange services for individuals and government agencies. In 1911, it lost its largest customer with the fall of the Qing government. Soon, it failed in competition with modern banks and finally exited the financial market. As traditional small local banks, qianzhuang led in the domestic financial market and kept up close cooperation with local merchants and enterprises by providing short-term micro-loans.

To meet the rapidly growing financial settlement needs of international trade, foreign banks opened branches and offices in treaty ports and became increasingly important institutions in China's financial market during the Treaty Port Era (1842–1943). For example, the British Oriental Bank, the first foreign bank in China, established its first branch in Hong Kong, what it called a ‘sub-office’ in Canton in 1845, and a branch in Shanghai in 1847. The foreign banks introduced new organizational forms, using concepts of ‘limited liability,’ closely related to the evolution of other modern capital-based organizations.

Distinct from the form of unlimited liability partnership used by China's traditional financial entities, a new form with legal personality and limited liability manifested itself as a then ‘modern’ bank. The new organizational form allowed banks to have more capital and a lower risk profile than qianzhuang or piaohao. Modern banks acquired relatively extensive businesses, including not only the deposit businesses shared with traditional financial institutions but also exchange and trade settlement services. As an indigenous (non-foreign owned) modern bank, the Imperial Bank of China was the first, established in 1897. During the 1897–1927 period, the government did virtually no bank regulation, so the banking industry was in a state of virtually completely free competition (Du, 2015). In this context, modern banks became important competitors of traditional financial institutions, particularly after 1911, when domestic-owned banks entered a period of rapid development.

Modern banks adopted a headquarters-branch management system. The headquarters were usually established in treaty ports such as Shanghai, Tianjin, and Beijing. In 1912, there were 51 bank headquarters and 215 branches nationwide in China. By 1936, the number of headquarters had increased to 310, and the total number of bank locations, including both branches and headquarters, had reached 1,148. Figure 1 presents the nationwide expansion of modern banks during the 1911–1935 period. From the perspective of ownership structure, there were several types of modern banks: foreign banks, central and chartered banks, provincial and municipal banks, and private banks. As foreign banks had extraterritorial rights in China and were not restricted by regulations of the Chinese government, we exclude these banks in our further discussions.

2.2 Modern Banks and the Railroad Network

The establishment of modern banks was closely associated with railroad construction from an early stage. The first modern bank in China, the Imperial Bank of China, was founded by Xuanhuai Sheng, who was also the Minister of Railroad Affairs. He wrote: ‘If China wants to open a steel factory, it has to run a railroad; if it runs railroads, it has to run a bank.’⁴ Another national bank, the Bank of Communications, was founded for the railroad, shipping, and postal service businesses, and related payments required by the Ministry of Posts and Communications. In addition, many private modern banks were also founded, largely catering to railroad companies. For example, Zhejiang Industrial Bank, one of the famous ‘southern three banks,’⁵ was founded to facilitate the opening of the Zhejiang Railway Company.⁶ Correspondingly, one-third of the total shares of Zhejiang Industrial Bank were held by that company.

Railroads connected coal mines with the sites of emergent modern industry, such as steel factories built by the Qing government during the ‘Self-Strengthening Movement.’ The lower transport prices of raw materials offered by railroads meanwhile stimulated merchants to establish new factories. Traditional banks, qianzhuang and piaohao were either sole proprietorships or operated under unlimited liability partnerships, and were unable to provide the large, capital-intensive and long-term financing required by modern industry (Cheng, 2007). The fundraising difficulties of modern industries thus brought about opportunities for modern bankers to develop their businesses. For example, a new business mode, the documentary bill service initiated by modern banks allowed modern industries or merchants to borrow in more complex ways, consolidating their freight financing into mortgage loans.

The mortgage loan business of modern banks meanwhile created a large demand for collateral storage. Out of convenience, railroad stations became regional cargo distribution centers, and modern banks-built warehouses near the stations to provide warehouse guaranty business to merchants. For instance, Kincheng Bank, one of the most influential banks in northern China,⁷ built the largest cotton warehouse in the country in Zhengzhou, the main rail hub in northern China, which was convenient for merchants to mortgage goods for loans. The cargo warehouses built by other modern banks were mostly located in the commercial streets near Zhengzhou train stations in the early 1930s (Liu, 2018). In 1933, the Bank of China operated 30 cargo warehouses that stored one fifth of the nation's crops including cotton, tobacco, and tea.

In addition, the railroad connected treaty ports with the rest of China, involving more of its regions in the global trade system of the 20th century (Köll, 2019). The potential profit opportunities growing from the expansion of trade also promoted the expansion of modern banks' businesses to regions along the railroad lines (Rawski, 1989). The Beijing office of Kincheng Bank, headquartered in Tianjin, was asked by customers to open offices along the Longhai railway⁸ to meet their demands for financing.

[4] Sheng X., Yu zhai cun gao (Yuzhai manuscript), volume 25, page 15.

[5] Southern Three Banks included the Shanghai Commercial & Savings Bank (Shanghai ChuXu Yinhang), the Zhejiang Industrial Bank (Zhejiang Xingye Yinhang), and the Chekiang Industrial Bank (Zhejiang Shiye Yinhang), private banks developed in modern China.

[6] The historical evidence is based on Zhejiang Xingye Yinhang Gubenzhang Tongji (The Stock of the Zhejiang Industial Bank), cited from Hong (1993), p. 180.

[7] There were four banks that played an influential impact on the finance markets after 1927, including the Yien Yieh Commercial Bank (Yanye Yinhang), the Kincheng Bank (Jincheng Yinhang), the China & South Sea Bank (Zhongnan Yinhang), and the Continental Bank (Dalu Yinhang). Those banks established the famous Joint Savings Bank Warehouse (Sihang Cangku) in 1931, which seized international attention at the beginning of the Sino-Japanese War.

[8] The historical evidence is based on the Jincheng Yinhang Shiliao (The History of the Kincheng Bank), compiled by the Institute of Finance in the Shanghai branch of the People’s Bank of China in 1983. See Rawski (2009) on page 164.

The empirical analysis below is based on prefecture-level panel data as county-level data are unavailable and missing during the study period from 1895 through 1937. The administrative level of prefectures is similar to that of contemporary cities. The average area of the prefectures in our sample is 13,680 square kilometers. The smallest is Haimen Hall in Jiangsu Province in eastern China, with an area of 8,000 square kilometers, while the largest is Anxi Prefecture in Gansu in northwestern China, with an area of 133,800 square kilometers.

Although the names and boundaries of prefectures were not usually changed before 1911, some prefectures might have been affected by several regime changes during that period. To make the sample unit consistent, we traced all the historical changes in administrative boundaries and names of prefectures, and then unified these prefectures based on the 1911 map taken from the China Historical Geographic Information System (CHGIS, version 6). There were total of 381 prefectures in 26 provinces on the 1911 map. Due to data limitations, our sample covers 250 prefectures in 18 provinces throughout the period, which included 94.22% of the total population in 1880, 91.82% in 1910 and 93.71% in 1920⁹. Figure 2 shows the distribution of our samples.

