University of Groningen MSc. International Business &
Management
Uppsala University MSc. Economics & Business
Macroeconomic and Institutional Determinants of Stock Market
Development in South Asia and the Asia-Pacific Region:
Implications for Advanced and Emerging Economies
March 2011
Abstract
This thesis intents to expose macroeconomic and institutional indicators of stock market development, by making use of the indicators and methodology as described in existing literature. This thesis contributes to the literature by focusing solely on the fast-developing region of South Asia and Asia-Pacific, and by making a distinction between advanced and emerging economies within the sample. The data shows that macroeconomic indicators are more accurate predictors of stock market development than institutional indicators. Real income is the single most accurate estimator of changes in levels of stock market development. Furthermore, I show that the banking sector and stock markets are complements in emerging economies, whereas this relationship does not hold for advanced economies. The GDP deflator is identified as an additional proxy for macroeconomic instability, and is found to be a significant indicator. Finally, measures of institutional quality are no considerable predictors, as they have a negative impact on stock market development in several cases.
Keywords
Table of Contents
1. INTRODUCTION 5
2. STOCK MARKET EVOLUTION IN THE ASIA-PACIFIC REGION 9
3. LITERATURE REVIEW AND HYPOTHESES DEVELOPMENT 13
3.1.THE FINANCIAL SYSTEM AND ECONOMIC GROWTH 13
3.2.DEVELOPMENT OF THE FINANCIAL SYSTEM 15
3.3.INDICATORS OF SMD 16
3.3.1.STOCK MARKET SIZE 17
3.3.2.INCOME LEVEL 18
3.3.3.SAVINGS AND INVESTMENT 19
3.3.4.BANKING SECTOR DEVELOPMENT 19
3.3.5.STOCK MARKET LIQUIDITY 20
3.3.6.MACROECONOMIC INSTABILITY 21
3.3.7.PRIVATE CAPITAL FLOWS FROM ABROAD 22
3.3.8.INSTITUTIONAL QUALITY 22
4. DATA AND METHODOLOGY 25
4.1.METHODOLOGY 25 4.2.SAMPLE 27 4.3.DATA SOURCES 27 4.4.MISSING DATA 27 4.5.NORMALITY 28 4.6.EXTREME OUTLIERS 28
4.7.RANDOM AND FIXED EFFECTS 28
4.8.CAUSALITY 29
4.9.MULTICOLLINEARITY 30
4.10.SIMPLE BIVARIATE REGRESSION 30
5. PANEL REGRESSION RESULTS AND DISCUSSION 31
5.1.MACROECONOMIC DETERMINANTS OF SMD 31
5.2.INSTITUTIONAL DETERMINANTS OF SMD 34
5.3.IMPLICATIONS FOR ADVANCED AND EMERGING ASIA 37
6. SENSITIVITY ANALYSIS: AN ABSOLUTE MEASURE OF STOCK MARKET SIZE 41
7. CONCLUSION 43
7.1CONCLUSION 43
1. Introduction
A well-functioning financial system is a vital ingredient for the economic development of countries (Levine and Zervos, 1996). This view has recently been confirmed by the outbreak of the financial crisis, as the causes of the crisis were to a certain extent rooted in imperfections of the financial system. While there were indications that the financial crisis would threaten global welfare in a way not observed since the Great Depression, resolute global actions may have averted this from happening. Yet, the current economic and financial situation in most parts of the world is far from stable. Moreover, considering estimates that emerging economies will be the primary engine of global economic growth in the years ahead, and the relative underdevelopment of financial systems in those economies, it is doubtful whether global financial systems will be more stable in the near future.1
For this reason, better understanding and awareness of financial development, and the factors that drive financial development, is desired. Above all, this thesis focuses on one aspect of financial development, namely stock market development (hereafter referred to as SMD). Multiple scholars (Levine and Zervos, 1996; Demirgüc-Kunt and Levine, 1996; Levine, 1999) have emphasized the positive effects that stock markets can have on long-run economic growth of countries, through the financial functions stock markets perform. These functions include among others the promotion of corporate control and governance, an efficient mobilization of savings, the acquisition and distribution of information, the promotion of specialization, and the diversification and pooling of risk (World Economic Forum, 2010).
