Pacific-Basin Finance Journal 59 (2020) 101262
Contents lists available at ScienceDirect
Pacific-Basin Finance Journal journal homepage: www.elsevier.com/locate/pacfin
Financial literacy and retirement preparation in China ⁎
Geng Niua, Yang Zhoub, , Hongwu Ganc
T
a
Research Institute of Economics and Management, Southwestern University of Finance and Economics, Chengdu, Sichuan 610074, China Department of Finance, School of Economics and Management, Wuhan University, Wuhan, Hubei 430072, China c Hanqing Advanced Institute of Economics and Finance, Renmin University of China, Beijing 100872, China b
A R T IC LE I N F O
ABS TRA CT
Keywords: Financial sophistication Retirement planning Financial planning Private pension
A growing body of literature, which primarily focuses on the developed world, investigates the implications of financial literacy (or lack thereof) on households' well-being. This paper examines the level of financial literacy and its impact on retirement preparation in China, a country that is growing old before getting rich. Drawing on internationally comparable survey questions, we find that a large proportion of Chinese people, especially the elderly, women, and under-educated, lack financial knowledge. The empirical results show that financial literacy has a strong and positive impact on various aspects of retirement preparation among Chinese people, including determining retirement financial needs, making long-term financial plans, and purchasing private pension insurance. Our findings suggest that concrete measures are needed to improve financial literacy so as to increase the awareness of retirement preparation in China.
JEL codes: D14 D91
1. Introduction Around the world, individuals are expected to bear more responsibility for their own financial well-being in old age. However, household surveys in different countries suggest that a substantial number of people are not capable of managing their financial affairs well (Agarwal et al., 2009; Calvet et al., 2007). In particular, individuals prepare inadequately for retirement. The literature suggests that financial literacy has a positive impact on individual households' propensity to plan for retirement. Although ample evidence of this relationship has been found in developed countries, such as the Netherlands and the United States (Van Rooij et al., 2011a; Lusardi and Mitchell, 2011), little is known about this relationship in the context of developing countries, where households are more likely to lack financial literacy and their well-being in retirement is under greater threat. The demographic challenge is more acute in Asia than in other regions of the world, and China, in particular, has the most rapidly aging population in Asia (United Nations, 2017). In this paper, we examine the level of financial literacy and its impact on retirement preparation in urban China. China presents an interesting and important case. Over the past several decades, China has experienced a significant increase in life expectancy coupled with a sharp fall in fertility, which indicates that the country is facing an unprecedented burden in supporting older persons. The dramatic social and economic transformations featured by the shrinking family size, changing social norms, and increasing labor mobility have made the traditional family support system for elderly people unsustainable. Moreover, as an effective welfare system has not been well established in China, government support for well-being in retirement is also inadequate. In particular, public pensions have limited coverage and provide few benefits. As a result, Chinese households are expected to be heavily self-reliant to sustain retirement. However, having undergone a long period with a centrally planned economy, in which the state guided economic decisions and market functions were heavily restricted, many Chinese people
⁎
Corresponding author. E-mail addresses:
[email protected] (G. Niu),
[email protected] (Y. Zhou),
[email protected] (H. Gan).
https://doi.org/10.1016/j.pacfin.2020.101262 Received 16 December 2018; Received in revised form 7 May 2019; Accepted 1 January 2020 Available online 07 January 2020 0927-538X/ © 2020 Elsevier B.V. All rights reserved.
Pacific-Basin Finance Journal 59 (2020) 101262
G. Niu, et al.
lack experience in financial decision making. As the financial markets in China are growing rapidly and sophisticated financial products are becoming accessible to households at an increasing speed (Liang and Guo, 2015), the importance of financial knowledge for Chinese households cannot be overstated. Based on a large-scale household survey implemented recently, we find that a substantial faction of Chinse households lack knowledge of basic economic concepts, let alone more sophisticated financial knowledge. Furthermore, there exists substantial heterogeneity in financial literacy among different social groups. In particular, women, the elderly, and people with lower levels of education show lower levels of financial literacy. After controlling for a large number of confounding factors, such as income and education, in multivariate analysis, we find that people with a higher level of financial literacy, especially sophisticated knowledge, are more likely to think about financial needs after retirement, have long-term financial plans, and hold private pensions. Moreover, we address the endogeneity of financial sophistication using instrumental variable analysis, which suggests that there is a causal impact of financial literacy on retirement preparation in China. Our study provides important implications for policy makers in China. As we focus on households in the urban areas of China, the level of financial literacy among individuals in our sample probably corresponds to the upper bound of the average level among the whole Chinese population. The widespread financial illiteracy among households should raise attention in both academic and political arenas. Given the large impact of financial knowledge on retirement preparation in China, financial education programs should be widely introduced and should be especially targeted at the groups of people who are most lacking in financial knowledge, such as the elderly and under-educated. The remainder of this paper is organized as follows. Section 2 introduces institutional background and the related literature. Section 3 describes the survey data and variables used in empirical analysis. In particular, the distribution of financial knowledge across demographic variables is presented and discussed. Section 4 performs the empirical analysis. Section 5 concludes the paper. 2. Institutional background and related literature 2.1. Demographic shift in China Over the past several decades, China has experienced a rapid decrease in mortality. Life expectancy at birth increased from 44 years in the 1950s to 76 years in 2015, only 4–5 years lower than the average for developed regions, such as Western Europe, and is projected to reach approximately 80 by 2050 (United Nations, 2017). Meanwhile, China has undergone a drastic fertility decline driven by both the one child policy and socio-economic development. The fertility rate in China decreased from 6.11 in the 1950s to 1.6 in the 2010s (United Nations, 2017). As a consequence of longevity improvement and fertility reduction, there are more elderly people and fewer children in Chinese families, and the average family household size shrank from 4.4 in 1982 to 3.1 in 2010 (National Bureau of Statistics, 2012). Correspondingly, the old-age dependency ratio increased from 7% in 1950 to 14% in 2014 and is expected to reach 44% by 2050 (United Nations, 2017). In China and other East and Southeast Asian countries, family has traditionally been the major source of old-age support. The elderly primarily rely on their adult children for material needs and personal care (Phillips, 2000). However, the traditional family support system is under pressure in China. The continuing decline in both fertility and mortality suggests that fewer children have to support more elderly in the future and therefore the family's ability to support its older members is weakening (Zhan et al., 2008). Moreover, the transition to the market economy in the last few decades of the 20th century has substantially raised work mobility and labor migration, which implies that many children have moved away from their parents. Furthermore, socio-economic development and exposure to Western culture are introducing values and norms that clash with the traditional code of intergenerational duty, which aggravates the strains on the family old-age support system (Phillips and Feng, 2015). 2.2. Pension system in urban China Until the mid-1980s, the pension system in China existed exclusively in the urban public sector, where workers were fully covered by the state budget for their pension (Croll, 1999). After a series of progressively introduced reforms, China has established a threepillar pension system that is heavily focused on the first pillar, i.e., the public pension. 2.2.1. The first pillar: public pension system China's public pension system remains decentralized and highly fragmented in practice (Li, 2014). Prior to 2015, the urban part of the system could be loosely divided into three schemes, namely the Urban Enterprise Pension Scheme (UEPS), the Pension Scheme for Civil Servant (PSCS), and the Urban Residents Pensions Scheme (URPS). The three schemes target different groups of the urban population and vary in their contribution and benefit formulas (Bateman and Liu, 2014; Fang and Feng, 2018). 2.2.1.1. The Urban Enterprise Pension Scheme (UEPS). The UEPS, established in 1997, is a compulsory pension scheme for urban enterprise workers in China. The UEPS consists of two elements: a pay-as-you-go defined benefit social pooling account and a funded defined contribution individual account. Employers are supposed to contribute 20% of employees' monthly salaries to the social pooling account. The employer's contribution base is normally capped at 300% of the average local salary and cannot be lower than 60% of the local average salary. At the same time, employees are required to contribute 8% of salaries to their individual accounts. When reaching the statutory retirement age, employees in this scheme with a contribution history of at least 15 years shall receive monthly pension benefits from both the social pooling account and the individual account. Monthly payments from the social pooling 2
