Summary：China has entered a stage of development led by the service economy. Traditional economic theories predict that economic growth will slow due to the “cost disease” caused by low productivity in service industries. Will China follow this “law” and enter a period of slow growth? This paper assesses the sports and cultural service industries as cases for an empirical study of this hypothesis. We chose the sports and cultural service industries for two reasons. The first is that they were previously typical examples of low-productivity services and are now also typical services in internet space; therefore, some general conclusions can be made from studies of them. The second reason is that it has long been debated in China whether the sports and cultural service industries can be treated as industries and operate under market mechanisms, and their development has, to some extent, been restrained by this debate. Therefore, their nature as industries must be justified so that they can develop more quickly.

The main contents and contributions of this paper are as follows. First, the two industries are briefly described, including their size, structure, supply and demand characteristics, competition patterns, influences of network technologies, growth prospects and so on. We emphasize the following two points. Sports and culture activities share the general nature of service industries, where market players need competition pressure to motivate innovation and market expansion, and keep developing by satisfying market demand. They also have their own characteristics, unique business models and market structures, reflecting the complex and pluralistic features of service industries. Theoretical case studies from both perspectives are rarely found in academic literature.

Second, we focus on the productivities of the sports and cultural industries in the internet age. For a long time, sports and cultural services have been considered purely labor-intensive and necessary to deliver face-to-face, which is typical of low productivity. However, with satellite TV and the internet, both have changed from field services to internet services, which has substantially improved their productivity. For example, a show or sports game can be broadcast live to billions of people around the world. The productivities of fitness, training and other industries have also been significantly improved through online services.

Third, this paper presents an econometric analysis of demand maximization for professional sports. Using a dataset of 1680 CSL football games from 2011—2017, we carry out an ordinary least squares (OLS) regression on the factors influencing attendance at CSL football games using R software. We find that the overall levels of the teams and competitive balances within the league are the two most important factors for attracting an audience, implying that a balanced and highly competitive league can obtain the biggest market share. One policy implication is that while we are boosting sport industries with deregulation reforms and opening-up to introduce foreign players, effective measures are needed to prevent the gap between clubs from becoming too big. Such econometric analysis on sports games is new in China.

Fourth, we propose that time is a “real constraint” on service industries in the internet age. This is a theoretical innovation; so far, we have not found any research on this topic in the literature. Internet space can store and provide an enormous amount of services and content, resulting in extremely substantial economies of scale and almost limitless potential for service provision growth. However, service consumption is experiential, and time must be spent on it in person. Thus, time becomes the resource of scarcity and a hard constraint. As the total amount of time available to every person and country in a year is fixed, how can service grow once time is used to its maximum? We are not sure of the answer, nor have we found any research that addresses this question. Furthermore, the development of service industries in the internet age will cause a significantly widened income gap.

We conclude that for sports, culture and other service industries that can make effective use of internet space, productivities have been greatly enhanced and “cost disease” problems solved. From this point of view, entering a stage led by a service economy does not necessarily mean slow growth. However, “gross national time” may be a hard constraint and ceiling on service economy growth. If this judgment holds, it will bring both uncertainties to long-term economic growth prospects and severe challenges to the applicability of existing economic theories.

Keywords： Service Industries in Internet Space; Productivities; Constraints; Prospects; Sports and Cultural Industries

Summary： Replacing BT with VAT has been the most important taxation reform since 1994. The reform not only restructures the distribution relationship of fiscal revenue between the central and local governments, but also impacts the production behavior of service enterprises because of its tax-reduction effect. Because of the latter, replacing BT with VAT has been given special significance in the structural reform of the supply side that is being promoted in China. Because replacing BT with VAT affects the tax burden of manufacturing enterprises by extending the chain of deduction, it is also hoped it will promote the upgrading of China's manufacturing industry. However, the literature on replacing BT with VAT has focused on the tax-reduction effects. Although a few studies have considered the effect of replacing BT with VAT on industrial upgrading, they have used samples of either service or manufacturing enterprises whose main businesses are directly affected by VAT. Even when these studies find that replacing BT with VAT drove those enterprises to upgrade, this does not prove that the manufacturing industry in general is upgrading. Mixed operations are common in manufacturing enterprises, and the evidence has indicated that businesses that originally paid the business tax significantly increased in manufacturing enterprises which main businesses have been levied the VAT. However, evidence of an upgrade in the manufacturing industry has not been sufficient.