In the existing literature, SMD is normally linked to the size of stock markets, as larger markets are generally less concentrated and more transparent (Yartey, 2008). Virtually all of those papers measure SMD by means of the market capitalization ratio, which is the total market capitalization of an economy relative to its nominal GDP. Papers that try to explain levels of SMD have found different indicators to be of importance. Those differences can partly be explained by considering the region and time period involved. For example, while Naceur et al. (2007) find a significant negative relationship between inflation changes and SMD for the Middle Eastern and North African region, Billmeier and Massa (2005) do not find evidence for this relationship for the Middle Eastern and Central Asian region.
It is currently an interesting period to study topics related to financial development, as the recent past has seen the growing significance of emerging economies on the world stage.
1
Especially the massive economic growth of emerging economies in Asia is remarkable. While total GDP of advanced economies was four times higher in 1995 than the value of their less developed counterparts, the gap has shrunk to almost zero thirteen years later (see Figure 1). 0 1000 2000 3000 4000 5000 6000 7000 8000 1995 1997 1999 2001 2003 2005 2007 Emerging Advanced
Figure 1: Development of total combined GDP (in billions of current USD) of the six largest advanced and the six largest emerging economies in Asia-Pacific for the period 1995-2008 (data retrieved from 2010 World Bank sources).
While stock markets in Africa and the Middle East have been directly linked to SMD (see for example Billmeier and Massa, 2007; Naceur et al., 2007; and Yartey, 2008), stock markets in South Asia and the Asia-Pacific region have so far not been the center of attention.2 An exception being Garcia and Liu (1999), who try to explain different levels of market capitalization observed in East Asia and Latin America. They find that stock markets in East Asia are more developed than in Latin America due to sustained economic growth, a higher savings rate, more liquidity, and a more developed banking sector.
The shifting of global economic gravity from the West towards the Far East is not the only reason why financial development is currently an appealing topic. As mentioned before, the recent financial crisis clearly exposed the flaws of the current financial system. The crisis led to a rise of claims asking for drastic financial reforms, in order to prevent a severe crisis from occurring again. This thesis intents to combine the two points of interests, by focusing on determinants of SMD in the Asia-Pacific region. Related studies have demonstrated that
2
macroeconomic and institutional factors are most likely to affect levels of SMD (see Billmeier and Massa, 2007; Yartey, 2008). These factors are subsequently used in this thesis to explain changes in the relative market capitalization of economies. Consequently, the research question of this thesis is stated as follows:
Which macroeconomic and institutional variables, as described in current literature, are related to levels of stock market development in the advanced and emerging economies of South Asia and the Asia-Pacific region?
This thesis aims to contribute to the existing literature by using a sample of advanced and emerging economies in the Asia-Pacific region, a region that has hardly been linked to the topic of SMD thus far. Moreover, it provides an up-to-date perspective on the topic of SMD, as several of the pioneering papers on the topic of SMD are currently outdated. This is the main reason why particularly the two major Chinese stock markets, which currently rank among the largest stock markets in the world, are commonly excluded.3 Finally, most of the indicators that are used in this thesis have been identified in existing literature. However, I argue that the GDP deflator is a more accurate proxy for macroeconomic instability than the macroeconomic proxies used in current literature, as the deflator quicker reflects up-to-date expenditure patterns caused by changing price levels. Therefore, the GDP deflator is added to the list of possible macroeconomic indicators of SMD.
A better understanding of the factors that drive SMD in Asia-Pacific will not only be beneficial for scholars and practitioners focusing on the topic of financial development and SMD, but also for politicians and officials. A better understanding of the linkages between components of the financial system, and the factors that influence their development, will allow them to create policies that will have a positive effect on an economy’s level of SMD, which in turn can lead to sustainable economic growth. This is particularly relevant at the present time, as reforms in the financial system are desired.