Pacific-Basin Finance Journal 59 (2020) 101262
G. Niu, et al.
account depend on the ratio of the individual's salary to the local average salary before the retirement year and the number of contribution years. The value of the individual account pension equals to the balance in the individual account divided by the actuarial month.1 In 2005 the Ministry of Human Resources and Social Security (MOHRSS) of China set the target replacement ratio of UEPS at 59.2% for an employee working for 35 years. Self-employed workers can voluntarily participate in the UEPS by funding the plan solely from their own pockets. They are allowed to choose their own contribution base. Limits of the base are usually 60% and 300% of the local average wage. The total contribution is 20% of the base: 12% of the base is contributed to the social pooling account and 8% to the individual account. Selfemployed workers in this scheme, with a minimum of 15 years of contribution, can claim pension benefits after retirement. The benefit formula is similar to the one that applies to paid employees. 2.2.1.2. The Public Servants Pension Scheme (PSPS). Until very recently, civil servants and employees in government-funded institutions were covered by the PSPS, a special defined benefit pension that was financed by the state. This scheme required no contribution from individuals and provided generous pension benefits. The pension replacement ratio can reach up to 80–90% of an individual's pre-retirement income. To counter mass resentment over pension inequality and ease the public financial burden, in 2015 the PSPS was merged into the UEPS and the government started to require pension contributions by public sector workers. After the reform, new entrants to the public sector are subject to the same contribution and benefit rules as those in the UEPS. Some transitional arrangements apply to those already in the PSPS before the reform. 2.2.1.3. Urban Residents Pension Scheme (URPS). Since 2011, urban residents who are non-students, aged 16 and above, and who are not entitled to the UEPS can voluntarily participate in the newly established URPS. The URPS combines government subsidies and individual contributions. In this scheme, individual make annual contributions, normally in the 100 RMB to 2000 RMB range, and local governments provide yearly subsides, the amounts of which vary greatly across regions due to the differential in fiscal capacity. Personal contributions and local government subsidies are added to the individual account. Participants with a contribution history of at least 15 years are entitled to receive a basic monthly pension provided by the central government in addition to the benefits paid from the individual account. The basic pension was originally 55 RMB per month and increased to 88 RMB in 2018. 2.2.1.4. Problems of China's public pension system. Despite noticeable improvement over recent years, there exist concerns over China's public pension with respect to coverage, adequacy, and sustainability. First, a non-trivial amount of China's urban residents is still not covered by the public pension system. For instance, the mandatory UEPS in practice mainly covers employees of state-owned-enterprise and large- to medium-sized private enterprises, while many small and micro enterprises, due to high costs and slack law enforcement, do not provide a pension for many of their workers (Liu and Sun, 2016). In addition, for self-employed workers, participating in the UEPS is not mandatory and participating cost is high, which results in low pension coverage for this group (Jiang et al., 2018). Most notably, rural-to-urban migrants (hereafter“rural migrants”), the total number of which exceeded 270 million in 2014 (National Bureau of Statistics, 2015), are considerably excluded from the public pension systems. Only a small portion of migrants, mostly those with formal labor contracts, have access to the UEPS.2 Many rural migrants work in the informal sector and take up uncontracted, temporary jobs, under which pension plans are mostly not offered. Rural migrants not enrolled in the UEPS can opt to join the New Rural Pension Scheme (NRPS) in their home village, which however provides rather low pension benefits.3 Data from the 2011 Migrant Dynamics Monitoring Survey in China shows that more than 70% of the rural migrants do not participate in any public pension scheme (Qin et al., 2015). Overall, based on recent survey data, only around 60% of the working age population in urban China is covered by the public pension system (Jiang et al., 2018). Second, for many Chinese the public pension system does not provide adequate benefits. The proportion of workers covered by the generous PSPS is small. In addition, as PSPS was merged into the UEPS in 2015, public sector workers are likely to receive lower benefits in the future. As regard to the UEPS, the target replacement ratio (59.2%) is often not attained in the real world, as many employers set the their contribution bases at levels much lower than legally required and many workers stop contributing before the statutory retirement age (Cai and Cheng, 2014). In fact, the average replacement ratio of the UEPS has declined from around 70% in 1998 to around 44% in 2016 (Zhao and Mi, 2019). For the large number of non-employed urban residents and rural residents, the public pensions they can access aim to providing only a minimum standard of living in old ages. Third, and more importantly, the sustainability of China's public pension system is seriously at risk. The system heavily relies on publicly provided pay-as-you-go arrangements, which are under increasing financial pressure given the quickly aging population. Moreover, returns on publicly managed individual accounts are generally low due to restrictions on investment strategies. Furthermore, in practice, local governments often “borrow” the contributions of workers to meet governments' current pension liabilities, which worsens the pension deficit. The sustainability challenges pose considerable policy risks and uncertainties: people might receive less pension benefits in the future than promised by the current system due to policy change (Bateman and Liu, 2014). In this sense, sustainability problems threaten retirement security for all the residents regardless of the types of pension plans they are 1 The actuarial month is based on the estimated mean residual lifetime of retirees. It is 139 for workers retiring at 60 and varies for different retirement ages. 2 In some regions the local governments allow for lower levels of contribution from employers and individuals under the UEPS if the participants are rural migrants. Expected pension benefits will be reduced accordingly. 3 The NRPS, which targets rural residents, is similar to the URPS in urban areas in terms of contribution and benefit rules.
3
Pacific-Basin Finance Journal 59 (2020) 101262
G. Niu, et al.
enrolled in. 2.2.2. The second pillar: occupational pension plan The second pillar pension is mainly based on two voluntary supplementary occupational pension programs, namely, the enterprise annuities scheme for enterprise employees and the occupational pension for civil servants. In principle, an employer can formulate an enterprise annuity plan through consulting its employees. Both the employer and its employees have to make contributions to the plan. Assets under the enterprise annuity are managed by a fiduciary, which can be a self-governing council or an external corporation such as an insurance company or a bank. However, due to the high costs and lack of tax incentives, only a small portion of employers, primarily government agencies and large state-owned enterprises, have implemented supplementary enterprise annuity schemes or occupational pension schemes. As a consequence, the second-pillar is unavailable to most workers in China. By the end of 2015, only 23.16 million employees, equivalent to 6.7% of workers covered by the public pension, had access to the supplementary enterprise annuity plans (Dong and Wang, 2016). 2.2.3. The third pillar: private savings and individually arranged private pension The traditional value of frugality and the weak social safety net make personal savings an important old-age security arrangement in China. Household saving accounted for more than 20% of China's GDP in 2013, much higher than the global average (Zhang et al., 2018). However, household saving in China has started to decline in recent years and the decline is expected to continue and even accelerate in the future, especially among the younger generations (Zhang et al., 2018). For example, saving rate (in percentage of disposable income) was well below 10% among those aged below 30 according to data from the 2013 China Household Income Project (CHIP) (Zhang et al., 2018). In fact, household saving rates in other East Asian economies, such as Japan and Korea, experienced a rapid decline after reaching the income level that China is soon to reach (Zhang et al., 2018). In addition, for many working-age Chinese, much of their savings might have been depleted before or shortly after retirement due to rising private burden of expenditures on housing, children's education, and healthcare (Chamon and Prasad, 2010). Moreover, Chinese households primarily invest savings in bank deposits that offer rather low returns, eroding the real value of savings (Glaeser et al., 2017). Therefore, despite the current high saving rates, retirement security should still be an important concern for many Chinese. To prepare for retirement, individuals can also purchase private pension products, such as annuity insurance, that are offered by insurance companies. Recently, annuity insurance has been expanding fast in China at an average annual growth rate of 16.9% between 2001 and 2014. In 2014, annuity insurance income totaled 282.2 billion RMB and there were more than 69 million in-force policies covering 100 million people (China Insurance Regulatory Commission 2015). The Chinese government is also trying to stimulate the development of private pension insurance. For example, in 2018 China's Ministry of Finance launched individual income tax deferral on private pension in certain pilot cities (Fang and Feng, 2018). Given the problems of public pension systems, private pension insurance can potentially plan an important role in improving retirement well-being in China. However, due to lack of knowledge and experience, many Chinese are still unaware of or unfamiliar with private pension plans. 2.3. Importance of retirement preparation and financial literacy Given the collapse of the traditional family support model and the inadequacy of the pension system, Chinese people have to rely more on themselves to accumulate enough wealth for well-being in retirement. It has been widely documented that planning has a strong positive impact on household wealth. For example, it is found that households with even a small retirement plan can accumulate a level of wealth that is double that of those without any plan (Van Rooij et al., 2011a; Behrman et al., 2012; Van Rooij et al., 2012). It has been documented around the world that financial literacy affects people's financial decisions in many aspects such as wealth management, stock holding, and insurance demand (Jappelli and Padula, 2013; Liao et al., 2017a; Lin et al., 2017; Van Rooij et al., 2011b). In particular, the literature proposes that financial literacy is a key factor that positively contributes to retirement planning, with empirical evidence documented in many developed countries, such as the Netherlands, the United States, Germany, and Canada (Bucher-Koenen and Lusardi, 2011; Lusardi and Mitchell, 2011; Van Rooij et al., 2011a; Boisclair et al., 2017). However, studies in the context of developing countries, where the provision of financial education is inadequate or even absent, are still scarce. In particular, the problem of financial illiteracy could be severe in China as the overall education level is relatively low. Financial literacy is becoming increasingly relevant for Chinese households, as China's financial market is developing fast and investment opportunities for households are expanding. For example, Chinese stock market has become the second largest in the world since 2014. The mutual fund industry is also one of the fastest-growing industries in China (Liao et al., 2017b), now offering to retail investors a wide range of products such as equity funds, bond funds, index funds, money market funds, hybrid funds, Qualified Domestic Institutional Investor (QDII) funds, and ETFs. In particular, China's money market funds (MMFs) had around 3.6 trillion RMB (520 billion USD) in net assets in the early 2017, marking China the second largest market for MMFs globally (McLoughlin and Meredith, 2017). In addition, investors in China have (limited but growing) opportunities to get exposure overseas due to opening-up arrangements such as QDII, Shanghai-Hong Kong Stock Connect and Shenzhen-Hong Kong Stock Connect (Zhu et al., 2018). Moreover, China's rapid development of internet finance/Fintech, in many fields such as peer-to-peer lending, equity crowdfunding, online wealth management, and online insurance, has greatly improved the accessibility of financial products to households (Chen, 2016). With the ongoing development and opening up of China's financial market, Chinese households are likely to have increasingly more opportunities to participate in the financial market. Compared to previous work on the role of financial literacy in retirement security in China (Niu and Zhou, 2018), this paper 4