To separate evidence of an upgrade in the manufacturing industry from mixed operations, this paper distinguishes pure manufacturing enterprises from mixed manufacturing enterprises. Then, according to the direct effects of replacing BT with VAT, two types of enterprises are identified as the treatment and control groups, respectively. A difference-in-differences method is used to identify whether replacing BT with VAT drives the manufacturing industry to upgrade. Because China implemented VAT reform before replacing BT with VAT and this replacing has also been implemented step by step. To reduce the impacts of other reforms and measures as much as possible, the years 2010 to 2015 are chosen. Manufacturing listed companies are selected as the sample. Following Becker et al. (2013), Alstadsater et al. (2017), Zwick & Mahon (2017) and others, manufacturing listed companies are divided into mixed and pure manufacturing companies that are respectively identified as the control and treatment groups according to the extent to which they are directly affected by replacing BT with VAT. We find that replacing BT with VAT has led to the upgrading of the manufacturing industry. There are heterogeneous reactions, with the specialization of pure manufacturing enterprises improving and the specialization of mixed manufacturing enterprises significantly declining. These findings mean that the reactions to replacing BT with VAT differ significantly. The upgrading path chosen by pure manufacturing enterprises is significantly different from the upgrading path chosen by mixed manufacturing enterprises identified by Chen & Wang (2016) and Fan & Peng (2017).

Further evidence indicates that there have even been significant differences in pure manufacturing enterprises since the replacement of BT with VAT. In pure manufacturing enterprises where the original technology level is relatively low, the proportion of outsourcing technical services significantly increases after the replacement of BT with VAT. Non-state-owned pure manufacturing enterprises are more inclined than state-owned pure manufacturing enterprises to increase their R&D investments after following the change. These conclusions are robust to a variety of analyses, such as different methods of identifying the treatment and control groups and changing the estimation methods.

Keywords： Replacing BT with VAT; Upgrading; Heterogeneity

Summary： The Chinese government has put a heavy emphasis on structural reforms on the supply side with economic restructuring at the core. Supply-side structural reforms ensure sustained economic growth. Demand management creates a favorable macro-environment for supply-side structural reforms by appropriately expanding aggregate demand. Therefore, the supply and demand sides interact with each other to jointly promote economic growth. Cooperation and coordination between the supply and demand sides needs to be considered when decomposing the trend and cycle of China's economic growth.

There are two main methods for decomposing the trend and cycle of economic growth: production function decomposition and statistical decomposition (Osman, 2011; Andrei & Paun, 2014). Production function decomposition starts from the supply side and has two advantages: it has a solid theoretical basis and can identify the structural input factors that affect the output trend. Statistical decomposition decomposes the data in a data-speaking manner. According to real business cycle theory, if technology progress is a drifting random walk process (King et al., 1988), there are common deterministic and random elements to the trend components of actual output. Based on this, King et al. (1991) propose a decomposition method, KPSW, that imposes a co-integration relationship constraint on each variable in the model. This paper uses the KPSW method of Hecq et al. (2000, 2006) to decompose the trend and cycle of GDP under the two-wheel drive of the supply and demand sides.

The sample period of this article is from 1984 to 2004. Labor input, physical capital, human capital, the upgrading of industrial structures, institutional changes and technological progress are selected as the driving forces on the supply side. Demand-side drivers include consumption, investment and exports. In this paper, the supply-side driving force, demand-side driving force and GDP are composed of co-integration vectors, creating an error correction model. We impose the Vahid & Engle (1993) co-dependence period constraint on the error correction model to obtain a quasi-error correction model. This paper decomposes the trend and period of GDP based on this quasi-error correction model and uses a generalized impulse response function to calculate the mutual influence of the two-wheel driving force and its effect on economic growth.

We find that the supply- and demand-side driving forces directly promote economic growth and jointly promote the long-term trend and short-term cycle of GDP. By the two-wheel drive, the downward trend of economic growth during China's new normal period is the result of a combination of trend components and periodic components. Looking at the ability of the supply and demand sides to support the economic growth trend, the demand trend is smaller than the supply trend, but the gap between the two is shrinking. This decrease is due to the supply-side driving force falling faster than the demand-side driving force. The accumulation of human capital is not conducive to the increase of the three major demands, and the accumulation of physical capital is not conducive to the increase of consumption. The remaining supply-side driving forces have a positive impact on demand-side driving forces. Additionally, consumption and exports have a sustained positive effect on the supply side, but investment is not conducive to institutional changes, industrial upgrade and human capital accumulation.

This study leads to several policy recommendations. Long-term macroeconomic control policies should focus on the supply side, concentrating on upgrading industrial structures and accumulating human capital. Short-term macro-control policy should strive to expand total demand, focusing on the stimulation of consumption.

Keywords： Economic Growth; Supply Side; Demand Side; Common Trend; Codependent Cycle

Summary： Most research on the middle-income trap adopts the political economy perspective with an emphasis on how large income gaps lead to dissatisfaction in the poor, a confrontation between the rich and poor and social unrest in Latin American and Asian countries when entering the middle-income stage (Cai & Wang, 2014). Others have a one-sided emphasis on middle-income countries' weak economic growth due to the loss of cost advantages and a lack of technological innovation (Yao & Han, 2015). In addition to extreme examples that seriously affect social stability, more mundane factors such as demand pull, income level and income gap intermediate the effect technological innovation has on the formation of a middle-income trap. This complex mechanism is the key point to explore for the innovation-driven development strategy currently used in China to get out of the middle-income trap.