Furthermore, Demirgüc-Kunt and Levine (1996) characterize developed stock markets as being generally larger in size, more liquid, less volatile, less concentrated, better able to price assets efficiently, and more integrated with world capital markets. Hence, the relationships revealed in this thesis can be linked to stock market characteristics as described by Demirgüc-Kunt and Levine. This information can be useful for investors, which can be either institutions
3
(insurance companies, pension funds, corporations, etc.) or private investors, as it can help them to create more diversified and less risky portfolios in this fast-developing region of the world.
2. Stock market evolution in the Asia-Pacific region
Care should be taken when investigating the evolution of stock markets in Asia-Pacific, as large deviations exist between the level of economic and financial development of the countries located in this region. While several advanced economies in Asia-Pacific rank among the most financially developed countries in the world, some developing economies in this region have experienced little progress in their financial sectors and still have underdeveloped financial systems. Trading on those stock markets occurs in only a few stocks, which together make up a large part of the total market capitalization. Other stocks on those markets suffer from serious disclosure and information deficiencies. Table 1 illustrates the deviations in levels of SMD between advanced (e.g. Hong Kong, Singapore, and Japan) and developing economies (e.g. Fiji, Nepal, and Mongolia).
Australia 65.0 1924 97.9 103.1 Bangladesh 8.4 290 11.6 137.3 China 61.6 1604 120.7 121.3 Fiji 15.9 16 0.1 3.2 Hong Kong 617.0 1017 755.1 81.8 India 53.2 4921 86.5 85.2 Indonesia 19.3 396 21.7 71.3 Japan 65.9 3299 120.3 153.2 Korea, Rep 53.1 1798 157.4 181.2 Malaysia 84.6 977 38.5 33.2 Mongolia 7.7 420 1.0 9.3 Nepal 38.8 149 2.9 7.5 New Zealand 20.9 149 27.4 46.1 Pakistan 14.2 653 32.9 116.0 Philippines 31.1 244 10.3 22.2 Singapore 93.1 455 140.1 101.3 Sri Lanka 10.6 234 2.5 17.2 Thailand 37.7 476 42.9 78.2 Number of listed firms Market capitalization ratio Value traded ratio Turnover ratio
Due to those substantial deviations, a distinction is made between advanced (developed) and emerging (developing) economies in the region.4 This distinction is based on the classification by the International Monetary Fund (2010). They classify China, India, Indonesia, Malaysia, Pakistan, Philippines, Sri Lanka, Thailand, Bangladesh, Fiji, Mongolia, and Nepal as emerging and developing economies, whereas Australia, Hong Kong, Japan, New Zealand, Singapore, and South Korea are listed as advanced economies.
Figure 2 shows that both advanced and emerging economies have experienced market capitalization growth in the relevant period. This increase in market capitalization coincided with a growth in the number of listed firms in the region from less than 13,900 firms in 1995 to over 19,000 firms in 2008 (World Bank, 2010). Furthermore, the figure shows that the relative share of market capitalization of emerging economies remained small until 2000, but started to take off substantially after that. Total market capitalization of emerging economies even overtook market capitalization of the region’s advanced economies in 2007.
An upward trend becomes less visible when total market capitalization is corrected for the size of the economy, as Asian economies experienced enormous GDP growth in the last two decades (see Figure 3). A missing upward trend is partly caused by the financial crisis, as an upward trend is visible until 2007.5 Figure 3 additionally illustrates that advanced economies have on average a higher market capitalization ratio than emerging economies, which is in line with the generally accepted view that advanced economies have relatively larger and more developed stock markets.
4
Strictly speaking, emerging and developing countries are not synonyms. While developing countries are generally described as nations with low levels of material well-being, emerging countries are characterized as developing nations that experience rapid growth and industrialization, and are therefore considered to be in a transitional phase between developing and developed status. However, I do use the two terms as synonyms in this thesis, as they are both ambiguous, and no single definition of the terms is recognized internationally. Additionally, the International Monetary Fund does also not make an explicit distinction between emerging and developing economies.
