2.1 A Working Definition of Financial Structure

A structure is defined as the ratio of different components that share common functions while having distinctive features. From the NSFE perspective, the financial structure refers to the relative weights of financial institutions with different features in the financial system. Though different financial institutions may have similar functions such as mobilizing savings, allocating resources, and managing risks, they have distinctive comparative advantages in meeting different demands of the real economy.

The existing literature has been primarily focused on two important aspects of financial structure: the relative importance of financial markets and banks in the financial system, and the distribution of banks of different sizes in the banking industry. Reference Beck and LevineBeck and Levine (1999) introduced a new database of indicators of financial development and structure across countries and over time. This database is unique in that it unites a wide variety of indicators that measure the size, activity, and efficiency of financial intermediaries vis-à-vis markets. The existing literature usually uses the proportion of the assets of the top N banks to measure the structure of big banks vis-à-vis small banks (Reference Beck, Demirgüç-Kunt and MaksimovicBeck et al., 2004). However, the aforementioned two prevailing measurements of financial structure are not exhaustive.

From the NSFE perspective, financial structure can be measured in alternative ways, which depend on the research question. If we examine the relative merit of capital market versus banks in serving the financing demands of the high-tech industries, we can focus on whether financial activities need to pass through financial intermediaries and measure the financial structure by the ratio of financial markets to financial intermediaries. If we examine whether financial activities are regulated by the government, we can divide financial institutions into formal finance and informal finance. If we examine whether financial institutions are development-oriented or profit-driven, we can measure the financial structure as the relative weight of development banks to commercial banks.

2.2 Does Financial Structure Matter?

The early literature maintains that financial structure matters, and debates which is better – bank-based or market-based financial systems (Reference Allen and GaleAllen & Gale, 2000; Reference DiamondDiamond, 1984; Reference Holmstrom and TiroleHolmstrom & Tirole, 1997). The principal-agent theory has often been used to underpin the models in the debates.Footnote ³ The bank-based view highlights the positive role of banks in acquiring information about firms and managers and thereby improving capital allocation and corporate governance, managing liquidity risk, and thereby enhancing investment efficiency, mobilizing capital to exploit economies of scale (Reference Allen and GaleAllen & Gale, 1999; Reference DiamondDiamond, 1984; Reference Siri and TufanoSiri & Tufano, 1985). The market-based view highlights the growth-enhancing role of well-functioning markets in fostering greater incentives to conduct research, enhancing corporate governance by easing takeovers, and facilitating risk management (Reference Holmstrom and TiroleHolmstrom & Tirole, 1993; Reference Jensen and MurphyJensen & Murphy, 1990; Reference LevineLevine, 1991). Since Reference GoldsmithGoldsmith (1969) tried to explore the impact of financial structure upon economic growth, a large number of economists have continued to discuss which is better, bank-based or market-based financial systems, but there is still no consistent conclusion.

Later on, the view that financial structure matters was challenged, as authors of empirical studies failed to find a statistically significant relationship between financial structure and economic performance using a more comprehensive dataset. Reference Allen and GaleAllen and Gale (2000) discussed financial systems in five industrial economies and found that the United States and the United Kingdom are market-based financial systems, while Germany, Japan, and France are bank-based financial systems. They noted that all five countries have similar long-run growth rates. The experiences of the five countries considered suggest that there is no optimal financial structure and a variety of financial structures can lead to high rates of growth in real per capita GDP.Footnote ⁴ Reference LevineLevine (2002) constructed a broad cross-country dataset to examine market-based and bank-based financial systems. The results indicate that although overall financial development is robustly linked with economic growth, there is no support for either the bank-based or market-based view. Reference LevineBeck and Levine (2002) used cross-country and cross-industry regression analyses and found that having a bank-based or market-based system per se does not seem to greatly affect financing the expansion of industries that depend heavily on external finance, facilitating the formation of new establishments, and improving the efficiency of capital allocation across industries.Footnote ⁵ The view that financial structure does not matter at the macro level echoes the classic theory of the MM theorem that capital structure is of little relevance in deciding the market value of firms at the microlevel. Reference Modigliani and MillerModigliani and Miller (1958) argued that in a perfect and frictionless financial system, the choice between debt and equity financing has no consequence for the market value of the firm or the cost of capital. The theory assumes that there is no income tax, information asymmetry, or agency cost in the financial market.Footnote ⁶

The prevailing view that financial structure does not matter is reinforced by the financial service view, which stresses that financial arrangements – contracts, markets, and intermediaries – arise to ameliorate market imperfections and provide financial services (Reference LevineLevine, 1997; Reference Merton and BodieMerton & Bodie, 2006). The financial services view argues that what matters is how the financial arrangements assess potential investment opportunities, exert corporate control, facilitate risk management, enhance liquidity, and ease savings mobilization. In short, the main issue is not banks versus markets. Instead, it is important to have different financial systems to provide efficient financial services to facilitate economic growth. Accordingly, the financial services view highlights how to create better-functioning banks or markets and relegates the bank-based versus market-based debate to the background.

An influential strand of the financial-services view is the law and finance view (Reference La Porta, Lopez-de-Silanes, Shleifer and VishnyLa Porta et al., 1997, Reference La Porta, Lopez-de-Silanes, Shleifer and Vishny1998). The law and finance view argues that finance is a set of contracts defined by legal rights and enforcement mechanisms. A good legal system can increase the confidence of investors and make them believe that investment can obtain reliable returns. As for whether investors invest through banks or stock markets, the channel itself is not important. This view has been supported by the empirical regression analysis that legal system efficiency boosts industry growth, new establishment formation, and efficient capital allocation (Reference LevineBeck & Levine, 2002).

2.3 Why Financial Structure Matters from the NSFE Perspective

While acknowledging that realizing the functions of financial arrangements requires a supporting environment, NSFE rethinks this mainstream approach by emphasizing that different financial arrangements intrinsically have distinct comparative advantages in meeting divergent financing needs of different industrial structures with different capital requirements and risk profiles for their operations and investments in the real economy. It also states that the structure of the real economy differs at different stages of development, causing the appropriate financial structure to be different accordingly.

One possible reason why the recent cross-country regression analysis has not identified a statistically significant association between financial structure and economic performance is that there is no optimal financial structure suitable for all countries. For instance, Reference Boyd and SmithBoyd and Smith (1998) argue that banks are particularly important at low levels of economic development. Reference Demirgüç-Kunt, Feyen and LevineDemirgüç-Kunt et al. (2013) empirically find that securities markets become more important as economies develop. Reference TadesseTadesse (2002) examined the relation between the financial structure and economic performance in the real sector. This author found that whereas market-based systems outperform bank-based systems among countries with developed financial sectors, bank-based systems fare better among countries with underdeveloped financial sectors. The preceding findings suggest that there is no one-size-fits-all optimal financial structure that can be applied across countries at different stages of economic development. From the NSFE perspective, while it is acknowledged that realizing the function of financial arrangements needs an enabling institutional environment, it maintains that the appropriate financial structure of economies at different stages of development hinges on distinctive and evolving financing needs of underlying production structures.