Using a vertical product differentiation model, this paper introduces a continuous symmetric trapezoid density function to describe the income distribution of consumers and to construct a mathematical model of the impact of income levels and income gaps on product innovation under innovation uncertainty. Combining this with a new product diffusion model, this paper develops a mathematical model of the impact of income level and income gap on economic growth from a macroeconomic perspective. Through model simulations, we depict the influence of income levels and income gaps on microeconomic independent innovation and macroeconomic growth under different growth patterns, revealing the formation mechanism and the escaping path of the middle-income trap. This paper uses panel data on 29 Chinese manufacturing industries from 1998 to 2008 from the China Statistical Yearbook on Science and Technology to verify the model.

This paper concludes that different combinations of income levels and income gaps have totally different impacts on the independent innovation behavior of domestic enterprises in the different stages of economic development. In the low-income stage, the expansion of the income gap will not inhibit independent innovation and economic growth. In the middle- and upper-income stages, if the income gap does not shrink when income levels increase, independent innovation will be inhibited and economic growth will stagnate. This microeconomic mechanism explains why many polarizing developing countries fail to progress to an innovation-driven growth mode and fall into the middle-income trap when they enter the middle-income stage. More importantly, if income gaps cannot be effectively reduced as income levels increase, developing countries cannot avoid economic stagnation even though R&D efficiency and the upper limit of product quality might increase. If income gaps can be narrowed when income levels increase, developing countries can get out of the middle-income trap successfully.

These results suggest that economic growth in developing countries does not necessarily bring about improvements in the independent innovation of domestic enterprises. To improve the ability of domestic enterprises to independently innovate, the Chinese government must focus on a policy of narrowing the income gap and adopt the following policy of demand pull and technology push. First, from the view of demand pull, the government should promote the reform of the income distribution system. Second, because the high Gini coefficient of China is mainly due to the large income gap between urban and rural areas, the government could narrow the income gap by raising the level of urbanization to achieve inclusive growth. Finally, in terms of technology push, compared with increasing the quantity of R&D input, the government should pay more attention to the institutional improvement of innovation to enhance R&D efficiency.

The most important innovation of this paper is the construction of a product innovation model of domestic enterprises with a continuous income distribution under innovation uncertainty, breaking the limitation of an income dichotomy and ignoring different types of enterprises and consumers matching. This clarifies that different combinations of income levels and gaps have different influences on innovation by domestic enterprises and economic growth, providing a reference for China to implement an innovation-driven development strategy and cross the middle-income trap.

Keywords： Income Level; Income Gap; Independent Innovation; Middle-income Trap

Summary： Financial structure is the medium of money policy transmission, so monetary policy transmission changes with variations in financial structure. Recently, internet finance has grown rapidly in China. The scale and growth speed of internet finance in China is the largest and fastest in the world. According to iResearch (2017), there are 500 million users of Chinese internet banking and 200 million users of the lending network. Undoubtedly, internet finance is a large shock to China's financial structure that will significantly influence monetary policy transmission.

This paper asks how internet finance influences the bank lending channel of China's monetary policy transmission. There are two reasons to care about this question. First, unlike in developed countries, China's finance sector is still in the process of full liberalization. Until 2016, bank loans accounted for 69.86% of total social financing, so the bank lending channel plays an important role in China's monetary policy transmission. Second, the relationship between internet finance and the bank lending channel is ambiguous. Internet finance may reduce financial frictions, weakening the transmission effects. However, internet finance would not affect bank credit transmission if there were no substitution between internet finance assets and bank credit. Therefore, further research into the relationship between internet finance and the bank lending channel is needed. In this paper, we build a general equilibrium model that includes representative economic agents and put forward four hypotheses about how internet finance can influence the bank lending channel. We then empirically test these four hypotheses.

The main empirical technology of this paper is the generalized method of moments/dynamic panel data (GMM/DPD). This method allows us to resolve the problem of endogeneity and improves the validity of the estimates. Our data include variables on bank loans, monetary policy, internet finance, financial frictions, shadow banking and bank size. Most of the macro and micro variables are available from Wind. The data on internet finance mainly comes from open network resources like iResearch and ERI. Missing values are completed with the moving average method.

Our analysis leads to the following conclusions. First, the empirical results prove that internet finance weakens the bank lending channel of monetary policy transmission by reducing frictions in the financial market. Second, through the optimal finance decision strategies of families, firms and commercial banks, internet finance influences the monetary policy lending channel. There are four effects in this process: a bank debt structure effect, a liquidity effect in securities markets, an effect resulting from the mismatch of financial resources and a corporate finance structure optimization effect. Third, the empirical results support there being significant effects of bank debt structure on the bank lending channel. Initial findings of no significance are due to the limited scale of securities markets in China and the effect resulting from the mismatch of financial resources offsetting the financing structure optimization effect. The policy implications of this research are that the implementation of monetary policy should be sensitive to the structure of financial markets and that more attention should be paid to the “catfish effects” to traditional financial institutions caused by internet finance.