0 2000 4000 6000 8000 10000 12000 14000 16000 18000 20000 1995 1997 1999 2001 2003 2005 2007 Emerging Developed Whole sample
Figure 2: Total market capitalization (in billions of current USD) of advanced and emerging economies in Asia-Pacific for the period 1995-2008 (data retrieved from 2010 World Bank sources).
0% 50% 100% 150% 200% 250% 1995 1997 1999 2001 2003 2005 2007 Emerging Developed Whole sample
Figure 3: Average market capitalization ratio of advanced and emerging economies in Asia-Pacific for the period 1995-2008 (data retrieved from 2010 World Bank sources).
costs, can be observed during the second half of the nineties. This finding can be linked to the IT and Internet boom that took place during this period, and ended with the dot-com bubble in 2000. Stagnating turnover ratios after this year provide further evidence for this linkage. Nevertheless, the average turnover ratio of the whole sample has roughly doubled during the full period. Furthermore, the figure makes clear that the turnover ratio of developed economies was not necessarily higher than the ratio of emerging economies until 2004.
0% 50% 100% 150% 200% 250% 1995 1997 1999 2001 2003 2005 2007 Emerging Developed Whole sample
Figure 4: Average value traded ratio of advanced and emerging economies in Asia-Pacific for the period 1995-2008 (data retrieved from 2010 World Bank sources).
0% 20% 40% 60% 80% 100% 120% 140% 1995 1997 1999 2001 2003 2005 2007 Emerging Developed Whole sample
3. Literature review and hypotheses development
The literature overview first provides a framework that explains the relationship between financial development and economic growth, before focusing on measures of financial development. Next, indicators of SMD are identified and the hypotheses are presented.
3.1. The financial system and economic growth
The financial system facilitates the crucial economic function of transferring money between savers and borrowers. The system consists out of a set of complex and closely interrelated financial institutions, markets, services, instruments, practices, and transactions. Financial markets and intermediaries form the central element of the system. While financial markets most obviously include stock markets, bond markets, foreign exchange markets, and derivatives markets, financial intermediaries can be divided into banking and nonbanking financial services (World Economic Forum, 2010).6
In the past disagreement existed about whether financial systems can positively contribute to economic growth.7 Nowadays ample empirical evidence confirms the useful functions that financial systems can serve economies. Table 2 provides a hierarchical illustration of the way in which financial markets and intermediaries contribute to economic growth. Economic theory suggests that the main reason why financial markets and intermediaries develop is because of the existence of market frictions (information costs and transaction costs). Financial markets and intermediaries deal with those market frictions by mobilizing savings, allocating resources, exerting corporate control, facilitating the trading, hedging, diversification, and pooling of risk, and facilitating the exchange of goods and services. These financial functions promote the rate of technological innovation and capital accumulation, which in turn can have a positive effect on an economy’s productivity growth and economic growth (World Economic Forum, 2010).
6
Nonbanking financial corporations include among others finance companies, mutual funds, and brokerage houses.
7
Table 2: How financial markets contributes to economic growth (Levine, 1997).
Additionally, and more specifically, a strong empirical association is also assumed between stock markets and long run economic growth (Levine and Zervos, 1996). Not only do stock markets improve the efficiency of the financial system by providing information and the practice of company valuation, they also provide investors with an exit mechanism (Naceur et al., 2008). Levine (1999) identifies three mechanisms through which stock markets can enhance economic growth. These mechanisms are in line with the functions that financial systems in general perform (as described in Table 2).
The first mechanism emphasizes the effect that stock markets can have on resource allocation and corporate control. Arbitrageurs have more incentives to use their resources for research purposes in case stock markets grow larger and become more liquid, as in this case it will be easier to realize a profit from research. A better supply of information regarding firms should improve resource allocation. Improved resource allocation can also be obtained by means of greater corporate control. Assuming that developed stock markets facilitate takeovers, poorly operating firms will be purchased by outsiders, and management will be replaced in order to increase profitability. Likewise, as stock markets make it easier to tie managerial compensation to stock price performance, interests of managers and owners will align.