One strategic underlying assumption behind the MM theorem is that firms can be divided into “equivalent return” classes such that the return on the shares issued by any firm in any given class is proportional to (and hence perfectly correlated with) the return on the shares issued by any other firm in the same class (Reference Modigliani and MillerModigliani & Miller, 1958: 266). This assumption neglects the substantial variation in the industrial structure to which firms belong and thus the capital requirement, size, and sources of risk of firms. The empirical evidence shows that firms from different industries have distinctive capital structures. For instance, large, integrated oil companies from technologically mature industries have relied mostly on debt for external financing, whereas the major pharmaceutical companies with huge technological risks typically operate at negative debt ratios (Reference MyersMyers, 2001). Although authors of the existing studies have explored how firm characteristics may affect capital structures, it is worth investigating the extent to which firm characteristics may be influenced by production structures where firms are embedded. Hence, it is promising to explore how industrial characteristics may affect the appropriate capital structure for firms.

In summary, NSFE argues that financial structure matters when examining how finance can best service the real economy, which has different production structures at different stages of development. NSFE maintains that different financial arrangements have their own advantages and disadvantages in mobilizing savings, dispersing risks, and allocating funds. NSFE is not defying the role of institutions in realizing the functions of financial arrangements, but it emphasizes that certain functions are not perfectly substitutable across different financial arrangements.

3.1 Variation in Financial Structures

There are substantial variations in the financial structure across countries and over time. Taking the financial structure measured by the relative weight of financial markets to banks as an example, Reference Lin, Sun and JiangLin et al. (2013) showed the evolution of financial structure across income groups and over time and identified the following stylized facts: first, the cross-country data shows that the more developed the economy, the higher the share of the stock market in the financial structure; second, the time-series data reveals that the share of the stock market in the financial structure tends to gain more weight as economies move to a more advanced stage of development. The history of the financial structure of the United Kingdom and the United States confirms that large banks and stock markets are playing an increasingly important role in the financial system as economies grow. During the English Industrial Revolution, industrial enterprises were typically extremely small. Interestingly, the English banks were also typically small and locally based with a limited number of offices until at least the mid-nineteenth century. The bank merger movement in England did not arise until the 1860s; it was not until 1918 that the “Big Five” banks came to dominate the English banking system. The London capital market, despite its current prominent role in the national and international financial system, did not make any significant direct contribution to capital formation in the English Industrial Revolution. It was not until the end of the nineteenth century that the capital market started to play an important role in financing industrial sectors. In the United States, before 1890, industrial firms were numerous, small, and closely owned by a limited number of shareholders. As late as the end of the ninete enth century, it was still rare for manufacturing or other non-infrastructural firms to be listed on exchanges. In fact, there was no large-scale market in common stocks prior to 1920. A variety of local financial institutions emerged to meet the needs of small and medium-sized firms in the economy (Reference Lin, Sun and JiangLin et al., 2013).

Furthermore, there are variations in financial structure among countries even within the same income group. Figure 1 shows substantial differences in the financial structure among the United Kingdom, the United States, Germany, and Japan. The financial structure of the United Kingdom and the United States was more dominated by the stock market, whereas the financial structure of Germany and Japan was more dominated by banks before 2000. Nevertheless, the financial structures in these four advanced countries converge to a similar profile after 2000.

Figure 1

Financial structure in major developed economies

Data source: World Bank, Global Financial Development Database.

3.2 The Existing Explanations of Variation in Financial Structures

To explain why the financial structure of Germany and Japan was bank-based whereas that of the United Kingdom and the United States was market-based, one existing explanation argues that legal origin matters – civil law countries have much weaker investor protection than common law countries, resulting in less developed capital markets (Reference La Porta, Lopez-de-Silanes, Shleifer and VishnyLa Porta et al., 1997). Using a sample of forty-nine countries, Reference La Porta, Lopez-de-Silanes, Shleifer and VishnyLa Porta et al. (1997) found that weak investor protection in civil law countries, as measured by both the possibility of legal rules in protecting investors and a survey-based estimate of the quality of law enforcement, is associated with the underdevelopment of capital markets.

However, the “legal origin” theory cannot explain the puzzle of why the ratio of France’s stock market capitalization to GDP was twice as high as that of the United States in 1913 (Reference Rajan and ZingalesRajan & Zingales, 2003) and why the financial strucure in the mjaor developed countries have been converging after 2000 as shown in Figure 1. If this theory holds water, we would expect that the stock market in France was much less developed than the US stock market. Furthermore, the “legal origin” theory is of little use in explaining changes in financial structure over time, given that the legal origin is a time-invariant variable. For instance, despite their current prominent role in the national financial system, capital markets did not start to play an important role in financing industrial sectors in the United Kingdom until the end of the nineteenth century (Reference CollinsCollins, 1988). In fact, the financial structure of the United Kingdom was bank-based during the Industrial Revolution: because industrial enterprises were typically extremely small, the English banks had been small and locally based until at least the mid-nineteenth century (Reference CameronCameron, 1992; Reference Cameron, Crisp, Patrick and TillyCameron et al., 1967). This bank-based pattern has also been applied to the early history of the US financial system (Reference Cull, Davis, Lamoreaux and RosenthalCull et al., 2006).

Another strand of theory holds that culture matters in explaining variations in the financial structure. Reference Stulz and WilliamsonStulz and Williamson (2003) argued that differences in cultures, proxied by differences in religion and language, cannot be ignored when examining why investor protection differs across countries. After controlling for the legal origin, they found that civil law Protestant countries have stronger creditor rights than civil law Catholic countries. Hence, they concluded that although legal origin matters in explaining variations in shareholder rights, culture plays a key role in explaining the variation in creditor right protection. Hence, they infer that Catholic countries may have less developed debt markets.Footnote ⁷ Similar to the legal origin theory, the culture theory maintains that the protection of investor rights, including both shareholder rights and creditor rights, is crucial for fostering certain types of financial arrangements.

Yet, the culture theory may be of little help in explaining changes in financial structure over time because culture, such as the principal region or language, is relatively stable. Another major limitation is that both the legal origin theory and the culture theory assume that a particular form of legal or cultural tradition is necessary for enforcing financial agreements. Consequently, they may neglect viable alternative institutional arrangements that can perform the same function (Reference LinLin 1989). For instance, Reference Miao, Niu and NoeMiao et al. (2021) documented that a large bank loan market – the trusted assistant loan market in nineteenth-century China – emerged even in the absence of contracting institutions. They argued that the repayment of trusted assistants was enforced through expertise leverage, which was sufficiently strong to implement enforceable financial agreements and support a large lending market.