An avenue for future research is to explore the broader impact of internet finance on monetary policy transmission.

Keywords： Internet Finance; Monetary Policy; Catfish Effects; Bank Lending Channel

Summary： Central banks communicate with the market about macroeconomic performance and monetary policy intentions to coordinate market expectations. While central banks in developed countries have placed a growing emphasis on communication and information disclosure, the People's Bank of China (PBC) does not pay enough attention to public communication. The Chinese economy is experiencing structural transformations in many areas, and there are calls for monetary policy to play a greater role. However, due to financial innovation, the effectiveness of quantitative monetary policy has decreased, while the price-based monetary policy has not been well established. Could the PBC improve the effectiveness of monetary policy by communicating more to the public? Prior studies recognize the positive role of central bank communication, but few investigate the mechanism through which a central bank's communication strategy impacts the effectiveness of monetary policy using DSGE models.

This paper introduces adaptive learning, public information and private information into a standard DSGE model to study the impact of central bank communication on public learning and expectations coordination in a dynamic setting. We assume that individuals know the correct structure of the economy but do not know the exact parameters. Therefore, they adaptively learn from public and private information to update their parameter estimates and form expectations. Due to information noise and learning biases, the public's expectations deviate from rational expectations. We show that if the central bank's public information is more accurate than private information, central bank communication can improve market expectations by reducing information and learning biases, accelerating the convergence toward the rational expectations equilibrium. This indicates that central bank communication can improve the effectiveness of monetary policy. Welfare analysis implies that the central bank can stabilize output and inflation through communication, reducing the gross social welfare loss. In our baseline calibration, central bank communication increases welfare by 15% relative to non-communication. The more accurate the central bank's information is, the less the welfare loss is and the more effective the monetary policy is. Moreover, when the central bank's public information is more accurate than private information, the public increases the relative weight they place on public information. This reflects a greater coordination motive and will further enhance the effectiveness of monetary policy. If public information is less accurate than private information, then central bank communication may magnify the information noise and decrease social welfare.

We contribute to the literature in two ways. First, we discover a new channel through which central bank communication affects the effectiveness of monetary policy. We decompose the public's expectation bias relative to rational expectations into an information bias and a learning bias. Central bank communication can reduce both biases by improving the precision of the information it discloses. Communication not only improves information quality, but also affects the public's learning process. The interaction between the public's adaptive learning and the central bank's communication is important. The model we construct and the channel we identify shed new light on the theory of how central bank communication affects the effectiveness of monetary policy. Second, we prove that this effect is limited. Prior studies often assume that the central bank communicates information that is superior to private information. We investigate the case when the central bank communicates less accurate information than the public in a dynamic macroeconomic environment and find that central bank communication in such a case tends to hamper the public's learning process and decrease the effectiveness of monetary policy. Therefore, to make monetary policy more effective, it is important for the PBC to increase the precision of the information it discloses.

Keywords： Central Bank Communication; Adaptive Learning; Monetary Policy; Public Information; Private Information

Summary： With the quick increase in direct financing in China, market interest rates are playing an increasingly important role in financial markets and the economy. How market interest rates are determined and their relation to economic fundamentals are key issues of concern to businessmen, academics and policymakers as China liberalizes interest rates. Departing from previous studies, we construct an affine interest rate model with three new factors as state variables, adopting characteristics of China's monetary policy and financial markets.

We use the one-year base deposit interest rate set by the central bank for commercial banks as our first factor. Because funding liquidity in the bond market causes the short-term market interest rate to differ from the official base interest rate, we use the part of the short-term market interest rate that cannot be explained by the one-year base deposit rate, a measure of funding liquidity in the bond market, as the second factor. The long-term market interest rate is determined by the average of the expected short-term rate in the future plus a risk premium. This risk premium is the third factor.

We fit the model to data on the government bond interest rates of various terms and the one-year base deposit rate from 2002 to 2017. The model fits well with standard deviations of fitting errors in the market interest rates less than 1.5 basis points. The model exposes how the market rates of various terms are determined and shows that the one-year base deposit rate is an important variable for determining both the short- and long-term market interest rates. A change in the base rate of 100 basis points corresponds with a 48-basis-point change in the one-year market rate and a 37-basis-point change in the 10-year market interest rate. The funding liquidity factor contributes to determining the change in short-term market interest rate, explaining 70% of the variance in the one-year market interest rate and only 3% of the variance in the 10-year market rate. The risk premium factor contributes most to changes in the long-term interest rate, explaining 80% of the variance. As the risk premium factor is so important for the long-term interest rate, it is natural that long-term bond excess returns are highly predictable. These three factors explain 63% of the variance in annual excess returns of the bond index portfolio consisting of government bonds with maturities of 7 to 10 years.