3.1 Modern Banks

We collected data on modern banks mainly from The Dictionary of Banking Institutions in Modern China (Jiang, 2014), which recorded financial institutions established in the 1850–1949 period, including their names, the years and locations of their establishment, etc. To confirm the accuracy and completeness of the dictionary, we cross-checked the information in The National Bank Yearbook (1933–1937), Liu (2007), and The Shun Pao Yearbook (1933). The sample covers a total of 1,108 domestically owned modern bank locations, including 589 headquarters and 519 branch locations.¹⁰

The measure of the development of modern banks was constructed by aggregating their number of separate locations at the prefecture level on an annual basis. Both headquarters and branches were included in the aggregation but foreign banks were excluded. Moreover, we use the number of branches at the prefecture level to further investigate the expansion of modern banks, which were established in more than 20% of prefectures by 1917 and in half by 1934. More than 18% of prefectures had bank headquarters as of 1927 and 40% of prefectures had branches by 1937.

3.2 Railroad Construction

Data on railroad construction, including the years of railroad operation and location of stations, were mainly collected from Traffic History compiled by Road Administration of Republic China (1925), the Railroad Yearbook (1933), The Statistical Data of Modern Chinese Economic History (Yan, 1955), and A Brief History of Railway Construction in China (Zhang, 1997).¹¹ We first use station location information to identify prefectures with railroads. We then use the year of initial railroad operation to determine when the railroad began to open to traffic. Combining the two pieces of information, we form two measures of railroad construction. One is a dummy variable to indicate whether the prefecture had at least one railroad station in a specific year (1 = Yes) and is used to capture railroad expansion.

The second measure, which aims to depict station growth, is the number of railroad stations in the prefecture in a specific year. By 1911, for example, 100 stations had been built in the Canton prefecture, which had the most famous treaty port. The railroad stations also increased in Jinan (in Shandong Province) and Tianjin prefecture (in Zhili Province).¹²

Figure 1 presents the spatial distribution of modern banks as well as railroads every decade in the 1900–1935 period. In 1907, there were already some of these banks located along the Huning railway or concentrated in terminal stations such as Beijing (the Jinghan railway), Jinan (the Jiaoji railway), and Hankou (the Jinghan railway). By 1927, China's railroad network was basically completed. As illustrated in the left sketch of Fig. 1, many modern banks had their locations along the railroad routes, particularly in the hubs including Beijing, Tianjin, Jinan, Zhengzhou, Hankou, Shanghai, Nanjing, and Canton. By 1935, more banks were further established across various prefectures; 22% of prefectures had both railroad stations and modern banks. In particular, 35 prefectures¹³ that had headquarters locations and 47 prefectures that had branches¹⁴ (there is overlap between these two groups) also had at least one railroad station.

3.3 Control Variables

As treaty ports represent trade openness, the more prosperous foreign and domestic trade activities, the more banks. Besides, the Chinese Maritime Customs (CMC) is beneficial to the development of modern banks in cities such as Shanghai since it provides good legal protection for public bonds traded by these banks (Ma, 2019).¹⁵ We thus control for these influences by using two measures, one is a dummy variable to indicate whether a prefecture has treaty ports or CMC houses (1 = Yes), and the other is the number of treaty ports or CMC houses a prefecture contains. We also control for population size as a proxy for the development of the local economy, which may be correlated with the distribution of the railroad and banks. Historically, prefectural population data was recorded in 1880, 1910, and 1920, and the population sizes in our model were adjusted to correspond to that data. That is, prefectural population sizes in 1895–1909, 1910–1919, and 1920–1937 are assumed to correspond to population sizes as measured in 1880, 1910, and 1920, respectively. The 1880 and 1910 data were collected from Cao (2001), and the 1920 data from The China Continuation Committee (1922).

In addition, we also control for the area of administration and the length of rivers in a prefecture, and the information was identified based on the China Historical Geographic Information System (CHGIS, V6). It is assumed that the larger the area of a prefecture was, the more modern banks there would be in it. Both variables are taken as the natural logarithm. Last, the impact of regime change is also controlled for, as political turmoil may affect the security of property rights, and inconsistent financial policies between regimes may also lead to changes in banking and finance (Ma, 2019; Du, 2014). We use two dummy variables, the Beiyang regime (1912–1926,1 = Yes) and the Nanjing regime (1927–1948,1 = Yes) to capture the effects. Descriptive statistics of all the variables are reported in Table 1.

[9] The population data of 1880 and 1910 were from Cao (2001), 1920 population was from the China Continuation Committee (1922).

[10] We eliminate the closed banks each year according to Jiang (2014) during sample periods. According to Liu (2007), there were at least 1038 modern bank locations (including head offices and branches) in 1934.

[11] We identify 68 railroads. There were 16 main lines that are usually illustrated on historical maps, including Jinshan Railroad, Tongpu Railroad,Pinghan Railroad, Pingsui Railroad,Zhengtai Railroad, Huning Railroad,Jinpu Railroad, Zhegan Railroad, Dianyue Railroad, Zhangxia Railroad, Gebishi Railroad, Yuehan Railroad, Jiaoji Railroad, Sujia Railroad, Daoqing Railroad, and Longhai Railroad.

[12] Tianjin prefecture was one of the most important cargo distribution centers and had approximately 17 warehouses established by the modern banks. See Tianjin Municipal Archives, 1997, Tianjin Shanghui Danganhuibian:1937-1945 (The Archives of Tianjin Chamber of Commerce, 1937-1945), Tianjin Renmin Chubanshe, Tianjin.

[13] For example, Shuntian (in Zhili province), Songjiang (in Jiangsu province), Tianjin, Hangzhou (in Zhejiang province), Canton (in Guangdo province).

[14] For example, Songjiang, Hanyang (in Hubei province), Tianjin, Canton (in Guangdong province) and Jiangning (Jiangsu province).

[15] The CMC was not only a tax agency but also an infrastructure management entity, for example, building lighthouses, providing aid to navigation, and managing ports. After 1895, The CMC became more profoundly involved in China's foreign debt affairs and helped China's modern banks participate more in the international financial market.

We first examine the relationship between railroads and modern banks, the regression specification is shown as follows:

where Banks_it denotes the number of modern banks in prefecture i in year t. Rail_it is either the railroad dummy or the number of railroad stations in prefecture i in year t. β₁ thus captures the relationship between railroad and modern banks. X_it refers to a vector of local economic characteristics: population size, Western colonial influence (measured by a dummy variable to indicate whether prefecture i has CMC houses or treaty ports in year t), administrative area, and the log of navigation river length. represents the prefecture fixed effect and δt represents the time. Robust standard errors are clustered at the prefecture level.

Columns (1) to (4) in Table 2 reports the basic results. Columns (1) and (2) are estimates using the railroad dummy, and columns (3) and (4) are estimates using the number of railroad stations. Besides, columns (1) and (3) only control railroad variables and fixed effect, while columns (2) and (4) include all control variables, where the effect of Western colonial influence is depicted by different measures. The results indicate that the number of modern banks is significantly associated with railroad construction. For example, after controlling other controls, a prefecture has 1.126 more banks if it has a station (or around 110% more relative to the mean number of modern banks, 1.025) or a prefecture has 0.369 more banks if it has one more station. Given that the prefecture had more opportunities to connect with other domestic or international markets after railroad construction, a railroad station was more likely to become a cargo distribution center, and more financial services were demanded there. Modern banks were incentive to place branches and offices near train stations to facilitate firm financing and mortgage freight management. This relationship corresponds to growth statistics found in historical materials. For example, Yan (2007) documented that freight transported by railroads increased from 2.43 billion ton-kilometers in 1912 in China to 6.27 billion ton-kilometers in 1936.