The second argument involves easier diversification of risk. In well-developed stock markets it is easier for investors to construct portfolios with a minimum of middlemen. Also, liquid markets promote long-term investments, as it is easier for investors to sell equities cheaply and quickly in those markets. In this way, firms can benefit from a permanent access to capital. Hence, long-term investments help to improve the allocation of capital and thereby boost productivity growth.
effective accounting standards, and the development of contracting systems that decrease obstructions to resource mobilization.
3.2. Development of the financial system
Financial development can best be defined as the factors, policies, and institutions that lead to effective financial intermediation and markets, as well as deep and broad access to capital and financial services (World Economic Forum, 2010). There are several ways to measure an economy’s level of financial development. Huang (2005) constructs a composite index of financial development that takes into account overall financial development, SMD, financial intermediary development, financial efficiency development, and financial size development (or financial depth). The World Economic Forum (2010) takes another approach. They identify seven measures of financial development, which in turn can be grouped into three broad categories (see Table 3).
Table 3: Elements of the financial system and pillars of financial development. The table is based on the report of the World Economic Forum (2010).
development are assumed to be positively related with each other in an intuitively plausible manner. As a result, an economy’s level of SMD can have a direct impact on the developmental stage of other financial markets, as well as on other pillars of financial development. The next section will take a closer look at the factors that influence the development levels of stock markets.
3.3. Indicators of SMD
Demirgüc-Kunt and Levine (1996) were among the first scholars to focus on the topic of SMD. They argue that SMD is a multi-dimensional concept, and they therefore use six estimators of SMD to create an aggregate development index of 44 national economies. These indicators include stock market size, stock market liquidity, stock market concentration, stock market volatility, stock market integration, and institutional development. They furthermore observe that the indicators show high correlations with each other. In line with this finding, follow-up studies consider stock market size as the optimal proxy for a stock market’s level of development, as this variable is less arbitrary than the other measures.8 Hence, stock market size, as measured by the market capitalization ratio, is normally used in related studies as the dependent variable in order to identify determinants of SMD.
Based on this approach, scholars have identified the saving rate, banking sector development, income levels, liquidity, macroeconomic instability, and private capital flows as the most important macroeconomic determinants of SMD. Table A1 in the appendix presents a comprehensive overview of the macroeconomic indicators that are used in related studies, and the indicators that are found to have a statistically significant effect on levels of SMD. Besides macroeconomic factors, Law and Habibullah (2009), Yartey (2008), and Billmeier and Massa (2007) stress the significant impact that institutional factors can have on levels of SMD. Cherif and Gazdar (2010) also investigate the relationship between institutional quality and SMD, but they do find this relationship to be significant for the Middle-Eastern and North African region.
Besides macroeconomic and institutional factors, several other factors have also been hypothesized to influence SMD directly, or indirectly through their impact on other components of the financial system. Among these are cultural factors, which include Hofstede’s (1980) power distance and masculinity dimensions (De Jong and Semenov, 2002), and language and religion (Stulz and Williamson, 2003). Geographical factors have not been directly linked to SMD, but it has been shown that they exhibit predictive power when it comes to explaining economic
8
development (see for example Sachs, 2003). Geographical factors can include an economy’s latitude and natural barriers (Huang, 2005), resource endowment levels (Billmeier and Massa, 2007), and settler mortality (Beck et al., 2003).
Table 4 gives on overview of the main factors. This thesis will focus on macroeconomic and institutional variables, as related studies on the topic of SMD normally focus on these factors. The most-widely used macroeconomic and institutional indicators of SMD are discussed in the next section. The hypotheses are presented subsequently.
Macroeconomic factors Institutional factors Other factors
Income levels Corruption Stock market concentration Domestic savings Political rights Stock market integration Domestic investment Public sector efficiency Stock market volatility Macroeconomic stability Regulatory burden Cultural factors Banking sector development Legal code origin Geographical factors Private capital flows Property rights
Stock market liquidity Accounting standards
Table 4: Variables hypothesized to affect levels of SMD.