The three-factor affine model also helps us to understand how the term structure of market interest rates is related to economic fundamentals. According to the Fisher hypothesis, the nominal interest rate has two components: the real interest rate plus the expected inflation. Cieslak & Povala (2015) find that expected inflation explains 71% to 89% of the variation in U.S. Treasury bond yields. We use the constant gain learning method of Sargent et al. (2002) to construct our measure of expected inflation. The empirical evidence shows that the relation between market interest rates of various terms and the expected inflation rate is weak. Expected inflation contributes 10% to the variance of market interest rates. This relation has not strengthened with the recent development of capital markets and changes in monetary policy. The three-factor model explains why. The official base rate increasingly follows expected inflation. However, the funding liquidity factor is negatively correlated with expected inflation in both the first and second halves of the sample period, and the risk premium factor is only weakly correlated with expected inflation. Funding liquidity is the main reason for the weak relation between market interest rates and expected inflation.

These findings are relevant to the current interest rate liberalization process in China. To have market interest rates closely reflect economic fundamentals, the control over interest rates must first be relaxed. The central bank can keep an eye on funding liquidity in the bond market and keep funding liquidity tense so that the short-term market rate increases when there is an expectation of high inflation.

Keywords： Market Interest Rates; Benchmark Interest Rate; Expected Inflation; Affine Model
 

Summary： In the current “new normal” economic period, many good firms in China face liquidity risks or go bankrupt because banks refuse to roll over loans or demand early repayment. To address this issue, practitioners have proposed loosening loan rollover standards for firms and reinstating the policy of “borrowing for repaying”. The government has listened to these suggestions and gradually eased regulations on rollover policies for bank loans. This study investigates whether loosening rollover standards is beneficial for firm value.

In 2007, China passed a law, the “Guideline for Loan Risk Classification”, prohibiting the practice of “borrowing for repaying”. The law stipulates that when firms borrow new loans to repay due loans, the new loans are classified as non-performing loans for commercial banks. Prior to 2007, loans used to repay due loans were treated as normal debt. Consequently, the 2007 law restricts banks' rollover practices and constitutes an exogenous shock to the loan rollover standard. We investigate how this rollover restriction affects firms.

The 2007 law change is a nationwide shock, allowing us to use the difference-in-differences methodology to identify the loan rollover restriction's effect on firms. We define treated firms as those whose dependence on short-term loans is higher than the median and control firms as those with below-median dependence. Our results show that the rollover restriction has three effects on firms: a liquidity risk effect, a corporate governance effect and a credit resources allocation effect. The liquidity risk effect is a reduction in loans, especially short-term ones, by firms and an increase in cash holdings and working capital. The corporate governance effect is a reduction in inefficient investment and less expropriation by the controlling shareholders. The credit resources allocation effect is good firms obtaining more loans and fewer zombie firms forming. The liquidity risk effect is harmful to firm value, while the other two effects benefit firm value. Overall, the rollover restriction has a positive effect and improves firm performance. Our results indicate that although loosening rollover standards can mitigate firms' liquidity risks, loosened standards disrupt banks' monitoring capacity and reduce the allocative efficiency of loans. This harms firm value at the micro-level and is not good for the structural transition and industry upgrade of a transitional economy.

Our work makes two contributions to the literature. First, it extends the research on the monitoring role of short-term loans. Myers (1977) and Stulz (2000) argue that banks use short-term loans to monitor firms. Short-term loans often mature before the final date of an investment project. When firms need to roll over short-term loans, banks get to decide whether to roll them over. If firms exhibit opportunistic behavior, banks will refuse to roll over the loans. Barclay & Smith (1995) and Datta et al. (2005) confirm that increasing short-term debt can enhance monitoring. We emphasize that changing the rollover policy of banks can also enhance the monitoring capacity of short-term loans.

Second, our work adds to the evidence of an effect of government policy on commercial banks. Bertrand et al. (2007) contend that government intervention can decrease the efficiency and performance of the banking industry. In contrast, we find that when the government tightens regulations on loan rollovers, bank loans can be allocated more efficiently and firm performance improves. Banks have more incentives to review and monitor borrowers under the new regulation. Therefore, we need to scrutinize how government policies affect the incentive structure of commercial banks before judging their value.

Our work also has implications for future research. We focus on the function of the maturity clause of short-term loans. Whether regulations can support the monitoring role of other loan clauses, such as collateral clauses and covenants, deserves further discussion.