However, the significant correlation between railroad construction and modern banks cannot be regarded as proving causality, due to possible reverse causation and omitted variables. There is no obvious reverse causation issue between railroad construction and the development of modern banking. However, omitted variable issues are unavoidable. For example, prefectures that were richer or better endowed with natural resources were usually more attractive for locating banks and building railroad stations. We are unable to capture such an effect due to data limitations. Moreover, the expansion of railroads increases the demand for financial services, such as exchange, remittance, and financing, and in turn, stimulates the building of branches. For example, The Shanghai Commercial & Savings Bank, one of the ‘Southern three banks,’ had its headquarters in Shanghai. It established an office at Bengbu station and provided exchange service for the Mobil Company and other foreign companies. It also located branches and offices near the Hankou and Zhengzhou train stations (Ma, 2015). Therefore, we examine the heterogeneous effect of railroad construction on headquarters and branch locations after solving the endogeneity issue in the next section.

4.1 Causality between Establishment of Railroads and Banks

We apply two-stage least squares (2SLS) and use the least-cost paths (LCP) of railroad as the instrumental variable following existing studies (Faber 2014; Jedwab and Moradi, 2017; Fenske and Kala, 2018). The main idea of constructing LCP as an instrumental variable is to find the lowest total cost path of railroad placement between two spots which are recognized as the nodes of the railroad network. Without considering other influences, we could calculate the construction costs of all possible railroad placement schemes between the two spots, finding the lowest total cost path (Faber, 2014). Assuming that the railroad planner is a rational decision-maker, it would prefer the lowest total cost path option, other things being equal. Any prefecture that is geographically located on the lowest total cost path of the two spots has a high probability of having railroad stations. A dummy variable, LCP, is thus constructed to indicate whether a prefecture is located on the lowest total cost path of two railroad network nodes. As the lowest total cost path is objectively determined by the algorithm, LCP meets the requirements for instrumental variables.

There are two considerations for railroad placements in China: one is to ensure national security¹⁶ and the other to connect important commercial cities, which help to decide the nodes of the railroad network. Based on this, we finally identified 34 nodes, of which 18 were the main military couriers and the others the transportation hubs since the Qing Dynasty.¹⁷ We then use ArcGIS10.2 to calculate the lowest total cost path between any two of these 34 nodes based on the topographical and hydrological information recorded in the China Historical Geography Information System (CHGIS V4, 2007). ¹⁸ Fig. 1 presents the lowest total cost path, and the distribution of the paths is basically consistent with that of the actual railroads. We then construct the dummy variable, LCP, to identify whether a prefecture is located on these lowest total cost paths.

The 2SLS specification is as follows:

where LCP_it represents the dummy variable to indicate whether prefecture i is located on the lowest total cost paths in year t, and is the estimation of railroad access. Other variables are the same as those in Eq. (1); β₂ captures the effect of railroad access on the number of modern banks, and β₃ captures the effect of LCP on railroad access.

Columns (5) to (8) in Table 2 reports the results. The Kleibergen and Paap F-statistics reported in columns (5) and (7) are significantly greater than 10, confirming the hypothesis that the LCP is not a weak instrument. As shown in Table 2, the results estimated by 2SLS are consistent with those estimated by fixed effect model, implying that railroad construction affects the development of the modern banking industry. More specifically, a prefecture has 3.063 more banks when it has a station (or around 299% more relative to the mean number of modern banks, 1.025), as shown in column (5); a prefecture has 0.248 more banks (or around 24% more relative to the mean number of modern banks), as shown in column (7). Compared with road transportation, railroads have higher capacity and lower costs. The more stations that are established, the greater the potential for trade and the higher the demand for trade settlement. The need for warehouse and exchange near stations is even greater than in other places, and the modern banks can meet the need through establishing bank branches.

Branches are institutions that provide specific financial services. The impacts of railroad construction on the headquarters and branches might be different. We therefore divide the sample into two for regression. As shown in Table 3, with railroad access, a prefecture has 1.352 more headquarters (or around 286% more relative to the mean number of modern banks,0.472) and 1.712 branches (or 310% more relative to the mean number of modern banks, 0.552). In terms of the number of stations, a prefecture can be predicted to have 24% more bank locations (10% more headquarters and 14% more branches) per additional station. The results show that railroad construction significantly promotes not only the creation of headquarters but also the expansion of branches, and the impact on branches is greater than that on headquarters. This suggests that the demand for specific financial services near the railroad stations promotes the establishment of branches.

During the period of construction, bankers usually have expectations about the future benefits brought by opening of the railroads and thus are likely to locate in advanced headquarters or branches in areas where stations are to be opened. To examine the expected effect of railroad additions, we regressed on the railroad variables in period t + 1¹⁹ and the results are reported in Table 4. All the estimates are in line with those in Table 3, implying the expectation that railroad stations promoted the development of modern banks. This finding suggests that railroad placement is likely to bring about greater business opportunities attractive to banks in the local market nearby (Rawski, 1989; Cheng, 2003).

4.2 Placebo Tests from Planned but Unconstructed Rail Lines

The previous section explores how the development of the banking industry was affected not only by completely built railroads but also railroads while under construction. In this section, we investigate the effects of ‘placebo’ railroad lines; that is, those planned to be constructed but never actually built.

In June 1919, Dr. Sun Yat-sen, the pioneer of the democratic revolution in the late Qing dynasty period who became interim president of the Republic of China in 1912, published the International Development of China, which put forward a long-term plan for the development of China's economy. The plan was a collection of Dr. Sun's thoughts on the modernization of China's economy, specifically in terms of industry, agriculture, and transportation. The plan focused on the construction of ports and railroads,²⁰ with Dr. Sun suggesting six major railroad systems including 106 lines across the country, extending over some 100,000 kilometers. In 1920, this work was widely publicized across China. Since railroad construction required a vast amount of money, Dr. Sun's solution was to borrow overseas. He sent the plan to ambassadors of the United States, France, Italy, and the United Kingdom. However, the proposed railroad construction plan went unrealized partly due to its size and financial requirements, and Dr. Sun’s resignation.

But the proposed lines in Dr. Sun's plan were useful for placebo testing. We sorted out the proposed Central and Southeastern railroad's route network, manually identified the locations of ‘placebo’ railroad stations, and matched them into 1911-based map data.²¹ There were 31 main railroad lines²² that were perceived at the time as likely to be built, extending into 105 prefectures in 17 provinces, serving as our ‘placebos.’

Referring to the design of Donaldson (2018), the regression can be specified as:

where Railplan_i is a dummy variable that indicates whether a prefecture has planned but eventually not constructed railroad stations (1 = Yes); Post1919_t is a dummy variable to indicate whether the year of data is after 1919 (similarly, Post1920_t equals 1 after 1920). The coefficients θ₁ and θ₂ capture the effects of the planned but unbuilt stations on the modern banks after 1919 (or 1920). Theoretically, we should not expect to see any significant impacts from these planned but unconstructed railroad stations on banks.

Table 5 presents the results, showing that the coefficients of these placebo railroads are statistically insignificantly different from zero. Meanwhile, the coefficients of the built railroads were not significantly changed by the inclusion of these placebo variables. We, therefore, confirm that the impact of railroads on the modern banking industry is not driven by accidental factors.

4.3 The roles of concessions and hubs on bank expansion

As mentioned in section 2, in foreign concessions, extra-territorial rights were retained by foreign powers. Such a regime was able to provide relatively stable property rights protection in the face of a larger national environment of political instability. In this way, railroad placement in a prefecture that had concessions was typically more conducive to the development of modern banks than in a prefecture with no concessions, which is confirmed in Panel A of Table 6. Regardless of whether a prefecture had concessions, railroad placement still significantly promoted the development of modern banks. It is just that the effect in concessionary prefectures is greater than that in non-concessionary ones.