A third strand of literature argues that history matters. Reference Monnet and QuintinMonnet and Quintin (2007) presented a dynamic, general equilibrium model in which past fundamentals of an economy (including the cost of bank intermediation, the cost of entry on the financial market, and the distribution of project characteristics) justify a unique, efficient path for its financial system. For instance, they argued that the reason why the United States is a market-based financial system whereas Germany is a bank-based one is that federal laws discouraged bank intermediation in the United States while Germany used to impose significant legal barriers to entry into financial markets. Therefore, despite the convergence in the legal framework, the financial structure in the United States and Germany continues to vary markedly. Although path dependency may play a role in the development of financial institutions, it still begs the question of why governments foster certain types of financial arrangement while depressing others at the initial phase and why they all converged to a similar financial structure once they reached the same advanced development stage.

3.3 The NSFE Perspective

From the NSE perspective, production structure, namely the combination of different industries and different technologies used in each industry, determines the financing demand and firm’s risk profile of the real economy. Production structure in an economy is endogenous to the economy’s factor endowments structure because the factor endowments structure determines the economy’s comparative advantages and thus the appropriate technologies and industries; the endowments structure is given at any specific time and changeable over time, and correspondingly, the production structure of the economy varies (Reference LinLin, 2012). From the NSFE perspective, a key determinant of financial structure is production structure because the role of finance is to serve the real economy’s operation and investment. If we define the financial structure as the relative weight of banks versus capital markets, NSFE would predict that high-income countries are likely to develop full-fledged capital markets to meet the financing needs of the industrial structure that is at the global technological frontier and relies on costly and risky R&D for technological innovation and industrial upgrading. The rationale is that banks are more appropriate for the financing of traditional asset-intensive industries, whereas capital markets favor innovative and risky projects. Reference Allen, Bartiloro, Gu and KowalewskiAllen et al. (2018) deployed a panel dataset of 108 countries over the sample period 1972–2015 to test whether countries with predominant tangible-asset-intensive industries are more likely to exhibit a bank-based financial system. Furthermore, they used a difference-in-difference strategy to conduct event studies by exploring how the financial structure responds to shocks to the industrial structure. They found that the stock markets experienced significantly faster growth than the banking system after the service sector gained more importance. From this perspective, the variation in the financial structure between the United States and Germany lies in the divergence in their industrial structure. Financial systems in countries such as Germany would remain bank-based as long as their economies are dominated by manufacturing industries. By contrast, the financial system in the United States will continue to be market-oriented as long as service and highly innovative companies constitute a large share of the economy. From the NSFE perspective, although we acknowledge that the industrial structure matters in determining the appropriate financial structure, we further highlight that technological structures are of equal importance. For instance, within the manufacturing sector, some industries are at the global technological frontier, which relies more on capital markets to mobilize large-scale and high-risk capital for innovations, whereas others use mature technologies that may depend on banks for financing.

In summary, although the legal origin, culture, and history play a role in explaining the variation in financial structure, they have limitations in explaining the changes in the financial structure over time. The NSFE draws our attention to the analytical angle of how changes in financing needs of the real economy drive the adjustment in financial systems. This perspective helps to explain the variation in the financial structure across countries and the evolution of the financial structure within a country that are unresolved by the existing theories.

4.1 Distinctive Demand of the Real Economy

From the NSFE perspective, grasping the financing needs of the real economy is of paramount importance in answering what the appropriate financial structure is and why there are substantial differences in the financial structure across different countries. This section delves deeper into the distinctive financing needs of the real economy which lays the foundation for examining comparative advantages of different financial arrangements in meeting specific financing needs.

When an economy develops from the early stage to the advanced stage in modern times, the proportion of agriculture in the industrial structure gradually falls, while that of manufacturing and services tends to rise (Reference KuznetsKuznets, 1946). During the process, the production structure evolves from labor-intensive to capital-intensive and technology-intensive. Different technologies and industries have different technical features, such as requirements for specific capital, skill and infrastructure, and economies of scale and risk, and thus they have distinctive demands for financing to meet the required size of investment and operation and to cope with the innate risks. For example, Reference OzawaOzawa (2007) noted that the industrial structure of Japan has undergone four stages of development dominated by labor-intensive industries, capital-intensive heavy industry and chemical industry, capital-intensive automobile industry, and R&D-intensive electronic industry, respectively, after World War II. Reference Ju, Lin and WangJu et al. (2015) showed that the capital intensity increased over time within each industry in the United States during the period 1958–2005.

4.2 Comparative Advantages of Different Financial Arrangements

Financial institutions serve the real economy by performing functions such as mobilizing savings, allocating capital, and diversifying risks. Different financial institutions have different comparative advantages in performing these functions because they have distinctive organizational forms, modalities, and ways to cope with information asymmetry and moral hazard. As a result, when one financial institution defies its comparative advantages in serving firms or industries, it would undermine the efficiency of this financial institution. For example, if a small bank provides loans to a firm engaging in R&D for a new technology, it will have difficulties to diversify risks. Hence, to enhance the efficiency of the financial system in serving the real economy, it is essential to identify the comparative advantages of different financial arrangements. In the following, we will examine selected financial arrangements to illustrate why different financial arrangements have distinctive comparative advantages.Footnote ⁸

The remaining section proceeds as follows: we first focus on the widely examined financial arrangements – banks versus the stock market. Then we examine the differences in comparative advantages between big and small banks, and shed light on how digital technology shapes the comparative advantage of digital finance. Finally, we analyze the comparative advantages of informal finance vis-à-vis formal finance in financing the real economy.

4.3 Match between the Financing Demand and Supply

The NSFE maintains that an appropriate financial structure matches the financing demand of firms within a certain production structure with comparative advantages of different financial arrangements. Because the production structure varies across different development stages, the appropriate financial structure should differ as well. The appropriate financial structure at the early stage of a developing country should enable financial sectors to meet the demands of labor-intensive industries such as garments and footwear. Firms in these labor-intensive industries are often small, opaque, and with a low-level innovation risk. Owing to the comparative advantages of harnessing soft information, small local banks are more appropriate to provide external financing for such firms. Hence the financial system of a developing country at the early stage should be characterized by the dominance of small banks.

As economies grow, firms in the labor-intensive industries often increase their capital inputs and upgrade their technology level, which induces the emergence of some large-sized enterprises. In this case, big banks and capital markets may emerge. However, such external financing demand is limited because small firms still dominate in the production structure. Hence, small banks will still play a critical role. Meanwhile, technological development might enable those small firms to produce and accumulate certain digital footprints. Under such a situation, small firms can rely on digital finance, such as fintech firms, to seek external finance.

As economies move to the advanced stage, capital-intensive industries and technology-intensive industries dominate. Firms in such industries tend to require large amounts of external financing and face high risks, including technology innovation and product innovation. The stock market and big banks are more suitable to serve those firms, whereas small banks are no longer able to diversify risks and provide large-scale financing. In addition, with the increasing digital level of enterprises, digital finance will play an increasingly significant role. Overall, the financial structure should be dominated by the stock market and big banks in advanced economies.