Pacific-Basin Finance Journal 59 (2020) 101262
G. Niu, et al.
makes several contributions. First, based on thirteen questions on financial knowledge in different domains, we comprehensively capture the concept of financial literacy by distinguishing between basic financial literacy and advanced financial literacy. In addition, we provide a detailed description of the distribution of the financial literacy levels among different social groups in China, which is helpful for making targeted policies. Furthermore, we explore the impact of financial literacy on retirement preparation from three different aspects, namely, thinking about financial needs after retirement, having long-term financial plans, and having private pension insurance. Finally, we include important confounding factors, such as measures of risk aversion, impatience, and overconfidence, to better capture the true effect of financial literacy. Above all, because whether people in developing countries, such as China, have sufficient financial knowledge to prepare for retirement is increasingly a scholarly and societal concern, our study fills an important research gap in this area. 3. Data and variables Our source of data is the China Family Panel Studies (CFPS), which were conducted by the Institute of Social Science Survey (ISSS) at Peking University in collaboration with the Survey Research Center at the University of Michigan. This data set is a nationally representative, annual longitudinal data set that includes information on Chinese communities, families, and individuals with a focus on the economic, as well as non-economic, wellbeing of the Chinese population. The survey was designed in a style similar to the Panel Study of Income Dynamics (PSID) in the USA and collects information on a variety of demographic and economic characteristics (Xie et al., 2014). The CFPS was first conducted in 2008. Consecutive waves of the survey were in the field in 2009 and every 2 years since 2010. In 2014, the ISSS added a new module to the survey that aimed at investigating and measuring various aspects of financial knowledge among Chinese people. Therefore, we restrict our sample to the 2014 survey, which covers 13,946 households across 29 provinces in China. Since only households residing in urban areas were requested to complete the financial literacy module, we restrict our sample to those households. The respondent to financial literacy questions in each household is the adult member that is most knowledgeable about the financial situation of the household. 3.1. Financial literacy The 2014 CFPS survey contains thirteen questions designed to measure financial literacy. Following the literature (Van Rooij et al., 2011a, 2011b, 2012), we split these questions into two sets to distinguish among different levels of financial knowledge. The first set of questions assesses basic financial literacy, which is a prerequisite for day-to-day financial transactions and decisionmaking. These questions explore the awareness of the current interest rate level, numerical skills, and the understanding of inflation, interest compounding, and time value of money. The exact wording of the five questions measuring basic financial literacy is provided in Appendix A. Table 1 summarizes the responses to basic literacy questions. Panel A reports the proportion of households providing correct, incorrect, and “do not know” answers to each of the five basic financial literacy questions, while Panel B reports the distribution of the number of correct, incorrect, and “do not know” answers to the five basic financial literacy questions. Somewhat fewer than seven out of ten respondents correctly answered the question on the time value of money. The percentage of correct answers drops to between 50 and 60% when we consider questions on interest rate level, interest compounding, and inflation. Approximately onethird of respondents provided incorrect answers to the question that relates to numeracy and a similar fraction did not know the current interest rate level in China. A striking observation is that only 15.5% of respondents were able to answer all five questions Table 1 Basic financial literacy. Panel A: Percentage of total number of respondents
Correct Incorrect Do not know
Interest rate level
Numeracy
Interest compounding
Inflation
Time value of money
55.37 9.00 35.63
44.89 32.82 22.30
50.49 25.02 24.50
58.80 18.16 23.04
68.28 13.85 17.86
Panel B: Summary of responses Number of correct, incorrect and do not know answers (out of five questions)
Correct Incorrect Do not know
None
1
2
3
4
All
Mean
12.46 38.77 53.17
12.23 33.85 17.77
15.34 19.09 6.99
20.42 6.54 5.99
24.08 1.55 6.80
15.47 0.19 9.29
2.78 0.99 1.23
Note: Panel A reports the proportion of households providing correct, incorrect, and “do not know” answers to each of the five basic financial literacy questions. Panel B reports the distribution of the number of correct, incorrect, and “do not know” answers on the five basic financial literacy questions. Percentages may not sum up to 100 due to rounding. See Appendix A for the exact wording of the questions on basic financial literacy. 5
Pacific-Basin Finance Journal 59 (2020) 101262
G. Niu, et al.
Table 2 Advanced financial literacy. Panel A: Percentage of total number of respondents
In general, investment opportunities with high returns have high risks. True or False? Buying a company stock usually provides a safer return than a stock mutual fund. True or False? Which bank formulates and implements monetary policy? Normally, which asset is riskiest? What happens if somebody buys the stock of firm B Which statement about mutual funds is correct? Which statement about banks' wealth management products is correct? Which statement describes the main function of the stock market?
Correct
Incorrect
Do not know
85.08 34.27 30.74 66.93 16.38 12.46 30.00 28.12
6.02 24.01 37.70 11.91 41.72 31.75 32.78 22.36
8.90 41.72 31.55 21.17 41.91 55.79 37.22 49.51
Panel B: Summary of responses Number of correct, incorrect and do not know answers (out of eight questions)
Correct Incorrect Do not know
None
1
2
3
4
5
6
7
All
Mean
7.83 22.75 28.87
14.60 20.42 13.07
20.32 19.48 10.06
19.55 15.11 9.06
16.08 11.88 8.16
9.87 6.76 9.06
7.15 2.98 8.22
3.27 0.58 7.83
1.33 0.03 5.66
3.04 2.08 2.88
Note: Panel A reports the proportion of households providing correct, incorrect, and “do not know” answers to each of the eight advanced financial literacy questions. Panel B reports the distribution of the number of correct, incorrect, and “do not know” answers on the eight advanced financial literacy questions. Percentages may not sum up to 100 due to rounding. See Appendix B for the exact wording of the questions on advanced financial literacy.
correctly. Overall, respondents display very limited understanding of basic financial concepts. To more comprehensively measure financial literacy, we employ a second set of financial literacy questions that capture more advanced financial knowledge related to investment and portfolio choice. Specifically, these questions assess the awareness of the central bank; knowledge of complex financial instruments, such as stocks, bonds, mutual funds, and banks' wealth management products; understanding of financial concepts, such as the tradeoff between risk and return and risk diversification; and the workings of the stock market. The exact wording of the eight advanced financial literacy questions is reported in Appendix B. A similar set of questions have been used to measure advanced financial literacy of Chinese households in Liao et al. (2017a), and the authors found that advanced financial literacy significantly affected households' stock holding behavior in China. Table 2 summarizes the responses to advanced literacy questions. Panel A reports the proportion of households providing correct, incorrect, and “do not know” answers to each of the eight advanced financial literacy questions, while Panel B reports the distribution of the number of correct, incorrect, and “do not know” answers to the eight advanced financial literacy questions. The pattern of answers turns out to be more alarming than that of the basic financial literacy questions. In general, the fraction of correct answers becomes much lower. For example, only 12.5% of respondents know how mutual funds work, and only 16.4% know how stocks work. A notable exception is that 85.1% of respondents are correct about the tradeoff between risk and return, which makes this question the easiest to answer out of all 13 financial literacy questions. In addition to the low level of accuracy, a substantial proportion of “do not know” responses is also observed. Approximately half of respondents do not know the answer to the question about the main function of the stock market and even more state that they do not know the answer to the question about the workings of mutual funds. As shown in Panel B, barely more than 1% of respondents are able to provide correct answers to all eight questions, while the proportion of incorrect answers or “do not know” responses on several questions is sizable. These results reveal the scarcity of advanced financial knowledge among Chinese households, which is in line with the results found in other countries, such as the Netherlands (Van Rooij et al., 2011a). To further assess the level of financial literacy among Chinese, given data availability, we compare distributions of households' responses to three most commonly used financial literacy questions across a number of developed and less developed countries. The three questions are the interest compounding question (Q3) and the inflation question (Q4) in our basic literacy question set, and the risk diversification question (Q2) in our advanced literacy question set. The countries selected include China, USA, Canada, Germany, the Netherlands, Japan, Russia, and India. Table 3 presents the corresponding results. Risk diversification seems to be the least understood concept among the three questions for households in all countries. Not surprisingly, Chinese households are less financially literate than households in developed countries such as USA, Germany, Canada, the Netherlands, and Japan. Russia ranks the lowest with respect to all three questions. India performs incredibly well, which is however due the fact that the corresponding sample includes primarily Indians with high socio-economic status (Agarwal et al., 2015). To capture the information on financial literacy from the responses to our two sets of questions, we follow the existing literature (Van Rooij et al., 2011a; Van Rooij et al., 2011b, 2012) and construct two types of literacy indices using factor analysis on the five basic and eight advanced literacy questions. The first index measures cognitive ability and very basic financial skills and has nothing to do with the stock market and complex financial instruments, while the second index measures more advanced financial knowledge. 6
Pacific-Basin Finance Journal 59 (2020) 101262
G. Niu, et al.
Table 3 Distributions of answers to three financial literacy questions across countries. Interest compounding
China USA Canada Germany Netherlands Japan Russia India
Inflation
Risk diversification
Correct
Incorrect
DK
Correct
Incorrect
DK
Correct
Incorrect
DK
50.5 64.9 77.9 82.4 84.8 70.5 36.3 81.0
25.0 20.5 13.2 6.7 5.2 15.4 30.8 14.0
24.5 13.5 8.8 11.0 8.9 12.5 32.9 4.0
58.8 64.3 66.2 78.4 76.9 58.8 50.8 79.0
18.2 20.2 17.7 4.7 8.4 10.8 23.1 6.0
23.0 14.2 16.1 17.0 13.5 28.6 26.1 16.0
34.3 55.5 59.4 61.8 51.9 39.5 12.8 79.0
24.0 13.3 9.4 5.9 13.3 2.8 51.8 6.0
41.7 33.7 31.3 32.3 33.2 56.1 35.4 16.0
Note: The table reports the percentage of households providing correct, incorrect, and “do not know” (DK) answers to each of the three financial literacy questions, namely interest compounding question, inflation question, and risk diversification question. Percentages may not sum up to 100 due to rounding and refusal to answer in some countries. The results for China are based on authors' own calculations. The results for other countries are drawn from Lusardi and Mitchell (2011) (USA), Boisclair et al. (2017) (Canada), Bucher-Koenen and Lusardi (2011) (Germany), Alessie et al. (2011) (Netherlands), Sekita (2011) (Japan), Klapper and Panos (2011) (Russia), and Agarwal et al. (2015) (India). It should be noted that the Indian sample includes mainly people with high socio-economic status.