A possible explanation is that the railroads increased market accessibility in any case. For bankers, the profit potential of the more extensive market facilitated by railroads may have offset any concerns about a potentially precarious operational environment in terms of property rights protection. For example, Ma (2019) shows that China's financial reform represented by Shanghai during the Republic of China mainly benefitted from institutional protections provided by public concessions and CMC. In considering that the legal protection effects of the concession are centered on consular courts but can be expanded up to 400 kilometers (Keller et al.,2021), it is worth emphasizing that the location selection of headquarters has been typically concerned with property rights protection, while that of branches is more based on potential profits from accessible customers. So, the effect of railroad construction in a prefecture with a concession includes not only property rights protection but also market access. For the prefecture without a concession, the impact of railroad construction on branches is greater than on headquarters.

A station is identified as a hub if it is a terminal of a line or an intersection station between two lines. Those transportation hubs are usually freight distribution centers that require vast capital for freight flow, driving more financing needs and in turn likely to thus promote the growth of modern banks. The results in Panel B support the perspective that hub stations are conducive to the development of modern banks. In particular, hub stations significantly promoted the establishment of headquarters, on which non-hub stations had no significant effects. Both hub and non-hub stations significantly promoted branch expansions. This is likely due to the different functions of headquarters and branches, as headquarters centrally gather and distribute financial resources, while branches expand the retail business.

Finally, we discuss the role of regional trade potential. We use a dummy variable indicating whether a prefecture had a significant silk or tea (both important trading goods for China in the 20th century as well as earlier ones) production area as a proxy for trade potential, due to data availability limitations. The results of Panel C in Table 6 suggest that the impact of railroad construction on the development of modern banks is associated with regional trade potential.

4.4 Mechanisms

In this section, we explore two potential mechanisms for promoting bank growth: modern industry and qianzhuang. The development of modern industry stimulated financial demand and qianzhuang was deeply involved in trade settlements due to the increasing trade promoted by railroads.

To explore the potential mechanism by which modern industry may have promoted bank growth, we manually collected information on industrial firms from Du (2014), which recorded the year of establishment and the location of each firm in the 1840–1937 period in China and can be used to calculate annual establishment of new firms by prefecture. Firms in the following six categories were included and delineated: agricultural, mining, manufacturing, construction, commercial, and electrical industries. We use the annually accumulated numbers of firms as the mechanism variables. Panel A in Table 7 report the results. The effect of railroads on industrial firms and the effect of such firms on modern banks are quite remarkable, both in headquarters and branches, which implying these firms are a potential mechanism through which railroads may promote the development of modern banks.

The result has some implications. The operation of railroads overcame distance barriers that hindered large-scale freight transportation and slashed transit costs, facilitating the delivery and purchase of raw materials, fuels, and products. Trade and business flourished, and modern industries spread and established factories along the railroads. However, many modern industries ran out of funds after purchasing equipment or lacked adequate raw materials procurement.²³ The Dasheng Cotton Mill factory, for example, continuously sought loans from all available financial institutions since the early period of its establishment. ²⁴Modern banks, such as The Shanghai Commercial & Savings Bank and The Kincheng Banks, sought such opportunities and introduced the mortgage loan that soon became one of the most popular transaction forms of early 20th-century Chinese financial markets (Cheng, 2021).

Next, we examine the mechanism of qianzhuang, and Panel B in Table 7 report the estimates. The results of the first-stage estimation show that a prefecture had 0.72 more qianzhuang when it had railroad access and 0.05 more qianzhuang per additional railroad station. In section 2.2, we have mentioned that traditional local banks, mainly qianzhuang, provided short-term microloans to local merchants and enterprises. Qianzhuang had relatively small amounts of capital and generally provided services such as currency exchange, issuance of commercial paper, and credit lending (Zhang, 1989). Our results imply that it is through qianzhuang, that the branches of modern banks met the growing need for trade financing after railroad construction.

From the above perspective, there are three possible explanations why the presence of qianzhuang might be another potential mechanism leading to the development of the modern banking industry. First, modern banks and qianzhuang were complementary in business to some extent. The establishment of railroads increased cross-regional trade and the distribution of products and raw materials from rural areas, prominently, raw silk and tobacco, which made them begin to enter the foreign trade network. The increase in trade volume required financial institutions to develop new payment methods, optimize credit issuance, and increase capital supply (Rawski, 1989). Thus, many modern banks began to establish branches in areas far from commercial centers. Qianzhuang was familiar with local business environments while modern banks could compensate for the weakness of the low-capitalqianzhuang. Therefore, it became the norm for modern banks and qianzhuang to jointly provide loans to industrial enterprises.²⁵

Second, qianzhuang's advantages in clearing and settlement also encouraged modern banks to cooperate with them. Since the opening of a series of treaty ports after the Opium War, the local currency (e.g., zhuangpiao) issued by qianzhuang at those locations became widely accepted by foreign merchants and became the main medium for foreign trade settlement. In addition, the currency used between the treaty ports and the mainland became the same currency in vogue for commercial transactions in major commercial cities. Qianzhuang was therefore closely related to both domestic and foreign trade. The development of trade greatly increased the demand for qianzhuang to use financial instruments such as zhuangpiao (cashier’s check issued by qianzhuang) and commercial acceptance bills for clearing and settlement. Bill settlement by domestic banks also relied more on qianzhuang to be completed before the establishment of the Shanghai Bills Exchange (by banks) in 1933. For example, the qianzhuang industry in Shanghai established the Shanghai Remittance and Exchange Association in 1890, which became the only intermediary for trade settlement. The railroad connected the treaty ports and the mainland market, promoting foreign trade and in turn the development of qianzhuang. To meet the needs for funds, qianzhuang issued cashier’s checks as credit instruments to establish cooperative relations with foreign banks, piaohao and Chinese-funded banks, forming an interbank lending market.²⁶

[16] The Qing government was defeated in the Sino-French War in 1885, while the efficient and rapid transportation of soldiers was an important driving force for the construction of railroads during the Qing dynasty (Wang et al. 2014).

[17] These military couriers included six Green Standard Army barracks of Qing dynasty and twelve important historical military affairs. The military data came from Luo (2011) and Lai (1988). The total 34 nodes consisted of Jun Xian, Kunming, Shanghai, Qiantang, Fengyang, Jiao Zhou, Tianjin, Licheng, Tongshan , Daxing, Hai Zhou, Zheng Zhou, Haicheng , Nanchang, Changsha , Huolu, Jiujiang , Wanping , Datong, Yangqu, Henei, Hekou, Yin Xian, Jiangning, Wuhu, Jiangpu, Linyu, Baoji, Panyu, Longxi, Xiangtan, Hanyang, Salaqi Ting, and Ruicheng.

[18] More specifically, we divide the historical map into 10*10km grids by ArcGIS, and the software algorithm can sum the weights of the topographic and hydrological indicators of each grid, in turn calculating the lowest total costs of railroad construction between any two grids. For any two nodes, we can find out the lowest total cost path by connecting all the lowest total cost paths of grids between two nodes.

[19] We also conduct regressions on railroad variables in t+2 and t+3, and the results are quite robust. The results are shown in Appendix Table A3 and Table A4.

[20] One of the goals was to build three world-class maritime ports and many other commercial ports within 10 to 20 years.

[21] There were 204 stations in the proposed Southeastern railroad's systems and 174 more in the Central system.