4.4 A Preliminary Discussion on Methods of Empirical Testing of Appropriate Financial Structure

As already explained, NSFE argues that only when the comparative advantages of financial structure match well with financing needs of the production structure determined by the economy’s endowment structure can the financial system effectively perform its functions of mobilizing funds, allocating funds, and diversifying risks so as to contribute to sustainable economic development. Therefore, as the financing needs of the underlying production structure differ, the appropriate financial structure would differ at different stages of development. This section explores how to build on the existing empirical literature to better test the theories on appropriate financial structure.

One strand of the existing literature regresses economic performance on financial structure by using the country-, industry-, and firm-level data. Reference LevineLevine (2002) constructs an assortment of measures for forty-eight countries over the 1980–95 period. The data analysis does not provide any evidence supporting the theories that favor either the bank-based or market-based perspectives. Categorizing countries based on their financial structure does not contribute to explaining the differences in long-term economic performance across nations. Instead, when countries are categorized based on their overall level of financial development, it finds a positive correlation between financial development and economic growth. Reference LevineBeck and Levine (2002) use the data of thirty-six industries in forty-two countries from 1980 to 1990 to empirically test the impact of financial development and financial structure on industry growth and capital allocation. The results show that the overall level of financial development and the legal environment play a crucial role in explaining the differences in the number of establishments within industries that rely on external financing across countries, and the impact of financial structure upon economic growth is not statistically significant. Reference Demirgüç-Kunt and MaksimovicDemirgüç-Kunt and Maksimovic (2002) investigate whether firms’ access to external financing to fund growth differs in market-based and bank-based financial systems. Using firm-level data for forty countries, the paper computes the proportion of firms in each country relying on external finance and examines how that proportion differs across financial systems, and finds that the development of a country’s legal system predicts access to external finance. Yet there is no evidence that the development of a market-based or bank-based financial system per se affects access to financing. The above literature mentioned in this paragraph assumes that there is an optimal financial structure for all countries. Yet from the NSFE perspective, the potential reason why the coefficient of financial structure is statistically insignificant is that there is no single optimal financial structure for all countries. Hence, statistical insignificance does not necessarily mean that financial structure lacks economic significance.

Another strand of literature holds that the relative importance of banks and financial markets depends on different stages of development. Reference TadesseTadesse (2002) uses a sample of thirty-six countries from 1980 to 1995, and finds that in countries with an underdeveloped financial system, bank-based systems outperform market-based systems. This paper constructs an interaction term between financial architecture and firm size, and then examines the relation between financial architecture and economic performance across countries of differing distribution of firm size, which finds that this correlation is significantly negative in the lowest quartile of countries on the average firm-size scale and significantly positive in the highest quartile of countries. Reference Demirgüç-Kunt, Feyen and LevineDemirgüç-Kunt et al. (2013) examine the evolving importance of banks and securities markets during the process of economic development. The paper finds that with the growth of economic activities, the association between an increase in real output and that in bank development decreases, while the association between an increase in real output and that in securities market development becomes larger. The results suggest that as economies develop, the services provided by securities markets become more important for economic activity, while those provided by banks become less important.

These efforts have primarily focused on association or correlation between financial structure and economic development, and have not adequately addressed the underlying causal mechanisms. In order to mitigate the potential endogeneity concerns, Reference Allen, Bartiloro, Gu and KowalewskiAllen et al. (2018) conduct event studies using a difference-in-difference strategy, and explore the shocks to the structure of the real economy and investigate how the financial structure responds to such shocks. They use different shocks that alter the economic structures in India, Finland, Sweden, and South Korea. In all these countries the shocks resulted in the significant development of the service sector relative to the industry sector. Consequently, these countries move toward more market-based financial systems. Yet this approach is not free from limitations. First, country selection may not be random. Second, for those policies that cause such shocks, they may not only affect the economic structure but also the financial structure.

To further mitigate the endogeneity problem, structural estimation may be deployed to make the counterfactual analysis to test the causal relationship. Structural estimation is an attempt to estimate an economic model’s parameters and assess the model fit, which allows it to answer counterfactual questions such as what would happen if we shocked this variable and how would the world look if we changed this parameter’s value. Although strategies such as difference-in-difference, regression discontinuity, and randomized controlled trials have been popular in empirical research in search for robust identification of causal effects, structural estimation can still be a potent complementary tool to help disentangle alternative mechanisms. Especially when there is a lack of data on the treatment or policy we are interested in (e.g., when this policy has never been previously implemented), structural estimation methods may be a viable alternative method of empirical testing.

In addition to enhancing econometric methodology, more efforts can be made to improve the measurement of financial structure. The existing literature mainly follows the approach of Reference LevineBeck and Levine (2002) and Reference LevineLevine (2002) to measuring the size, activity, and efficiency of the financial structure as defined by the relative importance of bank-based versus market-based financial system. This approach provides cross-country data, laying the foundation for empirical analysis. The size of financial structure is a measure of the size of stock markets relative to that of banks, which equals the logarithm of the market capitalization ratio divided by the bank credit ratio. The market capitalization ratio equals the value of domestic equities listed on domestic exchanges divided by GDP, and the bank credit ratio equals the value of deposit money that bank credits to the private sector as a share of GDP. The activity of financial structure is a measure of the activity of stock markets relative to that of banks, which equals the logarithm of the total value traded ratio divided by the bank credit ratio. The total value traded ratio equals the value of domestic equities traded on domestic exchanges divided by GDP. The efficiency of financial structure is a measure of the efficiency of stock markets relative to that of banks, which equals the logarithm of the total value traded ratio times overhead costs. Overhead costs are measured by the overhead costs of the banking system relative to banking system assets. While acknowledging these meaningful efforts, there is room for improvement to address the following limitations of indicators.

First, stock market prices may fluctuate substantially in some years, but this does not indicate that the functioning of the stock market has also changed substantially, nor does it indicate that the financial structure of the country has changed dramatically. This may just be a short-term deviation from the trend. For example, the ratio of US stock market trading value to private sector credit was 1.8 in 2000, and dropped to 0.9 in 2018. Obviously, we cannot think that the importance of the US financial market relative to banks has declined significantly in the past two decades, because the US stock market rose sharply due to the Internet bubble around 2000, and then fell sharply. Hence, we need to explore complementary measures that can mitigate the effect of short-run fluctuations and better reflect the long-term trend of financial structure.

Second, there may be excessive speculation in the stock markets of some countries, resulting in abnormally high values of their financial structures. Take Turkey as an example, from 1999 to 2007; the average value of Turkey’s stock market trading value to private sector credit ratio was 2.1, while that of the United States during the same period was only 1.1. It would be misleading to use this measurement to conclude that Turkey’s financial structure is more inclined toward market dominance compared to that of the United States. In fact, in most years, the value of Turkey’s financial structure is significantly lower than that of the United States.

Lastly, in the case of small economies serving as international financial centers, their stock market trading value to private sector credit ratio can be extremely high. However, the listed companies in these economies may be foreign companies, that is, the highly developed stock market may not mainly serve the financing needs of enterprises within the economy. In other words, existing theories usually assume that the more the financial structure is dominated by the financial market, the more favorable it is for the development of innovative industries in the economy, but obviously this is not fully true for small economies in international financial centers. Taking Hong Kong, China, as an example, from 2007 to 2017, the average ratio of stock market trading value to private sector credit was 3.4, indicating a typical financial structure dominated by the financial market. However, this has not led to rapid development in innovative industries in Hong Kong.