Given the importance of the “do not know” responses outlined above, we explicitly take into account the distinctions between incorrect answers and “do not know” answers in constructing the two indices. Specifically, we create two dummies for each of the 13 literacy questions to assess whether the response is correct and whether the response is “do not know” in the factor analysis. For both sets of literacy questions, we find that one factor can explain a large proportion of the variations in responses. From a policy point of view, it is instructive to examine financial literacy among subgroups to figure out who knows the least. Table 4 reports the distribution of the basic financial literacy index. Obviously, basic financial literacy varies considerably across demographic variables, such as education, age, and gender. As expected, the basic financial literacy level rises sharply with education. Approximately half of respondents with primary education are located at the bottom quartile of basic literacy. As we move to higher quartiles of basic literacy, the percentage of respondents with high levels of education increases; however, even for those with university education, only 47.2% are at the top quartile of basic literacy. In the meantime, more than 20% of respondents with a university degree are in the bottom two quartiles of basic literacy. Thus, even respondents with high levels of education can exhibit a low level of financial literacy. Basic financial literacy is negatively correlated with age, which stands in stark contrast to the hump shaped profile of basic financial literacy across age found in the developed countries (Bucher-Koenen and Lusardi, 2011; Lusardi and Mitchell, 2011; Van Rooij et al., 2011a; Boisclair et al., 2017). This phenomenon might be attributed to the cohort effects: the rapid economic development over recent decades enables the Chinese government to raise the quantity and quality of public education, thereby leading to a substantial increase in the years of schooling across generations. With regard to differences across gender, men display higher basic financial knowledge than women, which is consistent with findings in the literature (Van Rooij et al., 2011b; Fonseca et al., 2012). Table 5 reports the distribution of the advanced financial literacy index. Like basic literacy, advanced literacy is strongly
Table 4 Basic financial literacy across demographics. Basic financial literacy quartiles
Education Primary school Junior high school Senior high school University Age ≤30 years 31–40 years 41–50 years 51–60 years ≥ 61 years Gender Female Male
1 (low)
2
3
4 (high)
Mean
N
49.52 25.33 17.18 6.31
28.18 30.35 25.10 16.56
13.95 25.44 31.77 29.97
8.34 18.89 25.96 47.16
3.07 4.44 4.94 5.73
731 916 809 634
10.07 12.28 22.24 30.97 36.53
24.83 26.32 25.29 27.78 23.83
32.89 30.02 26.30 21.67 21.63
32.21 31.38 26.18 19.58 18.01
5.35 5.22 4.68 4.16 3.87
298 513 787 720 772
28.10 21.32
27.51 23.38
22.94 28.14
21.45 27.15
4.30 4.76
1683 1407
Note: This table reports the distribution of basic financial literacy index across different levels of education, different age groups, and across gender. We group basic financial literacy index in four quartiles and report for each subgroup of education, age, and gender the proportion of households in each literacy quartile as well as the mean quartile number. Percentages may not sum up to 100 due to rounding. 7
Pacific-Basin Finance Journal 59 (2020) 101262
G. Niu, et al.
Table 5 Advanced financial literacy across demographics. Advanced financial literacy quartiles
Education Primary school Junior high school Senior high school University Age ≤30 years 31–40 years 41–50 years 51–60 years ≥ 61 years Gender Female Male
1 (low)
2
3
4 (high)
Mean
N
53.63 23.58 15.70 5.84
25.44 32.42 24.60 14.35
15.32 25.87 30.90 27.13
5.61 18.12 28.80 52.68
2.98 4.69 5.42 6.69
731 916 809 634
9.73 12.67 19.82 30.83 38.86
18.46 20.66 25.54 26.39 28.63
29.53 29.04 25.92 23.75 20.60
42.28 37.62 28.72 19.03 11.92
6.14 5.92 5.24 4.44 3.78
298 513 787 720 772
28.52 20.75
25.85 24.02
25.73 24.02
19.90 31.20
4.57 5.27
1683 1407
Note: This table reports the distribution of advanced financial literacy index across different levels of education, different age groups, and across gender. We group advanced financial literacy index in four quartiles and report for each subgroup of education, age, and gender the proportion of households in each literacy quartile as well as the mean quartile number. Percentages may not sum up to 100 due to rounding.
positively related to education. More than half of respondents with primary education are at the lowest level of advanced literacy, while a similar fraction of those with university education fall in the highest level of advanced literacy. Elderly respondents display a low degree of advanced literacy: only 11.9% of elderly respondents (61 or older) are in the top quartile of the literacy distribution. As age declines, advanced literacy significantly increases. Women are more financially illiterate than men. Around 28% of women are at the lowest level of literacy, while only 19.9% are in the fourth quartile; the corresponding figures for men are 20.8% and 31.2%, respectively. To complement the questions measuring actual financial literacy, the 2014 CFPS also asks respondents to assess their own financial knowledge, which allows us to check the validity of the two indices by comparing them with self-perceived knowledge. The latter is constructed using the question in the survey: How would you asset your overall financial knowledge relative to peers' average level? There are six options available: Much higher than peers' average level; Higher than peers' average level; At peers' average level; Lower than peers' average level; Much lower than peer's average level; Do not Know. Table 6 reports the relationship between these the two objective literacy indices and the self-assessed literacy. Apparently, there is a strong positive correlation between objective literacy and subjective literacy. More than 45% of respondents who assess their financial knowledge as being much lower than average level fall in the bottom quartile of basic literacy distribution, while the majority of those with high self-assessed levels of financial knowledge Table 6 Basic and advanced financial literacy versus self-assessed literacy. Basic financial literacy quartiles
Self-assessed literacy
Much lower than average level Lower than average level At average level Higher than average level Much higher than average level Do not know
Self-assessed literacy
1(very low) 2 3 4 5(very high) Do not know
1 (low)
2
3
4 (high)
Mean
N
48.48 26.82 16.92 9.82 12.50 77.50
24.79 30.00 23.33 16.56 8.33 15.00
15.98 24.02 29.43 29.61 32.33 7.50
10.74 19.15 30.31 45.01 46.83 0.00
3.18 4.35 5.00 5.57 5.48 1.54
363 1253 1247 163 24 40
Advanced financial literacy quartiles 1 (low)
2
3
4 (high)
Mean
N
48.76 26.82 16.44 10.43 8.33 87.50
24.24 29.29 22.85 16.56 4.17 12.50
16.25 25.62 27.99 22.70 20.83 0.00
10.74 18.28 32.72 50.31 66.67 0.00
3.38 4.60 5.51 6.36 6.96 1.05
363 1253 1247 163 24 40
Note: This table reports the distribution of basic and advanced financial literacy indices across different answer categories of the self-assessed literacy question. We group basic and advanced financial literacy indices in four quartiles and report the proportion of households in each literacy quartiles as well as the mean quartile number for each category of self-assessed literacy. Percentages may not sum up to 100 due to rounding. 8
Pacific-Basin Finance Journal 59 (2020) 101262
G. Niu, et al.
are located in the top two quartiles of the basic literacy index. This relationship is even stronger for the advanced literacy index. Thus, in general, respondents have a reasonable assessment of their own financial literacy. 3.2. Retirement preparation and other variables We employ three variables to measure households' levels of retirement preparation. The first measure is a dummy variable Retirement planning, which is constructed using the following question in the 2014 CFPS: Have you ever tried to figure out how much you need to save for retirement? The same question was also included in the 2004 Health and Retirement Study (HRS) and the National Financial Capacity Survey. Retirement planning is important, as the literature has shown that planners are inclined to save more for retirement than non-planners (Behrman et al., 2012; Van Rooij et al., 2012). Most extant studies in this field focus on developed countries, but little is known about the role of retirement planning in developing countries. As explained in the introductory section, the relatively low public pension benefits and the collapse of the traditional family support system are the main causes of the increasing need for retirement planning in China. In addition, an important element in retirement preparation is developing longterm financial goals to cope with financial needs over a long retirement period. To this end, we construct the second measure of retirement preparation, which is a dummy variable Financial planning indicating whether a household has a long-term financial plan or not. The precise wording of the question for Financial planning is given in Appendix C. The third measure of retirement preparation is a dummy variable Private pension participation, which indicates whether a household has private pension insurance offered by commercial insurance companies. As the major component of the third pillar, private pension insurance plays an increasingly important role in ensuring retirement well-being, especially considering the inadequate financial support from public pensions. Thus, purchasing a private pension can be seen to be an active way of preparing for retirement. To account for individual heterogeneity that might affect the relationship between financial literacy and retirement preparation, we include a wide range of household background characteristics as control variables, such as age, gender, marital status, family income and so on. We also include the level of financial wealth of a household as a proxy for private savings. To control for the potential impact of expected retirement income, we also include the types of pension plans that the respondent has. Definitions of all the variables can be found in Table 7. Table 8 reports the descriptive statistics of the variables outlined above. Note that following the literature (Alessie et al., 2011; Lusardi and Mitchell, 2011), we restrict our sample to non-retired respondents in the empirical analysis, because retirement planning is more relevant for them. Both the share of people with retirement planning and the share with financial planning are around onethird. Only 3% of respondents have private pension insurance. This result indicates that the majority of respondents do not adequately prepare for retirement. The average age of the sample is approximately 44. There is a great deal of heterogeneity in educational attainment: 24% of respondents have a university degree, while 47% do not obtain even a senior high school diploma. Men are slightly underrepresented (45%), and the majority of respondents (86%) are married. 20% of respondents are self-employed. 28% of respondents are covered by the Urban Enterprise Pension Scheme (UEPS), 7% by the Public Servants Pension Scheme (PSPS), and 9% by the Urban Residents Pension Scheme (URPS). Only 4% of respondents have the supplementary Enterprise Annuity Scheme (EAS). In addition, 9% of respondents are covered by the New Rural Pension Scheme (NRPS). Those respondents are mainly ruralurban migrants. In our sample, the proportion of residents in urban China covered by the public pension system is barely more than Table 7 Variable definitions. Variable
Definition
Retirement planning
A dummy variable that equals one if the respondent chooses the option “yes” as the answer to the question “Have you ever tried to figure out how much you need to save for retirement?”, and zero otherwise. A dummy variable that equals one if the respondent has a long-term financial plan, and zero otherwise. A dummy variable that equals one if the respondent has private pension insurance, and zero otherwise. The respondent's age. A dummy variable that equals one if the respondent's education level is Junior high school, and zero otherwise. A dummy variable that equals one if the respondent's education level is Junior high school, and zero otherwise. A dummy variable that equals one if the respondent's education level is University or above, and zero otherwise. A dummy variable that equals one if the respondent is male, and zero otherwise. A dummy variable that equals one if the respondent is married, and zero otherwise. A dummy variable that equals one if the respondent does not belong to Han ethnic group, and zero otherwise. A dummy variable that equals one equals one if the respondent is member of the Chinese Communist Party, and zero otherwise. The number of family members A dummy variable that equals one if the respondent is self-employed, and zero otherwise. A dummy variable that equals one if the respondent is member of the Public Servants Pension Scheme, and zero otherwise. A dummy variable that equals one if the respondent is member of the Urban Enterprise Pension Scheme, and zero otherwise. A dummy variable that equals one if the respondent is member of the Enterprise Annuity Scheme, and zero otherwise. A dummy variable that equals one if the respondent is member of the Urban Residents Pensions Scheme, and zero otherwise. A dummy variable that equals one if the respondent is member of the New Rural Pension Scheme, and zero otherwise. A dummy variable that equals one if the household owns a home, and zero otherwise. Total income of the household. Total value of financial assets held by the household. Financial wealth includes cash, deposits, money market accounts, bond, stocks and mutual funds.