[22] The placebo main railroad lines include ChongQing - Fuzhou Line, Dongfang Dagang- Xinfeng Line, Fuzhou-Zhenjiang Line, Fuzhou - Wuchang Line, Fuzhou - Gulin Line, Wenzhou - Chenzhou Line, Xiamen - Jianchang Line, Xiamen - Canton Line, Shantou - Changde Line, Nanjing - Shaozhou Line, Nanjing - Jiaying Line, Dongfnag Dagang - Nanfang Dagang Line, Jianchang - Yuanzhou Line, Nanjing - Luoyang Line, Nanjing - Hankou Line, Xi`an - Datong Line, Xi`an - Ningxia Line, Xi`an - Hankou Line, Xi`an - Chongqing Line, Lanzhou - Chongqing Line, Beifang Dagang - Xi`an Line, Beifang Dagang - Hankou Line, Huanghegang - Hankou Line, Zhifu - Hankou Line, Haizhou - Jinan Line, Haizhou - Hankou Line, Haizhou - Nanjing Line, Xinyanggang - Nanjing Line, Lusigang - Nanjing Line, Coast Line, Huoshan - Jiaxing Line. See Sun (1920).

[23] Some firms ran out of funds after purchasing equipment, and some firms did not have enough funds in raw materials purchase. See Liu (1940) on page 66-69.

[24] The discussion about shortage of funds in Dasheng (a representative enterprise in early 20th century China) is described in detail by Elisabeth Köll (2004).

[25] Many studies provide evidence of this kind of cooperation, such as Jincheng Yinhang Shilia o(The History of the Kincheng Bank) and Shanghai qianzhuang Shiliao(The History of Shanghai qianzhuang). See Rawski (2009) on page 162.

[26] Interbank lending is a form of credit system in which qianzhuang borrows money from foreign banks through Zhuangpiao (a credit promissory note issued by qianzhuang). Qianzhuang can borrow funds as it needs (usually settled once every two days), but foreign banks can reclaim them at any time.

The modern transportation system represented by railroads has contributed to the economic development of semi-colonial countries like China. In the late 19th through early 20th century China, the emergence of railroads and modern banks were both prominent symbols and results of China's learning and adoption of science and technology from the West. The railroads connected the port cities and the vast hinterland, playing an important role in industrialization. The modern banks contributed to important innovations in the accounting system and management of the modern financial industry.

This study explores the impact of railroads on modern banks by sorting prefecture-level data of railroads and modern banks. In addition, it analyzes the economic effects of transportation infrastructure construction on the banking industry from a quantitative perspective. We also use plausible geographical characteristics to build exogenous instrument variables and estimated causal effects by 2SLS regression. Our study thus provides new evidence for the stimulative effects of modern large-scale transportation infrastructure construction on the rise of the new generations of banks. Industrial development along railroads brought about enormous financial demands and became an important mechanism by which the extension of railroads stimulated the development of the modern banking industry. Meanwhile, the development of trade has also enabled modern banks to develop new business based on cargo mortgages. These results not only provide a new perspective for understanding how the modern transportation system contributed to the rise of modern banks but also enrich the literature on understanding the development of the modern banking industry.

Disclosure statement: The authors declare that we have no relevant or material financial interests that relate to the research described in this paper.

Author Contribution Statement: 1) Jinsong Zhao: Conceptualization, Investigation, Writing- Original draft preparation, Funding acquisition; 2) Peilin He: Data curation, Software, Visualization; 3) Chin-Hsien Yu: Methodology, Supervision, Writing- Reviewing and Editing. All authors reviewed the manuscript.