Moving forward, we can at least make two efforts to enhance the empirical testing of appropriate financial structure. First, we can build on the existing static comparative analysis and explore how the change of production structure drives the evolution of financial structure over time. In the future empirical research, we can consider starting from the factor endowment structure and the corresponding production structure to explore the appropriate financial structure. Second, it is useful to explore the conditions under which appropriate financial structure can fulfill its potential for fostering the real economy. The efficiency of financial services for the real economy is affected by transaction costs, market environment, and government governance. Even though a country’s actual financial structure is theoretically appropriate, its impact upon economic development may be constrained by institutional environment. For example, the formal or informal systems for determining transaction costs such as corporate information disclosure, qualification review, investor protection, and bankruptcy law will directly affect the efficiency of different financial system arrangements to serve the real economy.

5.1 Establishing PDFIs to Address Market Failures of Financial Markets and Institutions

If the role of finance is to serve the real economy, market failures may exist in decentralized commercial banking systems and private capital markets for serving certain types of investments due to information asymmetry, positive externalities, and excessive risk aversion of commercial financial institutions. Governments may intervene to fix market failures and provide financial support for credit-rationed enterprises or disadvantaged clients such as SMEs and rural smallholders. Government interventions into financial systems can take many forms, including establishing specialized PDFIs, providing state guarantees to incentivize commercial banks to provide financial support for underserved clients, and setting up state-owned venture capital funds to incubate frontier markets.Footnote ⁹ Yet the mainstream perspective holds that governments should not directly intervene in financial systems but rather focus on indirect efforts to enhance the legal and institutional environments (World Bank, 2015). The rationale is that governments may suffer from corruption and rent-seeking so that the cost of government failures may exceed that of market failures. This echoes the debate on industrial policy: even though there is a basic consensus on the necessity of state intervention to address market failures, researchers disagree with each other on how to intervene (Reference Lin and XuLin & Xu, 2018). Although the popular perception that direct state intervention in financial systems fails miserably has tempted us into saying “no” to all kinds of state intervention, a more constructive approach is to delve deeper into comparative studies of both successes and failures to make prudent recommendations on how to make state intervention in financial systems work better in practice. Otherwise, we may risk turning our back on what could be genuinely helpful.

In this section, we will take PDFIs as an example to illustrate why it is important to go beyond unproductive confrontation on whether direct state intervention in financial systems is needed and to investigate how to make it work better. PDFIs are public financial institutions initiated and steered by governments with an explicit official mission to fulfill public policy objectives. In addition to equity- and guarantee-focused public funds, public development banks are a major category of PDFIs. They are potentially useful policy instruments for fixing market failures, incubating markets, and promoting structural transformation in an equitable and sustainable manner. Yet not all PDFIs are successful. It is important to explore how to make PDFIs work better.

The rest of this section proceeds as follows: we will first examine how the prevailing development thinking has shaped the evolution of PDFIs after World War II, ponder what insights the NSFE can offer to study the role of PDFIs, and then introduce the first global database on PDFIs to identify puzzles that the mainstream free-market development thinking fails to explain; we conclude with promising research questions on PDFIs.

5.2 Making Tailored Financial Regulations

From the NSFE perspective, there is no universal optimal financial regulatory scheme for all countries; instead, tailored financial regulation should be made based on the specific characteristics of financial markets, institutions, instruments, contracts, and so on.Footnote ¹² In this section, we tentatively explore why financial regulation needs to be tailored by examining distinctive characteristics of stock markets vis-à-vis banking systems, large banks vis-à-vis small banks, commercial banks vis-à-vis NDBs, and formal finance vis-à-vis informal finance.

Regulations of stock markets and banking systems should differ because stock markets and banks function differently. Stock markets create risky liquidity based on money providers’ price discovery actions, while banks create safe liquidity by transferring short-term debts such as demand deposits into longer-term assets such as mortgage loans (Reference Dang, Gorton, Holmström and OrdonezDang et al., 2017). Therefore, it is essential to enhance information disclosure in the stock markets through financial regulation. By contrast, banks need to keep the information they produce about backing assets, such as firm loans, secret to maintain outsiders’ expectation about the value of banks’ debts so that banks’ debts (deposits) can be easily traded at par like money because they are “riskless” to some extent. Hence, banking regulation places less importance upon information disclosure and puts more focus on mitigating moral hazards owing to information asymmetry among banks and depositors, such as banks’ capital requirements that increase “skin in the game.”

Banking regulation should consider how different types of banks can be better regulated. Even though in practice the Basel Accord pays special attention to global systemically important banks, banking regulation has not fully considered the heterogeneity within the banking system. Banking regulation is harmonized across countries even though countries vary in their banking structure, such as a mix of big banks and small banks. For example, although the capital requirements in Basel III were calibrated primarily for big banks in advanced economies, some key emerging-market economies have committed to implementing Basel III, and many other countries are in the process of adopting it to regulate big banks as well as small banks. They do so because they see it as being in their long-term interest even though the standards may pose new risks and challenges (Reference Beck and Rojas-SuarezBeck & Rojas-Suarez, 2019), and the Basel Accord has been perceived as the international best practice so that regulators in developing countries face peer pressure to adopt it (Reference Jones and ZeitzJones & Zeitz, 2017).

It remains unclear whether the regulatory framework designed for big banks is suitable for small banks. Taking capital requirements as an example, the existing literature proposes that when setting capital requirements, regulators should take into account the bank’s executive incentive schemes (Reference Eufinger and GillEufinger & Gill, 2017), the regulator’s screening reputation (Reference Morrison and WhiteMorrison & White, 2005), the bank’s ownership structure (Reference Laeven and LevineLaeven & Levine, 2009), and the country’s fiscal capacity (Reference StavrakevaStavrakeva, 2020). Yet it ignores the substantial difference between big banks and small banks and its impact on the effectiveness of regulatory frameworks. The latest research has begun to pay attention to the importance of bank size. For example, Reference Dávila and WaltherDávila and Walther (2020) showed that bank size is a key determinant of banks’ leverage choices when system-wide bailouts are possible. They argued that the optimal bank regulation should be size-dependent policies where capital requirements are identical for big banks and small banks, but big banks should be charged a supplement tax on borrowing. However, according to Reference Dávila and WaltherDávila and Walther (2020), the only difference between a big bank and a small bank is that the former is prone to the “too big to fail” moral hazard problem.