Financial planning Private pension participation Age Junior high school Senior high school University Male Married Minority CCP membership Family size Self-employed PSPS UEPS EAS URPS NRPS Homeowner Family income Financial wealth
9
Pacific-Basin Finance Journal 59 (2020) 101262
G. Niu, et al.
Table 8 Descriptive statistics. Variable
Observation
Mean
Median
Std. Dev.
Retirement planning Financial planning Private pension participation Age Junior high school Senior high school University Male Married Minority CCP membership Family size Self-employed PSPS UEPS EAS URPS NRPS Homeowner Family income Financial wealth
2333 2278 2331 2333 2333 2333 2333 2333 2333 2333 2333 2333 2333 2331 2331 2331 2331 2331 2333 2333 2314
0.33 0.34 0.03 44.49 0.29 0.29 0.24 0.45 0.86 0.05 0.13 3.33 0.20 0.07 0.28 0.04 0.09 0.09 0.73 81,844.16 91,846.93
0 0 0 45 0 0 0 0 1 0 0 3 0 0 0 0 0 0 1 55,000 20,000
0.47 0.47 0.18 10.94 0.46 0.45 0.43 0.50 0.35 0.21 0.34 1.44 0.40 0.25 0.45 0.19 0.29 0.29 0.44 181,136.31 239,609.83
50%, a figure similar to that reported in Jiang et al. (2018), who use another recent household survey. The lack of pension coverage further highlights the need for active retirement preparation. 4. Multivariate analysis of retirement preparation 4.1. Main results We run the following multivariate regression model to explore the relationship between retirement preparation and financial literacy,
Yi = α + β’Xi + δ′Ci + εi ,
(1)
where Y are retirement preparation variables, X are financial literacy indices, and C is a vector of control variables. All variables are defined and discussed in the previous section. Note that province dummies are also included in C to control for province fixed effects. α is the intercept and ε is the error term. We are mainly interested in the vector of coefficients β, which captures the relationship between financial literacy and retirement preparation measures, but we will also discuss δ, which are the coefficients on the control variables. Given the binary nature of our dependent variables, we estimate the models using probit regressions. Table 9 presents the estimation results for retirement planning. As explained above, each model also includes a constant and province dummies, the coefficients of which are not reported for concision. Holding all controls at the mean, the marginal effects are calculated and reported instead of the estimated coefficients. In columns (1) and (2), we first consider advanced and basic financial literacy in separate models, and then, in column (3), we include them jointly. When separately considered, both sophisticated and basic literacy increase the probability of retirement planning, indicating that financially knowledgeable households are more likely to realize the importance of retirement preparation. By contrast, when introducing both literacy indices jointly, advanced financial literacy remains strongly correlated with thinking about retirement, while the basic literacy index does not significantly affect planning for retirement. Given that the standard deviation of advanced literacy is standardized to one in our sample, the estimated marginal effect of advanced financial literacy (0.088) suggests that a one standard deviation increase in advanced literacy around the mean population advanced literacy is associated with 8.8% increase in the probability of retirement planning. Hence, the effect of advanced literacy on retirement planning is both statistically and economically significant. Based on these simple estimates, we cannot yet give a causal interpretation of the relationship between financial literacy and planning since the literacy variable might itself be endogenous due to omitted variable bias, reverse causality, and measurement errors (Van Rooij et al., 2011a; Van Rooij et al., 2011b). First, some individual characteristics that can affect planning are hard to measure (e.g., ability). We address the impact of the omitted variables, such as ability, by including a basic literacy index in the regressions as a proxy for ability. Additionally, if households that think more about retirement invest more in financial education, then the link may go the other way around: from retirement preparation to improvement of financial knowledge. For this reason, a positive relationship between planning and financial literacy could be contaminated due to reverse causality. Furthermore, because of the difficulty in measuring financial literacy, it is likely that there are measurement errors in the advanced financial literacy index, which could yield biased estimates (attenuation bias). To resolve the endogeneity problem, we adopt an instrumental variable probit (IVprobit) strategy. To be specific, following existing literature on this topic (Van Rooij et al., 2011a), we use the dummy variable indicating whether there is any family member 10
Pacific-Basin Finance Journal 59 (2020) 101262
G. Niu, et al.
Table 9 Effects of financial literacy on retirement planning.
Advanced financial literacy
(1)
(2)
(3)
(4)
Probit
Probit
Probit
IVprobit
0.050⁎⁎⁎ (0.012) 0.002⁎⁎ (0.001) −0.015 (0.030) −0.029 (0.031) −0.057 (0.037) 0.033⁎ (0.020) 0.005 (0.030) 0.033 (0.048) 0.026 (0.031) −0.001 (0.007) 0.071⁎⁎⁎ (0.025) −0.011 (0.040) 0.010 (0.025) 0.006 (0.052) 0.049 (0.035) 0.008 (0.036) −0.001 (0.023) 0.003 (0.011) 0.001 (0.002) 2289 0.082
0.088⁎⁎⁎ (0.013) 0.009 (0.013) 0.003⁎⁎⁎ (0.001) −0.036 (0.030) −0.059⁎ (0.032) −0.102⁎⁎⁎ (0.037) 0.019 (0.020) 0.009 (0.030) 0.028 (0.048) 0.017 (0.031) −0.001 (0.007) 0.072⁎⁎⁎ (0.025) −0.009 (0.039) 0.002 (0.025) −0.011 (0.050) 0.036 (0.035) 0.012 (0.036) −0.004 (0.023) 0.001 (0.011) 0.000 (0.002) 2289 0.097
0.295⁎⁎⁎ (0.113) −0.098 (0.066) 0.004⁎⁎⁎ (0.001) −0.082⁎⁎ (0.032) −0.124⁎⁎⁎ (0.039) −0.199⁎⁎⁎ (0.053) −0.016 (0.029) 0.022 (0.026) 0.007 (0.045) −0.009 (0.032) 0.002 (0.006) 0.058⁎⁎ (0.027) −0.006 (0.035) −0.019 (0.026) −0.050 (0.046) 0.002 (0.039) 0.015 (0.031) −0.009 (0.020) −0.005 (0.011) −0.002 (0.002) 2289
0.092⁎⁎⁎ (0.012)
Basic financial literacy Age Junior high school Senior high school University Male Married Minority CCP membership Family size Self-employed PSPS UEPS EAS URPS NRPS Homeowner Log(Family income+1) Log(Financial wealth+1) Observations R-squared Economics degree
0.003⁎⁎⁎ (0.001) −0.035 (0.030) −0.057⁎ (0.032) −0.099⁎⁎⁎ (0.037) 0.019 (0.020) 0.010 (0.030) 0.027 (0.048) 0.017 (0.031) −0.001 (0.007) 0.071⁎⁎⁎ (0.025) −0.009 (0.039) 0.003 (0.025) −0.011 (0.050) 0.037 (0.035) 0.013 (0.036) −0.003 (0.023) 0.002 (0.011) 0.001 (0.002) 2289 0.097
0.107⁎⁎⁎ (0.039) 46.81⁎⁎⁎
First stage F-statistic
Note: The table reports the marginal effects of the probit and IVprobit models, with retirement planning as dependent variables. Advanced financial literacy index is instrumented with Economics degree, which is a dummy variable indicating whether there is any family member with an economics/ management degree and zero otherwise. Province dummies are also included, but their marginal effects are not reported to save space. Standard errors are reported in parentheses. ⁎⁎⁎, ⁎⁎ and ⁎ indicate 1%, 5% and 10% significance levels, respectively.
with an economics/management degree as an instrument for financial sophistication.4 On the one hand, economics/management training could improve financial literacy. As financial knowledge might be shared within households, we consider not only the respondent's educational background but also other family members' educational background. On the other hand, having such an education in school should not be directly related to retirement preparation since retirement is a distant concern at schooling age. Column (4) of Table 9 presents the IVprobit estimation results. Consistent with our expectation, the estimated coefficient on the instrument Economics degree is positive and highly significant, suggesting that it is strongly correlated with advanced financial literacy
4 The classification of academic disciplines in China's education system is based on a hierarchy structure. At the very top are several broad academic categories. Each of the academic categories includes a number of level-one disciplines and level-two disciplines. Economics and management are identified as two academic categories in China, while microeconomic, macroeconomics, finance, accounting, marketing and other business-related disciplines are under the two categories. The data we use only has information on the academic category to which an individual's degree belongs.