Keywords： Bank Loans Rollover; Short-term Loans; Liquidity Risk; Corporate Governance; Credit Resources Allocation

Summary： Population aging is a severe challenge for China. Using data from “World Population Perspectives: The 2017 Revision”, we calculate that people aged 65 and above accounted for 9.68% of China's population in 2015. This percentage will increase to 26.30% in 2050 and 31.72% in 2100. Population aging results in greater need for elderly care. As China's fertility rate has declined, young people have devoted more time to work, and elderly people's needs for care have diversified. In this circumstance, family elderly care alone cannot meet the elderly people's care needs upgrading, and social elderly care can also play an important role in this process.

This paper focuses on two questions concerning China's reality of declining fertility and increasing elderly care needs. First, this paper investigates the effect of elderly care preference on fertility. If elderly care preferences decrease fertility, the population aging problem will worsen with the unbalance between elderly care supply and demand. Second, this paper investigates whether the market or government should provide social elderly care. We compare two modes of social elderly care provision—market and government provision—from the perspective of fertility and social welfare, giving us a clear understanding of the macro-economic effects of these two different provision modes.

This paper uses an overlapping generation model with both family elderly care and social elderly care. We find that both non-altruistic elderly care preferences and the structure of elderly care preferences have indeterminate effects on fertility. Altruistic elderly care preferences decrease fertility. Compared with government provision, the fertility rate is higher when the market provides social elderly care. The transmission mechanism is as follows. When the market provides social elderly care, young people transfer more to their parents, giving old people stronger incentives to bear more children during their youth. From the social welfare perspective, market provision also dominates government provision. More family elderly care results with market provision due to increased fertility. The higher transfer to parents with market provision also leads to more consumption in the old period and more social elderly care.

The paper contributes to the literature in three ways. First, we present a framework that includes both family elderly care and social elderly care. We differentiate between market and government provision for social elderly care, enriching the theoretical literature on elderly care provision. Second, we analyze the effect of non-altruistic elderly care preferences, altruistic elderly care preferences and the structure of elder care on fertility, extending the literature on fertility determinants. Third, comparison analysis between market and government provision of elderly care lends theoretical support for the marketization of elderly care provision, filling a gap in the public economics literature.

This paper confronts the negative effect of elderly care preference on fertility and solves the dilemma of elderly care needs increasing and fertility declining, aggravating the population aging problem as the unbalance between elderly care demand and supply worsens.

This paper has three policy implications. First, the government could adopt grant-in-aid and tax preferences for elderly care industries and enterprises to encourage lower prices for elderly care. This would reduce elderly people's expenditure on elderly care so they can devote more income to childbearing.

Second, the government could establish a special fund for encouraging young people to provide family elderly care for their parents. Young people can get both spiritual satisfaction and income compensation.

Third, the government could accelerate the process of marketization of elderly care provision to develop the specialized elderly care market to meet elderly people's individualized and diversified care needs. The government has an important role to play in basic service provision so that poor people and childless people can enjoy basic elderly care.

Keywords： Family Elderly Care; Social Elderly Care; Fertility; Social Welfare

Summary： Gender identity is a form of social identity. Social identities tell people how they are supposed to behave by providing them a reference group. Identity influences economic outcomes because deviating from prescribed behavior is inherently costly (Akerlof & Kranton, 2010). Bertrand et al. (2015) report that the norm in the U.S. that a man should earn more than his wife affects behavior in both marriages and labor markets. While studies of social norms have been conducted in Western countries (e.g., Bertrand et al., 2015; Wieber & Holst, 2015; Hederos & Stenberg, 2015; Karen et al., 2017), less has been learned about the role of social norms in China.

In feudal times, few Chinese women worked in the market, and a belief in female inferiority was prevalent. However, since the 1950s, the Chinese government has promoted a women's liberation movement. The slogan that “women hold up half the sky” has been widely used to encourage women to join the workforce. This policy has enhanced the status of women, and female participation in the labor force has reached an unprecedented level. This paper explores three questions. Is there any evidence to support the existence of gender identity in contemporary China? What impact does gender identity have on marriages and labor market behavior in China? What are the channels through which gender identity acts？

We answer these questions using data from the 2005 China census and 2010—2014 China Family Panel Studies. First, we calculate the distribution of relative household income earned by the wife. We find a sharp and statistically significant drop in the distribution where the wife's income equals that of the husband. This discontinuity appears in all subsamples, suggesting that gender identity is prevalent in China's society. Second, we estimate the impacts of gender identity on the marriage market, the labor market and household production, where relevant marriage markets and potential earnings are constructed based on demographic characteristics. We find that gender identity hinders matching in the marriage market by depressing the marriage rate and postponing first marriage. Gender identity also distorts wives' labor market behavior. Women tend to quit the labor market, choose jobs with lower salaries and do more housework. Once gender identity is broken in a family, there is a positive correlation between the divorce rate and a wife's relative income. The correlation is negative between the satisfaction level with respect to a partner's financial or housework contribution and a wife's relative income. Finally, preliminary findings hold that women with rural residence, low education level, long marriage duration and more children are more susceptible to pressures of gender identity.