From the NSFE perspective, the capital requirements in Basel III may not help small banks to achieve financial stability, given the fact that small banks are generally more opaque than big banks because the former usually lend more to small firms whose information is mainly “soft” (Reference Berger, Miller, Petersen, Rajan and SteinBerger et al., 2005; Reference HuberHuber, 2021; Reference Liberti and PetersenLiberti & Petersen, 2019). Consider a simple case where bank regulations take two forms: capital requirements that specify the level of equity that a bank must hold as a percentage of its total assets, and external regulations that specify the taxation or penalty on the bank if it invests imprudently and such a risk-taking behavior is discovered by the regulator. We argue that the optimal regulations for big banks and small banks should be different – capital requirements for small banks should be higher than those for big banks, and supervising big banks should rely more on external regulation compared with small banks – for the following reasons.

First, big banks have advantages in funding big firms and small banks have advantages in funding small firms owing to the information types of different firms and the organizational structures of different banks (Reference Berger and UdellBerger & Udell, 2002; Reference SteinStein, 2002). Funding big firms usually relies on transaction-based lending technologies such as financial statement lending, asset-based lending, or credit scoring because the information of big firms is mainly “hard” and standard, which is easy to identify, transmit, and verify. On the contrary, loans to small businesses are often relationship-based and rely heavily on “soft” information of small firms that cannot be directly verified by the bank manager who does not produce it. Therefore, big banks of large hierarchies may entail complex agency and information problems when funding small firms because a line manager inside a large hierarchy – who faces the risk that he cannot get a chance to act on his information ex post because he cannot credibly communicate this information to his superior and his effort of producing soft information will be wasted – does less soft information research ex ante (Reference Berger and UdellBerger & Udell, 2002; Reference SteinStein, 2002). Nevertheless, big banks can perform better when funding big firms whose hard information can be easily verified and transmitted. Besides, big banks can serve big firms without syndication because of their higher ability of liquidity production compared with small banks with fewer total assets. The scenarios for small banks are quite different. The agency problem and information problem of small business lending may not be a serious problem for a small bank with few managerial layers, because a line manager who produces soft information may also be in a position to make lending decision, and he will have a strong incentive to gather information ex ante. Therefore, there is an appropriate matching relationship between banks of different sizes and firms of different sizes; that is, big banks fund big firms, and small banks fund small firms.

Second, banking regulation may have a substantial impact on the matching relationship between banks and firms, especially on small banks when they choose which type of firms to fund. When a small bank chooses to fund big firms, it might be the case that small banks can only reach big firms of low quality and high default risk because big firms of high quality and low default risk might have already been funded by big banks. By contrast, when a small bank chooses to serve small businesses, it is more likely to fund small firms of high quality and low risk that are underserved by big banks owing to the difficulty of collecting and transmitting soft information. Besides, small banks are more opaque when providing small business lending or relationship lending based on soft information, which is costly for the “outsider” – the regulator – to verify. As a result, it is hard for the regulator to correctly identify the risk of small banks that serve small firms, let alone impose taxation or penalty on banks’ excessive risk-taking behaviors. Therefore, if the external regulation is relatively strict – that is, the taxation or penalty ex post is relatively high – small banks will choose to fund big firms other than small firms, although small banks are more suitable to fund small firms. On the contrary, the taxation ex post on big banks can be high because big banks are more transparent given they mostly fund big firms whose information is mainly hard, and it is less likely for the regulator to wrongly tax big banks.

Nevertheless, bank regulation should not only be limited to capital requirements and external regulation to the excessive risk-taking behavior but also target regulation of other potential risks in banking. For example, stated-owned banks may be inclined to grant loans to politically connected firms, and banks which are controlled by business groups are inclined to provide loans to affiliated enterprises within the business group (Reference Bertrand, Mehta and MullainathanBertrand et al., 2002; Reference Khanna and PalepuKhanna & Palepu, 2000; Reference Khwaja and MianKhwaja & Mian, 2005). Regulators should mitigate these potential risks by limiting the proportion of bank loans granted to politically connected enterprises, and strictly screening and supervising business groups holding bank shares when issuing banking licenses or approving bank mergers and acquisitions.

When it comes to the question of what the appropriate financial regulatory framework for NDBs is, we would like to investigate whether they should comply with the Basel Accord that was designed to regulate large, complex, and internationally active commercial banks. As discussed, unlike profit-maximizing commercial banks, NDBs are public financial institutions initiated and steered by national governments to fulfil public policy objectives. Unlike retail-deposit-taking commercial banks, NDBs often do not take household deposits; instead, they rely on the government support to mobilize long-term funds via market means, such as issuing bonds on capital markets (Reference Xu, Wang and RuXu et al., 2021).

Although NDBs differ from commercial banks, it is found that a majority of sampled NDBs comply with the Basel Accord or take the Basel Accord as the benchmark to emulate the perceived international best financial regulatory framework. Reference Gottschalk, Castro and XuGottschalk et al. (2022) conducted a survey among fifty NDBs to assess compliance with Basel III. For the purpose of representativeness, the survey took into consideration the income level of their countries, asset sizes, and mandate. Of all the banks included in the survey, nearly 50 percent have answered the questions. Among twenty-four banks, sixteen explicitly acknowledge that they are subject to the same regulatory rules as commercial banks are in their jurisdictions and that these rules are very closely aligned with the Basel Accord. This means that, in that sample, about two-thirds banks are subject to Basel Accord.

From the NSFE perspective, it is of crucial importance to formulate a financial regulatory framework tailored for NDBs to ensure its financial soundness and mitigate the risk of financial crises as financial institutions without undermining their ability to act as development-oriented public financial institutions. Compliance with the Basel Accord may undermine their capacity to fulfill public policy objectives. For instance, Basel credit risk models consider that the longer the term, the greater the risk. But the greater lending periods do not imperatively bring higher risks for NDBs in the same proportion as they do to private banks, because their funding is also held for longer terms such as long-term bonds. Basel III standards on market risks and operational risks impinge on equity finance. But NDBs use equity investments to support innovation, and climate finance, which, in many instances, involves betting on yet untested clean technologies. NDBs often possess industrial expertise to better assess the prospects of emerging industries and technologies. In terms of the capital adequacy requirement, Basel III demands banks to hold higher minimum capital requirements to reduce the risk of insolvency in situations of stress. Basel III has maintained the Basel II minimum capital requirements at 8 percent of risk-weighted assets, but, to that, it has added the requirement of capital conservation buffers at 2.5 percent of risk-weighted assets, to help banks withstand financial distress. For NDBs under Basel III, such standards could also affect disproportionately the availability of credit to those exposures that require more capital. NDBs are often mandated to provide long-term finance such as infrastructure financing, and the empirical study shows that they lend for much longer periods than commercial banks (Reference Hu, Schclarek, Xu and YanHu et al, 2022). But Basel III has in-built features that already require higher levels of capital for longer-term credit. If permission is granted by the national regulator, NDBs with more technical resources and data availability may try to mitigate the quest by adopting internal models for risk assessment. These models permit banks to assign their own risk weights to their portfolio of assets and thereby potentially save capital (Reference Gottschalk, Castro and XuGottschalk et al., 2022).