11
Pacific-Basin Finance Journal 59 (2020) 101262
G. Niu, et al.
Table 10 Effects of financial literacy on financial planning.
Advanced financial literacy
(1)
(2)
(3)
(4)
Probit
Probit
Probit
IVprobit
0.016 (0.013) −0.000 (0.001) −0.026 (0.032) −0.036 (0.033) −0.026 (0.040) 0.010 (0.021) −0.050 (0.031) 0.083⁎ (0.049) −0.010 (0.032) 0.013 (0.008) 0.050⁎ (0.027) −0.026 (0.043) −0.008 (0.026) 0.033 (0.055) −0.008 (0.036) 0.040 (0.039) 0.025 (0.024) 0.020 (0.012) 0.003 (0.002) 2257 0.025
0.058⁎⁎⁎ (0.014) −0.011 (0.014) 0.000 (0.001) −0.040 (0.032) −0.056⁎ (0.033) −0.057 (0.040) 0.002 (0.021) −0.046 (0.031) 0.079 (0.050) −0.016 (0.032) 0.013⁎ (0.008) 0.050⁎ (0.026) −0.025 (0.043) −0.015 (0.026) 0.021 (0.054) −0.016 (0.036) 0.040 (0.039) 0.023 (0.024) 0.018 (0.012) 0.003 (0.002) 2257 0.031
0.161 (0.117) −0.058 (0.121) 0.001 (0.002) −0.065 (0.065) −0.090 (0.086) −0.111 (0.137) −0.012 (0.042) −0.036 (0.042) 0.068 (0.060) −0.025 (0.039) 0.014⁎ (0.008) 0.049⁎ (0.027) −0.024 (0.042) −0.025 (0.036) −0.001 (0.076) −0.028 (0.045) 0.040 (0.037) 0.019 (0.026) 0.015 (0.016) 0.002 (0.004) 2257
0.054⁎⁎⁎ (0.012)
Basic financial literacy Age Junior high school Senior high school University Male Married Minority CCP membership Family size Self-employed PSPS UEPS EAS URPS NRPS Homeowner Log(Family income+1) Log(Financial wealth+1) Observations R-squared Economics degree
0.000 (0.001) −0.042 (0.032) −0.059⁎ (0.033) −0.060 (0.040) 0.002 (0.021) −0.046 (0.031) 0.079 (0.050) −0.015 (0.032) 0.013⁎ (0.008) 0.051⁎ (0.026) −0.026 (0.043) −0.016 (0.026) 0.021 (0.054) −0.017 (0.036) 0.040 (0.039) 0.023 (0.024) 0.018 (0.012) 0.003 (0.002) 2257 0.030
0.090⁎⁎ (0.039) 41.67⁎⁎⁎
First stage F-statistic
Note: The table reports the marginal effects of the probit and IVprobit models, with financial planning as dependent variables. Advanced financial literacy index is instrumented with Economics degree, which is a dummy variable indicating whether there is any family member with an economics/ management degree and zero otherwise. Province dummies are also included, but their marginal effects are not reported to save space. Standard errors are reported in parentheses. ⁎⁎⁎, ⁎⁎ and ⁎ indicate 1%, 5% and 10% significance levels, respectively.
index. Moreover, the F-statistics of the first stage regression is 46.81, which is well above the weak instrument critical value (Stock & Yogo, 2005). Thus, we conclude that the presence of a weak instrument is not an issue. More importantly, the marginal effect of financial sophistication remains positive and statistically significant and increases by a factor of 3.4 relative to the benchmark probit model, which confirms the finding of our benchmark model that financial sophistication boosts preparation for retirement. Inspecting the control variables yields some interesting observations. Older respondents are more likely to plan for retirement, perhaps because they are closer to retirement and thus find it more important to think about retirement life. After controlling for financial literacy and other household individual characteristics, education plays either an insignificant or a negative role in explaining retirement planning in China, which is similar to the findings in the Netherlands and the United States (Lusardi and Mitchell, 2011; Van Rooij et al., 2011a). A possible explanation in the Chinese context is that individuals with higher educational attainment have a better chance to be hired by state-owned enterprises, foreign companies and governments, which generally offer higher pension benefits and therefore crowd out individuals' own efforts to prepare for retirement. By the same token, the self-employed have a higher propensity to plan for retirement, since they are largely uncovered by the social safety net. Table 10 presents the results of regressions with financial planning. While both advanced and basic literacy are positively 12
Pacific-Basin Finance Journal 59 (2020) 101262
G. Niu, et al.
Table 11 Effects of financial literacy on private pension participation.
Advanced financial literacy
(1)
(2)
(3)
(4)
Probit
Probit
Probit
IVprobit
0.015⁎⁎ (0.006) −0.001⁎⁎⁎ (0.000) 0.001 (0.012) 0.023⁎ (0.013) 0.014 (0.015) 0.003 (0.008) 0.029⁎⁎ (0.015) 0.006 (0.023) −0.017 (0.014) −0.003 (0.004) 0.017⁎⁎ (0.008) 0.001 (0.015) 0.012 (0.009) 0.018 (0.022) 0.009 (0.014) 0.020 (0.019) 0.009 (0.009) 0.016⁎⁎ (0.006) 0.001 (0.001) 2100 0.128
0.005⁎⁎ (0.003) 0.012 (0.017) −0.001⁎⁎⁎ (0.000) 0.000 (0.012) 0.023⁎ (0.013) 0.011 (0.015) 0.002 (0.008) 0.030⁎⁎ (0.015) 0.006 (0.023) −0.017 (0.014) −0.003 (0.004) 0.017⁎⁎ (0.008) 0.001 (0.015) 0.011 (0.009) 0.017 (0.022) 0.007 (0.014) 0.020 (0.019) 0.008 (0.009) 0.016⁎⁎ (0.006) 0.001 (0.001) 2100 0.129
0.240⁎⁎ (0.121) −0.088 (0.080) −0.005⁎⁎⁎ (0.002) −0.052 (0.049) 0.068⁎ (0.037) 0.096 (0.089) 0.029 (0.028) 0.061⁎⁎ (0.028) 0.005 (0.047) −0.052⁎ (0.029) −0.001 (0.007) 0.033⁎⁎ (0.016) 0.001 (0.033) 0.001 (0.026) 0.020 (0.050) 0.017 (0.035) 0.044 (0.034) 0.005 (0.020) 0.020⁎⁎ (0.009) 0.001 (0.003) 2100
0.009⁎⁎⁎ (0.003)
Basic financial literacy Age Junior high school Senior high school University Male Married Minority CCP membership Family size Self-employed PSPS UEPS EAS URPS NRPS Homeowner Log(Family income+1) Log(Financial wealth+1) Observations R-squared Economics degree
−0.001⁎⁎⁎ (0.000) 0.001 (0.011) 0.023⁎ (0.013) 0.014 (0.015) 0.003 (0.008) 0.031⁎⁎ (0.015) 0.006 (0.023) −0.018 (0.014) −0.003 (0.004) 0.016⁎⁎ (0.008) 0.002 (0.015) 0.013 (0.009) 0.017 (0.022) 0.008 (0.014) 0.019 (0.019) 0.008 (0.009) 0.016⁎⁎ (0.006) 0.001 (0.001) 2100 0.125
0.106⁎⁎⁎ (0.040) 49.90⁎⁎⁎
First stage F-statistic
Note: The table reports the marginal effects of the probit and IVprobit models, with private pension participation as dependent variables. Advanced financial literacy index is instrumented with Economics degree, which is a dummy variable indicating whether there is any family member with an economics/management degree and zero otherwise. Province dummies are also included, but their marginal effects are not reported to save space. Standard errors are reported in parentheses. ⁎⁎⁎, ⁎⁎ and ⁎ indicate 1%, 5% and 10% significance levels, respectively.
correlated with the individual's propensity to have a long-term financial plan in separate models, only advanced literacy remains positive and significant in the joint model. This result indicates that financial sophistication is an important determinant of long-term financial planning, which might be attributed to a better understanding of the workings of financial market. The regression estimate of financial literacy under the IV estimation, though less significant, remains positive. Overall, we find some evidence that financial sophistication positively affects long-term financial planning. Turning to control variables, educational attainment is negatively related to financial planning, which might be explained by the fact that more educated people are more likely to have generous pension schemes, such as supplementary enterprise annuities, and earn stable income, and therefore have a lower incentive to plan for the future. For the same reason, the self-employed are more likely to have long-term financial plans. Table 11 reports the estimation results for private pension participation. The number of observations is slightly reduced due to missing responses to the private pension participation question. When considered separately in columns (1) and (2), both advanced and basic literacy have significantly positive effects on private pension participation. By contrast, basic literacy loses significance when it is considered together with sophisticated literacy in column (3). The IVprobit estimation result in column (4) confirms that 13
Pacific-Basin Finance Journal 59 (2020) 101262
G. Niu, et al.
Table 12 Subsample analysis: self-employed versus others.