This paper provides empirical support for the existence of a norm in China that a man should earn more than his wife. By focusing on the interaction of gender identity and relative household income, we provide evidence for a value to studies that use an interdisciplinary approach, including insights from sociology and psychology. Our paper contributes to the relevant literature and facilitates comparisons between cultures. The impact of gender identity on marital and job stability in China we find is less than that found by studies of Western countries. However, gender identity is universal and has implications for Chinese society. Finding ways to diminish the negative impacts of gender identity could help improve the quality of marriage matches, decrease labor market distortions and improve family relationships. Modifying gender attitudes could also play an important role in promoting labor market efficiency, family harmony and social stability. Our results suggest that modernization and education could help mitigate the negative impacts of gender identity. Families with fewer sunk costs, such as those with no children and shorter marriage durations, suffer less from gender identity.

Keywords： Gender Identity; Marriage Match; Labor Participation; Housework

Summary： The entrepreneurship-enhancing effect of urban agglomeration has attracted increasing attention in the economic literature over the last decade. Findings on this effect have been important motivations for the Chinese government's new urbanization strategy. However, studies have focused on the U.S.A. and European countries. Theoretical and empirical analysis for cities in developing countries, including those in China, is lacking. This paper fills this gap by assessing the extent to which urban agglomeration affects entrepreneurship in China.

We begin our analysis by proposing a mechanism-based framework to study the precise channels through which urban agglomeration contributes to entrepreneurship. Our theoretical framework shows that different types of urban agglomeration affect entrepreneurship in different ways. Specialization externalities promote entrepreneurship through specialized labor market pooling, specialized production networks, and within-industry knowledge spillovers. Diversity externalities promote entrepreneurship through inter-industry labor sharing, input-output linkages, and inter-industry knowledge spillovers. As a transforming economy, the challenges currently facing China's specialized industrial districts limit the entrepreneurship enhancing effect of specialization externalities, which makes diversity externalities rather than specialization externalities have a lasting effect on entrepreneurship in China.

Based on this theoretical framework, we combine individual-level data from the Chinese Household Income Project (CHIP) with city-level data from the China Urban Statistical Yearbook to estimate the effect of urban agglomeration on entrepreneurship. The empirical analysis shows that diversity externalities have a positive effect on individual entrepreneurial choice, whereas specialization externalities have no impact on individual entrepreneurial choice. This finding holds true in a series of robustness checks. Our heterogeneity analysis shows that smaller and subsistence entrepreneurs benefit from both specialization externalities and diversity externalities, while bigger and transformational entrepreneurs only enjoy diversity externalities. The diversity externality estimates are generally significant for male entrepreneurs and service industries. Using Population Census micro-data and input-output table, we gauge the relative importance of different diversity externality mechanisms, showing that input-output linkages are the main mechanism through which diversity externalities promote entrepreneurship in China.

Our paper contributes to the growing literature on urban agglomeration and entrepreneurship in three ways. First, theoretical analysis of the impact of urban agglomeration on entrepreneurship is in its infancy. This study constructs a mechanism-based framework to explore the ways in which specialization and diversity externalities influence entrepreneurship and conducts a dynamic analysis of the relationship between them based on the features of urban agglomeration economies in China.

Second, the literature on urban agglomeration and entrepreneurship normally takes the firms' perspective. Our paper examines the effects of urban agglomeration economies on individual entrepreneurial choice by combining individual-level data and city-level data, which provides a new perspective to study this very important phenomenon. Meanwhile, our paper also assesses the impacts of different types of urban agglomeration economies on different types of entrepreneurship, providing a better understanding of the entrepreneurship-enhancing effect of urban agglomeration economies.

Third, the empirical evidence is scant on how urban agglomeration economies contribute to entrepreneurship. By constructing relevant indices, our empirical analysis on the mechanisms through which diversity externalities contribute to entrepreneurship fills a void in the literature.

Keywords： Cities; Agglomeration Economies; Entrepreneurship; Mechanisms

Summary： The financial media plays an active role in capital markets. Media biases such as political bias, gender bias, regional prejudice, foreign bias and minority bias harm the objectivity and authenticity of news reports. This paper analyzes whether the media cover news with a regional bias. Regional biases are unfair and negative social attitudes held by groups and individuals living in other regions based on regional differences. The media may report news with bias due to regional differences.

Although the regional bias of the media has already gained the attention of academics, studies have been limited to case studies or small-scale explorations and lacked detailed theoretical analyses and empirical tests. This paper investigates the causes of media bias from the macro perspective of regional differences. The objective existence of regional differences leads the media to unconsciously affix regional labels when reporting the news. Long-standing regional differences become stereotypes and gradually take root as social norms. Under this social pressure, the media inevitably report the news with bias.