As for financial regulation of informal finance, NSFE argues that the financial regulation of formal finance is likely to be stricter than that of informal finance, which is also the case in practice (Reference Bell, Srintvasan and UdryBell et al., 1997). In the case of informal finance, lenders are often owners of funds. Hence, they have compatible incentives to monitor the behaviors of borrowers. By contrast, in the case of formal finance, banks are often afflicted with the moral hazard problem since the money they lend to firms belongs to depositors. Banks aim to maximize their own profits, not the utility of depositors, and sometimes the two goals conflict with each other. Consequently, the interests of depositors are ranked second, since depositors cannot supervise the bank’s behavior in an effective way due to the information asymmetry between the two parties. In general, the larger the size of a bank’s liabilities (deposits), the more severe its moral hazard. The scale of formal financing is often larger than that of informal financing. For instance, the size of loans provided by a small bank is generally larger than that of interpersonal borrowing. This entails a stricter financial regulation of banks with a larger size of deposits.

In a nutshell, the appropriate regulation frameworks for banks of different features should differ. As far as the size of banks is concerned, supervising big banks should rely more on external regulation – that is, taxation on the excessive risk-taking behavior – and lower big banks’ capital requirements to improve their abilities of liquidation production. For small banks, capital requirements should be higher and external regulations should be less strict so as not to distort the appropriate matching relationship between small banks and small firms. As for NDBs, the compliance with the Basel Accord, which is designed for regulating large and complex commercial banks from advanced economies, would undermine the ability of NDBs to achieve their development-oriented mandate. A tailored regulatory framework is needed to ensure financial soundness of NDBs while unleashing their potential for fulfilling public policies. Regarding the informal finance, the financial regulation of informal finance is likely to be looser than that of formal finance because the lender of informal finance is often the owner of capital and the size of transaction is relatively small.

5.3 Implementing Pragmatic Financial Reforms to Address Financial Depression

Financial repression is the opposite of financial liberalization. It generally refers to a wide array of government policies that restrict the activities of the financial sector; that is, governments interfere with the full potential operation of the financial sector by introducing numerous interventions, regulations, laws, and other nonmarket restrictions to the behavior of banks and other general financial intermediaries. For instance, financial repression includes control on deposits and loan interest rates, exchange rates, and capital accounts; entry and competition in banking; forcing banks (especially state-owned banks) to provide credit allocation to support “leading sectors” (mainly refers to state-owned enterprises (SOEs) or the industries that the government wants to develop); and requiring banks and other intermediaries to hold more government bonds than they would (Reference Chari, Dovis and KehoeChari et al., 2020; Reference Giovannini and De MeloGiovannini & De Melo, 1993; Reference McKinnonMcKinnon, 1973; Reference Roubini and Sala-i-MartinRoubini & Sala-i-Martin, 1995; Reference ShawShaw, 1973). Countries worldwide, developing countries in particular, have introduced a whole host of restrictions and controls on the behavior of financial intermediaries (Reference Roubini and Sala-i-MartinRoubini & Sala-i-Martin, 1995).

The existing literature has pointed out that financial repression results in inefficiencies and has made policy recommendations to end financial repression (see Reference FryFry, 1980, Reference Fry1997; Reference McKinnonMcKinnon, 1973; Reference ShawShaw, 1973). Many developing economies followed these policy prescriptions and removed controls on international asset trade as well as price and quantity rationing in domestic financial intermediation, under the auspices of international financial institutions such as the IMF and World Bank (Reference Giovannini and De MeloGiovannini & De Melo, 1993).

However, the results of financial liberalization reforms in many developing countries are disappointing. For example, the IMF and World Bank sponsored the structural adjustment programs since the 1980s, which mainly contained the items of fiscal adjustment, getting the prices right, trade liberalization, and, in general, a movement toward free markets and away from state intervention. A study shows that none of the top twenty recipients of repeated structural adjustment lending, as measured by a total number of adjustment loans from the IMF and World Bank over 1980–99, were able to achieve reasonable growth. On the contrary, about half of the adjustment loan recipients showed severe macroeconomic distortions (Reference EasterlyEasterly, 2005). For another example, several Latin American countries such as Chile, Argentina, and Uruguay during the 1970s implemented financial reforms as part of the Structural Adjustment Program to free domestic banks and other financial institutions from government-induced distortions, but these financial reforms yielded widespread bankruptcies, massive government interventions or nationalizations of private institutions, and low domestic savings in domestic financial sectors in the 1980s (Reference Diaz-AlejandroDiaz-Alejandro, 1985). Similar results of financial liberalization have been observed in other countries such as Mexico and Russia in the 1980s and 1990s as documented by Reference Arestis, Nissanke and SteinArestis et al. (2005).

The frustrating results of financial liberalization sparked widespread discussion about the causes of financial and economic instability (Reference LoizosLoizos, 2018). It has been argued that two main factors – macroeconomic fluctuation and institutional underdevelopment – are attributed to instability in these countries when they go through a financial liberalization process. Therefore, the policy implications include two types of policy sequencing – macroeconomic stabilization and institutional development before financial liberalization. For example, Reference McKinnonMcKinnon (1993) proposed to achieve fiscal balance before financial liberalization. It is argued that domestic capital markets should be liberalized, and free international capital mobility should follow once fiscal balance and price stability are established. Reference Mishkin, Crockett, Dooley, Ahluwalia and FeldsteinMishkin et al. (2003) proposed to establish an effective regulatory and supervisory framework before financial liberalization to avoid financial crises because financial deepening may worsen the problem of information asymmetry without prudential supervision and regulation.

From the NSE perspective, financial repression in developing countries is often endogenous to the distorted production structure. Over the period of the 1950s and 1960s, governments believed in the necessity of development planning as a policy prescription to develop advanced capital-intensive industries for the industrialization and modernization, especially in the so-called backward countries. Giving strong support to this conviction in practice was the seemingly successful experience of adopting the Stalinist system in the Soviet Union during the 1920s and 1930s even as the Great Depression plagued the Western market economies. Therefore, numerous governments introduced interventions and distortions to guide the process of economic development and industrialization after World War II, especially in developing countries. Capital is the most important input and cost for the development of capital-intensive industries. To implement the catching-up development strategy, in the financial sector, many governments imposed interest rate ceilings together with intervention on the credit allocation to support “leading sectors” (Reference LoizosLoizos, 2018). Other financial repression policies were also put into effect, albeit with minor variations across countries, such as controls on international capital flows and restrictions on entry and competition in the financial sector.

Financial repression is endogenous to the need for making distorted production structure feasible, which should not be overlooked when implementing financial reforms. Specifically, these “leading sectors” established in developing countries are mainly capital-intensive, not in line with these countries’ comparative advantages, and cannot survive without governments’ protection and subsidies given the fact that capital is scarce in these countries. As a result, financial repression, such as interest rate controls and other restrictions on financial intermediaries, is treated as an indirect way for governments in these developing countries to support these capital-intensive leading sectors. As a result, these leading sectors can be financed at lower lending rates and with higher credit accessibility than they would be without such policies. The aforementioned argument can shed new insights into the economic instability and financial crisis in these developing countries when they went through a financial liberalization process in the 1970s and 1980s. When focusing on the possibility that financial repression is inherent in the distorted economic structure – that is, to support the nonviable leading sectors from the financial aspect – the radical removal of financial repression will lead to a rapid rise in borrowing cost as well as a decline in the credit accessibility in those leading sectors, resulting in large-scale bankruptcies, stagnant economic growth, and economic instability, as well as financial crises.