Advanced financial literacy Controls Observations R-squared
(1)
(2)
(3)
(4)
(5)
(6)
Retirement planning
Retirement planning
Financial planning
Financial planning
Private pension participation
Private pension participation
Self-employed
Others
Self-employed
Others
Self-employed
Others
⁎⁎⁎
0.114 (0.024) Yes 455 0.162
⁎⁎⁎
0.085 (0.013) Yes 1857 0.102
⁎⁎⁎
⁎⁎⁎
0.074 (0.027) Yes 437 0.084
0.045 (0.014) Yes 1820 0.033
⁎⁎
0.044 (0.018) Yes 455 0.324
0.006 (0.005) Yes 1857 0.179
Note: The table reports the marginal effects of the probit models. Dependent variable is retirement planning in columns (1) and (2), financial planning in columns (3) and (4), and private pension participation in columns (5) and (6). Self-employed represents the subsample for the selfemployed respondents. Others represents the subsample for the respondents who are not self-employed. All regressions have the same controls as in column (1) of Table 8, but their marginal effects are not reported to save space. Standard errors are reported in parentheses. ⁎⁎⁎, ⁎⁎ and ⁎ indicate 1%, 5% and 10% significance levels, respectively.
financial sophistication raises the likelihood of purchasing a private pension. Overall, these results indicate that advanced financial literacy can induce individuals to actively prepare for retirement by participating in private pension plans, which could be explained by the fact that because private pension insurance is a complex financial product, its functioning and values for securing financial wellbeing in retirement can hardly be understood by people without sufficient financial knowledge. 4.2. Subsample analysis Our baseline results show that the estimate of self-employed is positive and significant in all the regressions, implying that the self-employed are more likely to prepare for retirement. This can be attributed to the lack of social security covering this group of people. Given the higher inclination for retirement preparation, it is tempting to investigate whether the effects of financial literacy differ between the self-employed and others. For this purpose, we split our sample into the self-employed versus others, and rerun the regressions separately using these two subsamples. Table 12 reports the results of subsample analysis. Given that advanced literacy is found to be more important for retirement preparation than basic literacy for retirement preparation in the above analysis, we only consider advanced literacy for simplicity. The estimate of advanced financial literacy is larger for the self-employed group than for the other group in all regressions. The coefficient on advanced literacy even becomes insignificant for the other group in the private pension participation regression. Overall, these results indicate that the effects of financial literacy on retirement preparation are more pronounced for the selfemployed people than for the other group of people. The results are consistent with the fact that retirement planning is more important for the self-employed people as they are not well covered by the public pension system. 4.3. Including additional controls In this section, we check the robustness of our results by including more control variables. Personal preferences, such as risk aversion, impatience, and overconfidence might drive the correlation between financial literacy and retirement planning, thereby biasing our estimates (Van Rooij et al., 2011a). First, risk aversion may induce individuals to invest in financial education and, at the same time, to prepare for retirement. Second, people who do not care much about the future might have a lower incentive to invest in becoming financially literate. Third, overconfident individuals are less likely to actively acquire financial knowledge and to prepare for the future. To exclude these possibilities, we construct an ordinal variable Risk aversion, a dummy variable Impatience, and a dummy variable Overconfidence and include them as additional controls in our regression model. The exact wording of the questions used to construct Risk aversion, Impatience, and Overconfidence, and the construction methods are provided in Appendix C. Table 13 presents the estimation results of regressions that include measures for risk aversion, impatience, and overconfidence. The IVprobit estimation results show that risk aversion has a significantly negative effect on retirement planning, while overconfidence has a positive effect on retirement planning. Impatience significantly increases private pension participation. More importantly, the inclusion of these variables does not alter the sign and significance of the advanced literacy index, which demonstrates that the positive effects of financial sophistication on retirement planning remain robust. 5. Conclusion In this paper, we examine financial literacy among urban residents in China. Based on thirteen questions that capture different dimensions of financial knowledge, we distinguish between basic financial literacy and sophisticated financial literacy. We find that the level of financial literacy is rather low in China. Barely more than half of respondents in urban China can provide correct answers to simple questions about topics such as interest rate and inflation. Moreover, only approximately one-third of respondents can 14
Pacific-Basin Finance Journal 59 (2020) 101262
G. Niu, et al.
Table 13 Including risk aversion, impatience and overconfidence.
Advanced financial literacy Risk aversion Impatience Overconfidence Controls Observations R-squared Economics degree
(1)
(2)
(3)
(4)
(5)
(6)
Retirement planning
Retirement planning
Financial planning
Financial planning
Private pension participation
Private pension participation
Probit
IVprobit
Probit
IVprobit
Probit
IVprobit
⁎⁎⁎
0.095 (0.012) −0.002 (0.011) −0.051⁎⁎ (0.021) 0.091 (0.093) Yes 2258 0.097
First stage F-statistic
⁎⁎⁎
0.302 (0.092) −0.053⁎ (0.028) −0.027 (0.026) 0.188⁎⁎ (0.089) Yes 2258
⁎⁎⁎
0.049 (0.013) 0.024⁎⁎ (0.012) 0.011 (0.022) 0.166⁎ (0.097) Yes 2242 0.032
0.096⁎⁎ (0.042) 27.69⁎⁎⁎
0.177 (0.282) −0.007 (0.075) 0.016 (0.023) 0.226 (0.142) Yes 2242 0.075⁎ (0.042) 26.48⁎⁎⁎
⁎⁎
0.007 (0.003) 0.007 (0.005) 0.019⁎⁎ (0.009) 0.016 (0.054) Yes 2059 0.132
0.239⁎⁎ (0.112) −0.041 (0.040) 0.046⁎⁎ (0.018) 0.256 (0.274) Yes 2059 0.100⁎⁎ (0.044) 29.95⁎⁎⁎
Note: The table reports the marginal effects of the probit and IVprobit models. Dependent variable is retirement planning in columns (1) and (2), financial planning in columns (3) and (4), and private pension participation in columns (5) and (6). Advanced financial literacy index is instrumented with Economics degree, which is a dummy variable indicating whether there is any family member with an economics/management degree and zero otherwise. All regressions have the same controls as in column (1) of Table 8, but their marginal effects are not reported to save space. Standard errors are reported in parentheses. ⁎⁎⁎, ⁎⁎ and ⁎ indicate 1%, 5% and 10% significance levels, respectively.
understand more sophisticated financial concepts, such as the function of stock markets. In line with findings in the existing literature, women, less-educated people and the elderly are more likely to have little financial knowledge. More targeted efforts and programs are therefore needed to improve financial literacy for those vulnerable groups. Based on multivariate analysis controlling for a large number of confounding factors, we document that financial literacy is strongly associated with retirement planning, financial planning, and private pension participation in China. In particular, sophisticated literacy remains an important driver of retirement preparation under different model specifications. Moreover, using information on the education history of household members, we perform instrumental variable analysis and provide evidence on the causal impact of financial sophistication. Our findings in the context of urban China echo the conclusions for developed countries, such as the United States, the Netherlands, Germany, and Canada. In light of the fragile family support systems and the inadequate social safety nets in contemporary China, it is more urgent than ever for policy makers to take measures that can improve financial literacy so as to promote people's awareness of retirement preparation. Acknowledgements This work was supported by the National Natural Science Foundation of China [grant number 71703114, 71904160], the Fundamental Research Funds for the Central Universities in China [grant number JBK170148], and the 111 Project [grant number B16040]. Appendix A. Basic literacy questions (Q1) Interest rate level. What is the current 1 year deposit rate? (1) Below 1%; (2) Between 1% and 5%; (3) Between 5% and 10%; (4) Above 10%; (5) Do not know. (Q2) Numeracy. Suppose you had ¥10,000 in a savings account and the interest rate was 3% per year and you never withdraw money or interest payments. After 1 year, how much would you have in this account in total? (1) Exactly ¥10,300 (2) More than ¥10,300; (3) Less than ¥10,300; (4) Do not know. (Q3) Interest compounding. Suppose you had ¥10,000 in a savings account and the interest rate was 3% per year. After 2 years, how much do you think you would have in the account if you left the money to grow? (1) Exactly ¥10,600 (2) More than ¥10,600; (3) Less than ¥10,600; (4) Do not know. (Q4) Inflation. Imagine that the interest rate for your savings account was 3% per year and inflation was 5% per year. After 1 year, how much would you be able to buy with the money in this account? (1) Exactly the same; (2) More than today; (3) Less than today; (4) Do not know. (Q5) Time value of money. Suppose San Zhang receives a bequest of ¥100,000 today, while Si Li will receive a bequest of ¥100,000 in 3 years. Which one's bequest is worth more? (1) San Zhang's bequest; (2) Si Li’s bequest; (3) Worth the same; (4) Do not know. 15