Atack, J., Bateman, F., Haines, M., & Margo, R. A. (2010). Did railroads induce or following economic growth? Urbanization and population growth in the American Midwest, 1850–1860. Social Science History, 34(2), 171–197.
Atack, J., Jaremski, M., & Rousseau, P. (2014). American Banking and the Transportation Revolution before the Civil War. The Journal of Economic History, 74(4), 943–986.
Banerjee, A., Duflo, E., & Qian, N. (2020). On the road: Access to transportation infrastructure and economic growth in China. Journal of Development Economics, 145, 1–36.
Berger, T., & Enflo, K. (2017). Locomotives of local growth: The short-and long-term impact of railroads in Sweden. Journal of Urban Economics, 98, 124–138.
Bogart, D., & Chaudhary, L. (2013). Engines of growth: the productivity advance of Indian railways, 1874–1912. Journal of Economic History, 73(2), 339–370.
Cantoni, D., & Noam, Y. (2020). Historical Natural Experiments: Bridging Economics and Economic History, NBER Working Papers, No.26754.
Cao, S. (2001). ZhongguoRenKou Shi: Qing Shiqi (A Demographic History of China: Qing Period). Shanghai: Fudan University Press.
Chen, Z., Ma, C., & Sinclair, A. J. (2021). Banking on the Confucian Clan: Why China Developed Financial Markets So Late. The Economic Journal, 132(644), 1378–1413.
Cheng, L. (2003). Banking in Early 20th century China: Entrepreneurs, Professional Managers, and the Development of Chinese Banks, 1897–1937, Published by Cambridge University Press.
Donaldson, D. (2018). Railroads of Raj: Estimating the Impact of Transportation Infrastructure. American Economic Review, 108(4–5), 899–934.
Donaldson, D., & Hornbeck, R. (2016). Railroads and American Economic Growth: A Market Access Approach. The Quarterly Journal of Economics, 131(2), 799–858.
Du, X. (2014). MinzuZibenzhuyi Yu JiuzhongguoZhengfu (1840–1937) (National Capitalism and the Old Chinese Government (pp. 1840–1937). Shanghai: Shanghai Renmin Chubanshe.
Du, X. (2015). The Financial Development Models and the Transforming Society of Early 20th century China, ZhongguoJingjishiYanjiu (Research in Chinese Economic History), 3, 14–24 + 143.
Du, X. (2019). Zhongguo De MinzuZibenzhuyi 1927–1937 (The National capitalism in China 1927–1937), Shanghai University of Finance and Economics Press, Shanghai.
Köll, E. (2004). From Cotton Mill to Business Empire: The Emergence of Regional Enterprises in Early 20th century China. Cambridge, Mass.: Harvard University Press.
Köll, E. (2019). RailRoads and the Transformation of China. Cambridge, Mass.Harvard University Press.
Faber, B., & Integration, T. (2014). Market size, and Industrialization: Evidence from China`s National Trunk Highways system,Review of Economic Studies, 81(3),1046–1070.
Fenske, K., J. and, & Namrata (2018). Linguistic Distance and Market Integration in India. The Journal of Economic History, 78, 619–619.
Fogel, R. W. (1964). Railroads and American Economic Growth: Essays in Econometric History, Published by Baltimore. Johns Hopkins University Press.
Fu, R. (1963). ZhongguoJindaiTielushiZiliao (Quan sance) (The Materials of Railway History in Early 20th century China), ZhonghuaShuju, Beijing.
Fu, R. (1980). Imperialism and China`s Railway (pp. 1847–1949). Shanghai: Shanghai Renmin Chubanshe.
Greenwood, J., & Jovanovic, B. (1990). Financial Development, Growth, and the Distribution of Income. Journal of Political Economy, 98(5), 1076–1107.
Harvard Yenching Library. China Historical Geographic Information System. Cambridge, MA: China Historical GIS (2016). Available at i>http://www.fas.harvard.edu/chgis/
Hong, J. (1993). ZhongguoJinrongshi (The History of Chinese Finance). Sichuan: Southwest University of Finance and Economics Press.
Jedwab, R., Kerby, E., & Moradi, A. (2017). History, Path Dependence and Development: Evidence from Colonial Railways, Settlers and Cities in Kenya. The Economic Journal, 127(603), 1467–1494.
Jiang, J. (2014). JindaiZhongguoYinhangyeRenmingDacidian (the Dictionary of Banking Institutions in Early 20th century China). Shanghai: Shanghai GujiChubanshe.
Lai, X. (1988). Beiyang Junfa. Shangai: Beiyang Warlords), Shanghai renminchunbanshe.
Lan, R. (2013). A Study on the Characteristics of Chinese Modern Banking Institution Change: Based on the Banking Institution Structure. Lanzhou Xuekan, 5, 47–51.
Levine, R., & Zervos, S. (1998). Stock Markets, Banks, and Economic Growth. American Economic Review, 88(3), 537–558.
Levine, R., Lin, C., Ma, C., & Xu, Y. (2021). The Legal Origins of Financial Development: Evidence from the Shanghai Concessions, NBER Working Papers 28794, National Bureau of Economic Research, Inc.
Lin, C., Ma, C., Sun, Y., & Xu, Y. (2021). The Telegraph and Modern Banking Development, 1881–1936. Journal of Financial Economics, 141(2), 730–749.
Li, N. (2010). Railway Development and Migration: A Natural Experiment Evidence in Manchuria of China during 1891–1935. ZhongguoRenkouKexue (Chinese Journal of Population Science), 4, 54–66.
Li, Z. (1994). ZhongguoTielushi (The History of Chinese Railroads). Shantou: Shantou University Press.
Liang, R. (2015). Treaty Ports, Railroad, and Industrialization in Early 20th century China, JingjiYanjiu. (Economic Research Journal), 4, 178–191.
Liu, D. (1937). ZhongguoGongyeDiaochaBaogao (Chinese Industrial Survey Report). Shanghai: Shanghai Jingji Tongji Yanjiusuo.
Liu, D. (1940). Shanghai GongyehuaYanjiu. Shanghai: The Commercial Press. The industrialization of Shanghai.
Liu, K. (2007). Restatistics of Chinese Modern banks in 1927–1937, ZhongguoJingjishiYanjiu (Research in Chinese Economic History), 1, 48–59.
Luo, E. (2011). Luying Bingzhi (The Biography of Green Standard Army). Shanghai: The Commercial Press.
Ma, D. (2019). Financial Revolution in Republican China During 1900–37: A Survey and a New Interpretation. Australian Economic History Review, 59(3), 242–262.
Ma, Q. (1983). ZhongguoTieluJianzhuBiannianJianshi 1881–1981 (A Brief History of Railway Construction in China in 1881–1981), ZhongguoTieluChubanshe, Beijing.
Nunn, N., & Nancy, Q. (2011). The Potato's Contribution to Population and Urbanization: Evidence from a Historical Experiment. The Quarterly Journal of Economics, 126(2), 593–650.
O’Brien and Patrick (1977). The New Economic History of the Railways. Published by London: Croom Helm.
Patrick, H. T. (1966). Financial Development and Economic Growth in Underdeveloped Countries. Economic Development and Cultural Change, 14(2), 174–189.
Rawski, T. G. (1989). Economic Growth in Prewar China. Berkeley, CA: University of California Press.
Robinson, J. (1953). The Production Function and the Theory of Capital. The Review of Economic Studies, 21(2), 81–106.
Robinson, J., James, A., & Diamond, J. (2010). Natural Experiments of History, Published by Cambridge, MA,Harvard University Press.
Rajan, R. G., & Zingales, L. (1998). Which capitalism? Lessons form the East Asian Crisis. Journal of Applied Corporate Finance, 11, 40–48.
Sun, Y. (1920). International Development of China. Shanghai: The Commercial Press.
Tan, Q. (1996). ZhongguoLishiDituji: Qingdai Juan. Qing Period), ZhongguoDituChubansh, Beijing: Historical Map of China.
Tianjin, M. A. (1997). Tianjin Shanghui Danganhuibian:1937–1945 (The Archives of Tianjin Chamber of Commerce in 1937–1945). Tianjin: Tianjin Renmin Chubanshe.
Wang, H., Liu, C., & Yan, S. (2014). Impact of Railroad Construction in Late Qing Dynasty and Early Republic China on Long-term Economic Development, JingjiXuebao, 3, 1–20.
Wolfgang Keller, C. H., Shiue, & Wang, X. (2021). Capital Market Development in China and Britain, 18th and 19th Century: Evidence from Grain Prices. American Economic Journal: Applied Economics.
Xu, D., & Wu, C. (2003). ZhongguoZibenZhuyiFazhanshi (Diyijuan) (History of Chinese Capitalism I). Beijing: Renmin Chubanshe.
Yan, H. (2015). Government Debts and Financial Development in Early 20th century China. CaijingYanjiu (Journal of Finance and Economics), 9, 108–120.
Yan, H. (2020). The Financial History of China. Shanghai: Shanghai University of Finance and Economics Press.
Yan, S., & Xu, M. (2015). Railways and Domestic Market Development in the Late Qing Dynasty, JingjiXue (Jikan) (China Economic Quarterly), 2015, 14, 779–800.
Yan, Z. (1955). ZhongguoJindaiJingjishiZiliaoXuanji (Economic History in Early 20th century China). Beijing: KexueChubanshe.
Yearbook Compilation Committee of Railway Ministry of Nanjing Government (1933). TiedaoNianjian (the Railroad Yearbook), Shanghai HanwenZhengkaiYinshuju, Shanghai.
Yin, T. (2005). WanqingTielu Yu WanqingShehuiBianqianYanjiu (Railway and Social Change in late Qing Dynasty), JingjiKexueChubanshe, Beijing.
Zhang, Z., Xiong, Y., & Shen, Z. (2002). Changjiang YanjiangChengshi Yu ZhongguoJindaihua. Shanghai Renmin Chubanshe, Shanghai: Cities along the Yangtze River and China's modernization).
Zhang, Y. (1997). A BriefHistory of Railway Construction in China, ZhongguoTiedaoChubanshe, Beijing.
Zhang, R. (2020). Railroads in Early 20th century China: Political Aspects of Railroad Administration, 1876–1937. Beijing: ZhonghuaShuju.
Zhao, J., & Pang, H. (2018). Surviving Unstable Property Rights in Early Early 20th century China: A Case Study of Young Brother Bank. Frontiers of Economics in China, 13, 505–530.
Zheng, B. (2000). MinguoShiiZhengQuYange (The evolution of political districts during the Republic of China). Hubei: Hubei JiaoyuChubanshe.
Zhou, J., & Zeng, J. (2020). The Substitution of Banks for piaohao and Late Qing and Early Republican Financial Reform, ZhongguoShehuiKexue (Social Sciences in China), 8, 178–203 + 208.