As an empirical test, we manually collect 203,860 news reports about Chinese listed companies from eight mainstream financial newspapers, covering a period from 2004 to 2014. We read the reports and make judgments on their attitudes, whether positive, neutral or negative. Regional differences are measured in terms of economic development levels, system construction processes and social trust levels. This allows us to explore in detail the relationship between regional differences and media bias. Our results indicate that the media tend to give more positive coverage to listed companies in regions with more advanced economies, better constructed systems or higher levels of social trust. However, the results could also be interpreted to show that well-managed companies in developed regions deserve positive coverage while those in backward areas deserve criticism. To address this ambiguity, we introduce terms interacting company performance with regional difference variables. This helps us to verify that there are striking differences in the coverage for listed companies with similar performances in developed and backward regions. This paper measures company performance from two aspects: economic performance and non-economic business performance. For economic performance, we use accounting and market performance as our measurement. For non-economic business performance, we use corporate social responsibility as our measurement. Our results indicate that the media report news with a double standard. For listed companies in developed regions, the media tend to overstate their performance, while firms in backward regions struggle to receive recognition from the media even when they perform well.

Our paper contributes to the literature in several ways. First, the role of the media in corporate governance, capital markets and government action has become a focus of attention of scholars. However, studies have been based on the assumption that the media report without bias. This paper uses media bias as the breakthrough point to review the role of the media. Second, there is a lack of rigorous evidence on whether regional differences evolve into regional biases. This paper extends the research by analyzing regional bias from the perspectives of the economy, systems and trust. Third, by adopting a regional comparative perspective, this paper regards regional differences as the key factor in media bias. Our results indicate that even in the same country with similar economic systems and cultural backgrounds, regional differences profoundly affect media coverage, deepening our understanding of consequences of regional differences. Lastly, this paper innovates in the use of analytical methods for empirical research and event studies, which are rarely seen in mass communication studies.

Keywords： Media Bias; Regional Differences; Limited Recognition; Social Norm

Summary： When he constructed the scientific labor theory of value, Marx solved the biggest puzzle in the Ricardian School: the value of a commodity as decided by the socially necessary labor time is not consistent with the production price as decided by the equal return to factors in market competition. In Volume III of Capital, Marx used the categories of profit rate equalization and production price to construct transformation theory, connecting commodity value to production price. This proved that the production price of a commodity was based on the labor theory of value.

Marx's analysis did not gain wide acceptance and became known as the “one-hundred-year puzzle”. Critics questioned whether the logic of transformation theory would remain consistent if the transformation process were generalized into cost price. This paper follows the research method of Marx's transformation theory, generalizing the model of transformation to include cost price without finding any contradiction. Transformation theory became a one-hundred-year puzzle because researchers diverged from Marx's original research method. The key to solving the puzzle is to return to Marx's method and look for the answer in his logic.

In his study of value transformation, Marx used a five multi-sectors numerical example to explain the mathematical relation behind commodity value's transformation into production price through profit equalization. However, he did not continue to use numerical models to further study transformation because they did not provide a proper way for calculating invariable capital and variable capital transformation. As the full information about product technique composition and the input-output relation between sectors were unknown, the production price of invariable capital could not be calculated. This calculation is necessary to derive the connection between the aggregate value and aggregate production price of net products. As such, the transformation of variable capital is not understood. We extend transformation theory by including these missing tools, following Marx's research motivation.

This paper makes three major contributions. First, by analyzing Marx's logic, we find that Marx had to use the multi-sectors numerical example in his research. Although Marx's lack of today's advanced mathematical tools kept him from constructing a generalized transformation model, his numerical example is consistent with the theoretical perspective of the labor theory of value. This implies that any extension of transformation theory must rely on the integral logic of the labor theory of value. Any model that introduces superfluous conditions such as restrictions on reproduction and the real wage vector will certainly diverge from Marx's theory.

Second, we point out that according to Marx's logic, the net product production price is bound to deviate from its value. This is the key point for advancing research. If we admit that the net product production price will deviate from its value but insist on the surplus value not changing, then the surplus value rate must change. There is a lack of evidence for this in Marx's theoretical set-up. One requirement of Marx's transformation theory is that the aggregate surplus value rate, which reflects the comparison of powers in a given period, must remain unchanged. Profit equalization is a process of market competition between capitalists, and the power comparison between labor and capital and the rate of exploitation will not and should not change. Following this argument, the generalized transformation model of a C system we develop in this paper forces an unchanged aggregate surplus value rate. We give the equations for the production price in comparative static and dynamic iterations and prove the existence of solutions to these equations.

Third, we empirically test the value transformation model. We use China's 2012 input-output table to calculate a series index in the value and production price systems and compare the result with other transformation theories. Preliminary evidence shows that the production price of the C system is closer to the market price than other solutions.

Keywords： Value Transformation; Marx's Research Access; Material Consumption Coefficient Matrix; Production Price of Net Products; General Rate of Surplus Value