Pacific-Basin Finance Journal 59 (2020) 101262
G. Niu, et al.
Appendix B. Advanced literacy questions (Q1) In general, investment opportunities with high returns have high risks. True or False? (1) True; (2) False; (3) Do not know. (Q2) Buying a company stock usually provides a safer return than a stock mutual fund. True or False? (1) True; (2) False; (3) Do not know. (Q3) Which bank formulates and implements monetary policy? (1) Bank of China; (2) Industrial and Commercial Bank of China; (3) The People's Bank of China; (4) China Construction Bank; (5) Do not know. (Q4) Normally, which asset is riskiest? (1) Savings account; (2) Bonds; (3) Stocks; (4) Mutual funds; (5) Do not know. (Q5) What happens if somebody buys the stock of firm B? (1) He lends money to firm B; (2) He owns a part of firm B; (3) If he holds the stock for a short period of time, he lends money to firm B; if he holds the stock for a long period of time, he owns a part of firm B; (4) None of the above; (5) Do not know. (Q6) Which statement about mutual funds is correct? (1) Mutual funds with low prices will have better performance in the future; (2) Mutual funds can invest in several assets, for example invest in stocks and bonds; (3) Normally, mutual funds pay a guaranteed rate of return depending on their past performance; (4) None of the above; (5) Do not know. (Q7) Which statement about banks' wealth management products is correct? (1) Buying banks' wealth management products may cause losses; (2) Buying banks' wealth management products never causes losses; (3) The expected returns of banks' wealth management products are realized returns; (4) None of the above; (5) Do not know. (Q8) Which statement describes the main function of the stock market? (1) The stock market helps to predict stock earnings; (2) The stock market results in an increase in the price of stocks; (3) The stock market brings people who want to buy stocks together with those who want to sell stocks; (4) None of the above; (5) Do not know. Appendix C. Wording of the questions and construction of the variables used in empirical analysis (Q1) Retirement planning. Have you ever tried to figure out how much you need to save for retirement? (1) Yes; (2) No. We use the question to construct a dummy variable Retirement planning (1 = yes; 0 = no). (Q2) Financial planning. Do you agree with the following statement? I have a long-term financial plan. (1) Completely disagree; (2) Disagree; (3) Neither agree nor disagree; (4) Agree; (5) Completely agree. We use the question to construct a dummy variable Financial planning (1 = agree, or completely agree; 0 = neither agree nor disagree, disagree, or completely disagree). (Q3) Risk aversion. If you are going to make an investment, which projects would you like to choose? (1) High risk, high return projects; (2) Average risk, average return projects; (3) Low risk, low return projects; (4) Not willing to take on any risk. We use the question to construct a categorical variable Risk aversion (1 = high risk, high return projects; 2 = average risk, average return projects; 3 = low risk, low return projects; 4 = not willing to take on any risk). (Q4) Impatience. Do you agree with the following statement? I prefer instant gratification over delayed gratification. (1) Completely disagree; (2) Disagree; (3) Neither agree nor disagree; (4) Agree; (5) Completely agree. We use the question to construct a dummy variable Impatience (1 = agree, or completely agree; 0 = neither agree nor disagree, disagree, or completely disagree). (Q5) Self-assessed financial literacy. How would you assess your overall financial knowledge relative to peers' average level? (1) Much higher than peers' average level; (2) Higher than peers' average level; (3) At peers' average level; (4) Lower than peers' average level; (5) Much lower than peer's average level; (6) Do not know. We follow Kramer (2016) and use the question to construct a dummy variable Overconfidence, which equals one if the respondent believes that his/her overall financial knowledge is higher or much higher than peers' average level, but his/her measured advanced financial literacy is below median measured advanced financial literacy. References Agarwal, S., Driscoll, J.C., Gabaix, X., Laibson, D., 2009. The age of reason: financial decisions over the life-cycle with implications for regulation. Brook. Pap. Econ. Act. 40 (2), 51–117. Agarwal, S., Amromin, G., Bendavid, I., Chomsisengphet, S., Evanoff, D.D., 2015. Financial literacy and financial planning: evidence from India. J. Hous. Econ. 27, 4–21. Alessie, R., Van Rooij, M., Lusardi, A., 2011. Financial literacy and retirement preparation in the Netherlands. J. Pension Econ. Finan. 10 (4), 527–545. Bateman, H., Liu, K., 2014. Pension Reform in China: Racing against the Demographic Clock. UNSW Australian School of Business Research Paper No. 2013ACTL2022. Behrman, J.R., Mitchell, O.S., Soo, C.K., Bravo, D., 2012. How financial literacy affects household wealth accumulation. Am. Econ. Rev. 102 (3), 300–304. Boisclair, D., Lusardi, A., Michaud, P., 2017. Financial literacy and retirement planning in Canada. J. Pension Econ. Finan. 16 (3), 277–296. Bucher-Koenen, T., Lusardi, A., 2011. Financial literacy and retirement planning in Germany. J. Pension Econ. Finan. 10 (4), 565–584. Cai, Y., Cheng, Y., 2014. Pension reform in China: challenges and opportunities. J. Econ. Surv. 28 (4), 636–651. Calvet, L.E., Campbell, J.Y., Sodini, P., 2007. Down or out: assessing the welfare costs of household investment mistakes. J. Polit. Econ. 115 (5), 707–747. Chamon, M., Prasad, E.S., 2010. Why are saving rates of urban households in China rising. Am. Econ. J. Macroecon. 2 (1), 93–130. Chen, L., 2016. From fintech to finlife: the case of fintech development in China. China Econ. J. 9 (3), 225–239. China Insurance Regulatory Commission, 2015. Annual Report for the Chinese Insurance Market 2015. Retrieved from. http://bxjg.circ.gov.cn/web/site0/tab5257/ module14498/page3.htm. Croll, E.J., 1999. Social welfare reform: trends and tensions. China Q. 159 (159), 684–699. Dong, K., Wang, G., 2016. China’s pension system: achievements, challenges and future developments. Econ. Polit. Stud. 4 (4), 414–433. Fang, H., Feng, J., 2018. The Chinese pension system. In: NBER Working Paper No. 25088. Fonseca, R., Mullen, K.J., Zamarro, G., Zissimopoulos, J.M., 2012. What explains the gender gap in financial literacy? The role of household decision-making. J.
16
Pacific-Basin Finance Journal 59 (2020) 101262
G. Niu, et al.
Consum. Aff. 46 (1), 90–106. Glaeser, E.L., Huang, W., Ma, Y., Shleifer, A., 2017. A real estate boom with Chinese characteristics. J. Econ. Perspect. 31 (1), 93–116. Jappelli, T., Padula, M., 2013. Investment in financial literacy and saving decisions. J. Bank. Financ. 37 (8), 2779–2792. Jiang, J., Qian, J., Vincent, W.Z., 2018. Social protection for the informal sector in urban China: institutional constraints and self-selection behaviour. J. Soc. Pol. 47 (2), 335–357. Kramer, M.M., 2016. Financial literacy, confidence and financial advice seeking. J. Econ. Behav. Organ. 131, 198–217. Klapper, L., Panos, G.A., 2011. Financial literacy and retirement planning: the Russian case. J. Pension Econ. Finan. 10 (4), 599–618. Li, B., 2014. Social pension unification in an urbanising China: paths and constraints. Public Adm. Dev. 34 (4), 281–293. Liang, P., Guo, S., 2015. Social interaction, internet access and stock market participation—an empirical study in China. J. Comp. Econ. 43 (4), 883–901. Liao, L., Xiao, J.J., Zhang, W., Zhou, C., 2017a. Financial literacy and risky asset holdings: evidence from China. Account. Finance 57 (5), 1383–1415. Liao, L., Zhang, X., Zhang, Y., 2017b. Mutual fund managers’ timing abilities. Pac. Basin Financ. J. 44, 80–96. Lin, C., Hsiao, Y., Yeh, C.Y., 2017. Financial literacy, financial advisors, and information sources on demand for life insurance. Pac. Basin Financ. J. 43, 218–237. Liu, T., Sun, L., 2016. Pension reform in China. J. Aging Soc. Pol. 28 (1), 15–28. Lusardi, A., Mitchell, O.S., 2011. Financial literacy and retirement planning in the United States. J. Pension Econ. Finan. 10 (4), 509–525. McLoughlin, K., Meredith, J., 2017. The Rise of Chinese Money Market Funds. RBA Bulletin, pp. 75–84 March. National Bureau of Statistics, 2012. Tabulation on the 2010 Population Census of the People's Republic of China. Chinese Statistics Press, Beijing. National Bureau of Statistics, 2015. Monitoring Report on Rural Migrant Workers 2014. Retrieved from. http://www.stats.gov.cn/tjsj/zxfb/201504/t20150429_ 797821.html. Niu, G., Zhou, Y., 2018. Financial literacy and retirement planning: evidence from China. Appl. Econ. Lett. 25 (9), 619–623. Phillips, D.R., 2000. Ageing in the Asia-Pacific Region: Issues, Policies, and Future Trends. Routledge, London and New York. Phillips, D.R., Feng, Z., 2015. Challenges for the aging family in the People’s Republic of China. Can. J. Aging-revue Can. Vieilliss. 34 (3), 290–304. Qin, M., Zhuang, Y., Liu, H., 2015. Old age insurance participation among rural-urban migrants in China. Demogr. Res. 33 (37), 1047–1066. Sekita, S., 2011. Financial literacy and retirement planning in Japan. J. Pension Econ. Finan. 10 (4), 637–656. Stock, J., Yogo, M., 2005. Testing for Weak Instruments in Linear IV Regression. In: Andrews, D.W.K. (Ed.), Identification and Inference for Econometric Models. Cambridge University Press, New York, pp. 80–108. United Nations, 2017. World Population Prospects: The 2017 Revision. Department of Economic and Social Affairs, Population Division. Van Rooij, M., Lusardi, A., Alessie, R., 2011a. Financial literacy and retirement planning in the Netherlands. J. Econ. Psychol. 32 (4), 593–608. Van Rooij, M., Lusardi, A., Alessie, R., 2011b. Financial literacy and stock market participation. J. Financ. Econ. 101 (2), 449–472. Van Rooij, M., Lusardi, A., Alessie, R., 2012. Financial literacy, retirement planning and household wealth. Econ. J. 122 (560), 449–478. Xie, Y., Hu, J., Zhang, C., 2014. The China family panel studies (CFPS): design and practice. Chin. J. Sociol. Shehui 34, 1–33 (in Chinese). Zhan, H.J., Feng, X., Luo, B., 2008. Placing elderly parents in institutions in urban China: a reinterpretation of filial piety. Res. Aging Int. Bimonthly J. 15 (5), 1695–1712. Zhang, L., Brooks, R., Ding, D., Ding, H., Lu, J., He, H., Mano, R., 2018. China’s high savings: drivers, prospects,and policies. IMF Work. Pap. 18 (277). Zhao, Q., Mi, H., 2019. Evaluation on the sustainability of urban public pension system in China. Sustainability 11 (5), 1–20. Zhu, J., Guo, K., Ai, M., Zhao, Y., Bai, X., 2018. The further opening up of China’s financial sector. China Econ. J. 11 (1), 44–52.
17