Table 1

Summary Statistics
VARIABLES	Obs.	Mean	Sd. Dev.	Min.	Max.
The number of banks	10750	1.025	5.363	0	135
The number of headquarters	10750	0.472	2.903	0	76
The number of branches	10750	0.552	2.676	0	59
Whether has a station (1 = Yes)	10750	0.195	0.396	0	1
Number of stations	10750	2.445	8.307	0	114
Whether has treaty ports or customs (1 = Yes)	10750	0.151	0.358	0	1
Number of treaty ports or customs	10750	0.176	0.525	0	7
Population size (log)	10750	4.629	0.985	1.526	7.141
Length of navigational rivers (log)	10750	3.799	2.665	0	7.659
Area of administration	10750	1.368	1.180	0.085	13.38
Modern Firms	10750	4.440	29.60	0	992
Qianzhuang	10750	1.561	7.938	0	155

Table 2

The Basic Results
	Fixed Effect Model Estimation				2SLS Estimation
	Banks	Banks	Banks	Banks	Banks	Whether has a station (1 = Yes)	Banks	Number of stations
VARIABLES	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)
Whether has a station (1 = Yes)	1.256***	1.126***			3.063**
	(0.408)	(0.385)			(1.238)
Number of stations			0.347***	0.369***			0.248**
			(0.103)	(0.123)			(0.102)
LCP						0.645***		6.340***
						(0.050)		(0.757)
Whether has treaty ports or customs		1.149			0.454	0.143***
(1 = Yes)		(0.756)			(1.009)	(0.039)
Number of treaty ports or customs				-0.966			0.613	11.598**
				(1.359)			(1.251)	(4.914)
Population size		0.616		0.603	0.561	0.008	0.623	-0.035
		(0.570)		(0.432)	(0.563)	(0.025)	(0.443)	(0.702)
Area of administration		1.959**		1.940***	1.878**	0.012	1.969***	-0.052
		(0.839)		(0.635)	(0.828)	(0.037)	(0.652)	(1.033)
Length of navigational rivers		-0.411*		-0.406**	-0.390*	-0.003	-0.414**	0.014
		(0.220)		(0.166)	(0.217)	(0.010)	(0.171)	(0.270)
Beiyang regime (1 = Yes)		1.313**		0.581**	0.905***	0.061***	0.789***	0.228
		(0.533)		(0.283)	(0.343)	(0.017)	(0.284)	(0.534)
Nanjing regime (1 = Yes)		2.095***		0.933**	1.491***	0.102***	1.294***	0.909
		(0.650)		(0.371)	(0.387)	(0.021)	(0.315)	(0.701)
Constant	-0.620**	-4.633	-0.058	-2.901*	-2.917	-0.074	-3.100*	0.150
	(0.251)	(2.819)	(0.124)	(1.616)	(2.079)	(0.097)	(1.636)	(2.382)
Prefecture FE	Y	Y	Y	Y	Y	Y	Y	Y
Year FE	Y	Y	Y	Y	Y	Y	Y	Y
Observations	10,750	10,750	10,750	10,750	10,750	10,750	10,750	10,750
R-squared	0.538	0.539	0.610	0.611	0.520	0.837	0.593	0.846
KP F					163.875		70.118

Robust standard errors in parentheses are clustered at the prefecture level. *** p < 0.01, ** p < 0.05, * p < 0.1.

Table 3

The Effects on Headquarters and Branches
	2SLS Estimation
VARIABLES	Headquarters	Whether has a station (1 = Yes)	Headquarters	Number of stations	Branches	Whether has a station (1 = Yes)	Branches	Number of stations
	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)
Whether has a station (1 = Yes)	1.352**				1.712**
	(0.602)				(0.675)
Number of stations			0.108**				0.140**
			(0.051)				(0.055)
LCP		0.645***		6.340***		0.645***		6.340***
		(0.050)		(0.757)		(0.050)		(0.757)
Whether has treaty ports or	0.363	0.143***			0.091	0.143***
customs (1 = Yes)	(0.588)	(0.039)			(0.551)	(0.039)
Number of treaty ports or customs			0.450	11.598**			0.163	11.598**
			(0.599)	(4.914)			(0.734)	(4.914)
Population size	0.444	0.008	0.471*	-0.035	0.117	0.008	0.152	-0.035
	(0.301)	(0.025)	(0.253)	(0.702)	(0.311)	(0.025)	(0.251)	(0.702)
Area of administration	0.274	0.012	0.314	-0.052	1.604***	0.012	1.655***	-0.052
	(0.443)	(0.037)	(0.373)	(1.033)	(0.457)	(0.037)	(0.370)	(1.033)
Length of navigational rivers	-0.153	-0.003	-0.164*	0.014	-0.237**	-0.003	-0.250***	0.014
	(0.116)	(0.010)	(0.098)	(0.270)	(0.120)	(0.010)	(0.097)	(0.270)
Beiyang regime (1 = Yes)	0.385*	0.061***	0.331*	0.228	0.520***	0.061***	0.459***	0.228
	(0.205)	(0.017)	(0.171)	(0.534)	(0.154)	(0.017)	(0.133)	(0.534)
Nanjing regime (1 = Yes)	0.689***	0.102***	0.599***	0.909	0.802***	0.102***	0.696***	0.909
	(0.248)	(0.021)	(0.212)	(0.701)	(0.165)	(0.021)	(0.138)	(0.701)
Constant	-1.480	-0.074	-1.559*	0.150	-1.436	-0.074	-1.541	0.150
	(1.094)	(0.097)	(0.925)	(2.382)	(1.163)	(0.097)	(0.938)	(2.382)
Prefecture FE	Y	Y	Y	Y	Y	Y	Y	Y
Year FE	Y	Y	Y	Y	Y	Y	Y	Y
Observations	10,750	10,750	10,750	10,750	10,750	10,750	10,750	10,750
R-squared	0.474	0.837	0.534	0.846	0.573	0.837	0.644	0.846
KP F	163.875		70.118		163.875		70.118

Robust standard errors in parentheses are clustered at the prefecture level. *** p < 0.01, ** p < 0.05, * p < 0.1.

Table 4

The Effects of Under Constructing Lines
	2nd -Stage Results of 2SLS Estimation
	Banks		Headquarters		Branches
VARIABLES	(1)	(2)	(3)	(4)	(5)	(6)
Whether has a station (1 = Yes) t + 1	3.344**		1.448**		1.896**
	(1.384)		(0.667)		(0.759)
Number of stations		0.257**		0.109**		0.148**
		(0.108)		(0.054)		(0.059)
Whether has treaty ports or customs (1 = Yes)	0.408		0.364		0.044
	(1.046)		(0.599)		(0.571)
Number of treaty ports or customs		0.443		0.414		0.030
		(1.347)		(0.631)		(0.798)
Population size	0.516	0.591	0.419	0.452*	0.096	0.139
	(0.559)	(0.444)	(0.298)	(0.253)	(0.311)	(0.254)
Area of administration	1.596**	1.699***	0.182	0.227	1.414***	1.472***
	(0.760)	(0.604)	(0.405)	(0.344)	(0.423)	(0.345)
Length of navigational rivers	-0.344*	-0.372**	-0.137	-0.149	-0.207*	-0.223**
	(0.206)	(0.164)	(0.110)	(0.093)	(0.115)	(0.094)
Beiyang regime (1 = Yes)	0.847**	0.774***	0.365*	0.329*	0.482***	0.445***
	(0.331)	(0.276)	(0.202)	(0.169)	(0.145)	(0.128)
Nanjing regime (1 = Yes)	1.364***	1.237***	0.679***	0.619***	0.685***	0.618***
	(0.376)	(0.313)	(0.249)	(0.218)	(0.156)	(0.133)
Prefecture FE	Y	Y	Y	Y	Y	Y
Year FE	Y	Y	Y	Y	Y	Y
Observations	10,500	10,500	10,500	10,500	10,500	10,500
R-squared	0.508	0.585	0.462	0.525	0.549	0.627
KP F	151.273	67.229	151.273	67.229	151.273	67.229

The first stage results can be found in Appendix. Robust standard errors in parentheses are clustered at the prefecture level. *** p < 0.01, ** p < 0.05, * p < 0.1.