As for financial reform, NSFE emphasizes the necessity of identifying the root cause of the distortions in the financial system and implementing pragmatic financial reforms, that is, formulating appropriate financial reform goals and paths. In terms of reform goals, several things should be taken into consideration. The first one is the endowment structure, which specifies the mix of different production factors, such as capital, labor, and natural resources, and determines the (latent) comparative advantages of an economy. The second one is production structure, which specifies the mix of industries/sectors of different capital intensity and technology intensity used in each industry. Financial reform goals should meet the requirements of the real economy, which may differ at different development stages (Reference Allen, Bartiloro, Gu and KowalewskiAllen et al., 2018), rather than blindly treating the financial structure of developed countries as a benchmark and pursuing the so-called advanced financial institutions in developed countries. For instance, if an economy remains in a relatively agriculture-oriented or labor-intensive production structure, it is better for the financial system to be bank-based, and the effort to create or further develop a capital market is not appropriate (Reference Allen, Bartiloro, Gu and KowalewskiAllen et al., 2018).

In terms of reform paths, it is better to implement gradual reforms by taking into account the economy’s endowment structure, industrial structure, and nonviable leading sectors’ profitability and competitiveness. As for financial repression and financial liberalization, if distortions in the real economy are induced by the nonviability of leading sectors and such sectors cannot be abandoned due to national security or social stability concerns, then it is not appropriate to eliminate financial repression policies – interest rate controls, credit allocations, and so on – simultaneously and aggressively in a short time like the big bang or shock therapy (Reference Lin and TanLin & Tan, 1999).

China’s transition to a market economy is a pertinent example in support of the gradual approach of financial reforms. China introduced its first economic reforms in the late 1970s. At that time, China’s production structure was severely distorted because the majority of firms were state-owned, which dominated industrial sectors in every aspect and were mainly either capital-intensive, defying China’s factor endowment structure and comparative advantage, or inherited some types of policy burdens, such as employing redundant workers to provide jobs from the previous development planning system. Consequently, these SOEs were nonviable and could not survive without external assistance (Reference Lin and TanLin & Tan, 1999; Reference Lin, Cai and LiLin et al., 1998). To support these nonviable SOEs for strategic purposes, the government persistently allocated additional credits and provided other assistance through the financial system to the SOEs (Reference Lin and TanLin & Tan, 1999). That is why a highly centralized banking system was created with strict access restrictions, universal interest rate controls, along with an allocation of credit that was guided by the government to SOEs.

Over the past forty years, China has undergone a spectacular economic and financial transformation involving fast economic growth, sustained capital accumulation, gradual financial liberalization, and sustained financial deepening (Reference Song, Storesletten and ZilibottiSong et al., 2011), in sharp contrast to transitional countries that have introduced “shock therapy.” From the NSE perspective, the key ingredient of China’s success lies in its pragmatism in both reform goals and reform paths. As for reform goals, China did not take the financial structure of Western industrial countries as its benchmark; instead, China’s financial liberalization process had always tried to avoid the collapse of unviable but essential SOE sectors that were established when implementing the heavy industry-oriented strategy from the 1950s to the late 1970s. As for reform paths, the distinctive characteristics of China’s transition were “economic reform comes first, and financial reform follows closely” and “feel the stones to cross the river” (i.e., instead of following a blueprint to undertake sweeping reforms, policymakers should pragmatically conduct experiments first to create a feedback loop to adjust the plan). For one thing, China promoted the development of labor-intensive industries that were in line with its comparative advantages to accumulate capital, create jobs, and maintain social stability while still supporting nonviable SOEs through all kinds of financial repression policies. Capital accumulation under the “dual-track” reform system induced sustainable changes in China’s endowment structure and comparative advantage so that more and more SOEs in the leading sectors became viable in the competitive market. Later on, as the viability in the SOE sectors was improved, the government subsidies for SOEs were no longer necessary. Hence, financial repressions aimed at supporting SOEs also became unnecessary, and financial restrictions and distortions were eliminated step by step. For instance, Figure 3 depicts the gradual removal of bank entry restrictions as well as interest rate controls, where the vertical axis represents banking concentration in China’s banking system.Footnote ¹³ As Figure 3 illustrates, between 1990 and 2017, the concentration in China’s banking system decreased from 90 percent to 35 percent, and interest rate controls were also eliminated step by step. In 1990, for example, the total credits of the largest five banks accounted for 88 percent of the entire banking system, and two years before that (in 1988) banks were allowed to charge loan interest rates 20 percent higher than the base rate for the first time. In 1999, the fraction of the largest five banks’ total credits decreased to 63 percent, and the interest rate controls were relaxed at the same time; that is, the interest rate on loans to small- and medium-sized firms can be 30 percent higher than the base rate and 10 percent higher for big firms. In 2004, the upper bound on the loan interest rates and the lower bound on the deposit rates were removed. In 2013, the lower bound on the loan interest rates was freed, and two years after that the regulator eventually removed the deposit interest rate cap.

Figure 3

Gradual removal of bank entry restrictions and interest rate controls and the decline in the banking concentration in China

Data sources: WIND; official website of the People's Bank of China.

China’s pragmatic financial reforms have also encountered some difficulties. For example, in 1990s, the stated-owned large banks of China severely suffered from bad loans and fell into the dilemma of technical bankruptcy. The primary cause was that in the late 1970s there remained many SOEs, which were nonviable but inherited some types of policy burdens from the previous state planning system (Reference Lin and TanLin & Tan, 1999; Reference Lin, Cai and LiLin et al., 1998). To support nonviable SOEs for strategic purposes, the government had to persistently allocate additional credits and provided other assistance through the financial system to the SOEs, so that banking faced high risk of bad loans. Despite these problems, the financial repression under the pragmatic reform has played a significant role in maintaining national defense security, safeguarding national economy and the people’s livelihood and maintaining social stability in the early stage of the transition from the planned economy to the market economy. This helps to support economic stability and growth in the long run. By contrast, adopting shock therapy to eliminate the financial distortions is likely to threaten economic and financial stability if policy burdens imposed upon enterprises are not solved. Although it is necessary to adopt a dual-track approach to reforming financial systems, it is important to make timely reforms to gradually liberalize the financial system to solve the problems of bad loans.

In summary, financial distortions or repressions may cause the financial structure to deviate from what is appropriate. From the NSFE perspective, the financial distortion originates from the catching-up strategy of developing capital-intensive industries that defy comparative advantage. Hence, it is necessary for the government to adjust the financial structure through gradually eliminating the financial distortions along with the amelioration of production structure distortions rather than eliminating such distortions overnight.