Table 5

Placebo Tests
	2nd -Stage Results of 2SLS Estimation
	Banks		Headquarters		Branches
VARIABLES	(1)	(2)	(3)	(4)	(5)	(6)
Panel A: Post year = 1919
Whether has a station (1 = Yes)	3.036**		1.336**		1.700**
	(1.250)		(0.611)		(0.678)
Number of stations		0.247**		0.107**		0.139**
		(0.103)		(0.052)		(0.056)
Whether has a station planned but unbuilt (1 = Yes)	-0.434	-0.320	-0.173	-0.120	-0.261	-0.199
	(0.340)	(0.316)	(0.192)	(0.182)	(0.159)	(0.146)
Whether has a station planned but unbuilt (1 = Yes)	0.589	0.287	0.333	0.195	0.256	0.093
# post1919 (1 = after 1919)	(0.841)	(0.798)	(0.471)	(0.457)	(0.406)	(0.379)
R-squared	0.521	0.593	0.475	0.535	0.562	0.634
KP F	164.471	69.296	164.471	69.296	164.471	69.296
Panel B: Post year = 1920
Whether has a station (1 = Yes)	3.040**		1.339**		1.701**
	(1.249)		(0.610)		(0.678)
Number of stations		0.247**		0.107**		0.139**
		(0.103)		(0.052)		(0.056)
Whether has a station planned but unbuilt (1 = Yes)	-0.410	-0.308	-0.156	-0.109	-0.254*	-0.199
	(0.323)	(0.300)	(0.184)	(0.175)	(0.149)	(0.137)
Whether has a station planned but unbuilt (1 = Yes)	0.564	0.276	0.311	0.178	0.254	0.098
# post1920 (1 = after 1920)	(0.849)	(0.806)	(0.480)	(0.465)	(0.406)	(0.379)
Controls	Y	Y	Y	Y	Y	Y
Prefecture FE	Y	Y	Y	Y	Y	Y
Year FE	Y	Y	Y	Y	Y	Y
Observations	10,750	10,750	10,750	10,750	10,750	10,750
R-squared	0.521	0.593	0.475	0.535	0.562	0.634
KP F	164.341	69.410	164.341	69.410	164.341	69.410

The first stage results can be found in Appendix. Robust standard errors in parentheses are clustered at the prefecture level. *** p < 0.01, ** p < 0.05, * p < 0.1.

Table 6

The Role of Concessions, Hubs, and Silk/tea
	2nd -Stage Results of 2SLS Estimation
	Banks		Headquarters		Branches
VARIABLES	(1)	(2)	(3)	(4)	(5)	(6)
Panel A: whether a prefecture has concessions
Has a station × Has concessions	20.448**		9.533**		10.916**
	(9.909)		(4.827)		(5.212)
Has a station × Has no concessions	0.863**		0.322		0.541**
	(0.364)		(0.228)		(0.213)
Number of stations × Has concessions		0.770**		0.357**		0.413**
		(0.318)		(0.159)		(0.166)
Number of stations × Has no concessions		0.159**		0.066*		0.093***
		(0.063)		(0.035)		(0.034)
R-squared	0.482	0.574	0.452	0.529	0.508	0.600
KP F	14.292	12.569	14.292	12.569	14.292	12.569
Panel B: whether a prefecture has hubs
Has a station × Has hubs	9.284**		4.350**		4.934**
	(4.040)		(1.984)		(2.172)
Has a station × Has no hubs	0.578		0.165		0.413**
	(0.373)		(0.210)		(0.188)
Number of stations × Has hubs		0.489**		0.229**		0.260**
		(0.210)		(0.108)		(0.111)
Number of stations × Has no hubs		0.095**		0.032		0.063***
		(0.043)		(0.024)		(0.023)
R-squared	0.516	0.609	0.472	0.557	0.554	0.640
KP F	176.203	122.698	176.203	122.698	176.203	122.698
Panel C: whether a prefecture has silk or tea
Has a station × Has silk or tea	4.371**		2.075**		2.296**
	(1.843)		(0.945)		(0.956)
Has a station × Has no silk or tea	1.639		0.566		1.073
	(1.467)		(0.616)		(0.877)
Number of stations × Has silk or tea		0.370**		0.177**		0.194**
		(0.162)		(0.087)		(0.082)
Number of stations × Has no silk or tea		0.153		0.056		0.098
		(0.111)		(0.048)		(0.066)
R-squared	0.520	0.593	0.473	0.533	0.560	0.634
Control	Y	Y	Y	Y	Y	Y
Prefecture FE	Y	Y	Y	Y	Y	Y
Year FE	Y	Y	Y	Y	Y	Y
Observations	10,750	10,750	10,750	10,750	10,750	10,750
KP F	93.763	87.556	93.763	87.556	93.763	87.556

The first stage results can be found in Appendix. Robust standard errors in parentheses are clustered at the prefecture level. *** p < 0.01, ** p < 0.05, * p < 0.1.

Table 7

Mechanism Tests
							1st -Stage Results
	Bank		Headquarters		Branches		Mechanism	Mechanism
VARIABLES	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)
Panel A: Mechanism is Number of Modern Firms (in logarithm)
Whether has a station (1 = Yes)	1.100**		0.266		0.834**		12.503**
	(0.528)		(0.268)		(0.357)		(6.075)
Number of stations		0.091*		0.020		0.071**		1.027**
		(0.050)		(0.028)		(0.032)		(0.512)
Mechanism	0.157***	0.152***	0.087***	0.086***	0.070***	0.067***
	(0.010)	(0.008)	(0.005)	(0.004)	(0.006)	(0.005)
R-squared	0.897	0.909	0.862	0.867	0.866	0.882	0.512	0.548
KP F	162.983	68.323	162.983	68.323	162.983	68.323	163.875	70.118
Panel B: Mechanism is Number of qianzhuang (in logarithm)
Whether has a station (1 = Yes)	2.042**		0.807*		1.235**		0.381***
	(0.948)		(0.460)		(0.527)		(0.120)
Number of stations		0.173**		0.067		0.106**		0.032***
		(0.084)		(0.042)		(0.047)		(0.012)
Mechanism	2.684***	2.350***	1.431***	1.285***	1.252***	1.065***
	(0.776)	(0.704)	(0.421)	(0.395)	(0.397)	(0.353)
R-squared	0.556	0.606	0.507	0.544	0.593	0.647	0.782	0.785
Control	Y	Y	Y	Y	Y	Y	Y	Y
Prefecture FE	Y	Y	Y	Y	Y	Y	Y	Y
Year FE	Y	Y	Y	Y	Y	Y	Y	Y
Observations	10,750	10,750	10,750	10,750	10,750	10,750	10,750	10,750
KP F	148.806	60.875	148.806	60.875	148.806	60.875	163.875	70.118

No competing interests reported.

Appendix.docx

Download PDF

Version 1

posted

You are reading this latest preprint version

Transportation Revolution and Banking Development：Evidence from Prewar China

Status:

Version 1

Abstract

Figures

1. Introduction

2. Historical Background

2.1 Transportation Revolution from Railroad Construction

2.2 China's Financial System: from Traditional to Modern

2.2 Modern Banks and the Railroad Network

3. Data And Variables

3.1 Modern Banks

3.2 Railroad Construction

3.3 Control Variables

4. Railroad Construction In Promoting Banking Development

4.1 Causality between Establishment of Railroads and Banks

4.2 Placebo Tests from Planned but Unconstructed Rail Lines

4.3 The roles of concessions and hubs on bank expansion

4.4 Mechanisms

5. Conclusion

Declarations

References

Tables

Additional Declarations

Supplementary Files

Status:

Version 1