3.3.1. Stock market size
According to Demirgüc-Kunt and Levine (1996), stock market size can be measured by either the market capitalization ratio (a relative measure of size) or by the number of listed firms (an absolute measure of size). Large variations can be observed in the size of stock markets across economies, which makes this measure useful for identifying different stages of SMD.9 As the absolute measure of stock market size does not make a correction for the size of an economy, it is highly correlated with an economy’s population and absolute income level. As a consequence, most papers on the topic of SMD perceive the market capitalization ratio as the most accurate estimator of SMD, and use this variable to explain other potential indicators of SMD.
The assumption behind the market capitalization ratio is that stock market size is positively related with the ability to mobilize capital and diversify risk (Demirgüc-Kunt and Levine, 1996). In addition, it can be argued that large stock markets are being more observed than smaller ones, while at the same time the stakes are higher. This makes it likely that those markets are more efficient and more developed compared to their smaller counterparts.
9
The market capitalization ratio is defined as the total market value of all listed shares within an economy divided by its GDP. The ratio serves as the dependent variable in this thesis. Hence, the market capitalization ratio can be interpreted as a synonym for SMD. Additionally, the number of listed firms variable is used as the dependent variable in the sensitivity analysis in chapter 6 to see whether outcomes are statistically different for this absolute measure of stock market size.
3.3.2. Income level
Income has been found to have a positive and significant impact on SMD.10 Yartey (2008) argues that the expansion of an economy will create new demand for financial services, which in turn exerts pressure to establish larger and more sophisticated financial institutions. In addition, as income level often goes hand in hand with better education, better defined property rights, and a better general environment for business, an economy’s income level is positively related to its institutional level of development (Garcia and Liu, 1999). Finally, income levels correlate strongly with the development of the macroeconomic environment, as economic downturns can have a considerable effect on the market capitalization levels of stock markets. Hence, income levels can also be an approximation of the macroeconomic situation of an economy (Beck and Levine, 2003).
Income level is measured by an economy’s real GDP and real GDP per capita variables.11 While the former takes into account changes in the absolute size of an economy, the latter measures the scaled size of an economy, and can therefore be interpreted as a proxy for welfare levels.12
H1a: An economy’s level of income, as measured by the absolute size of the economy, is positively related to the economy’s level of SMD in South Asia and the Asia-Pacific region.
H1b: An economy’s level of income, as measured by the scaled size of the economy, is positively related to the economy’s level of SMD in South Asia and the Asia-Pacific region.
10
Cherif and Gazdar (2010), Billmeier and Massa (2007), and Garcia and Liu (1999) provide evidence for this positive relationship.
11
Income (in)equality is additionally likely to affect levels of SMD. Unfortunately, World Bank data concerning this issue contains multiple missing observations for the sample used in this thesis. Hence, I am unable to include this factor.
12
3.3.3. Savings and investment
Like banks, stock markets facilitate the flow of savings to investment projects. Generally, the larger the saving rate, the higher the flow of capital to stock markets. As the investment rate depends largely on the saving rate, a positive relationship is expected between saving rate, investment rate, and SMD. Huang (2005) investigates the relationship between financial development and private investment, and finds positive causal effects going in both directions. While Yartey (2008) confirms this finding in his research covering 42 emerging markets, Naceur et al. (2007), Billmeier and Massa (2007), and Cherif and Gazdar (2010) do not find this relationship to be significant for economies located in the North Africa, Middle East and Central Asia. Instead, Cherif and Gazdar (2010), Naceur et al. (2007), and Garcia and Liu (1999) conclude that the domestic saving rate is a significant indicator of SMD. Yartey’s findings again contradict the others, as he does not find this relationship to be significant.
H2a: An economy’s saving rate is positively related to its level of SMD in South Asia and the Asia-Pacific region.
H2b: An economy’s investment rate is positively related to its level of SMD in South Asia and the Asia-Pacific region.
3.3.4. Banking sector development
As banks and stock markets both have the ability to supply firms with funds, it can be argued that stock markets and banks are substitutes. Garcia and Liu (1999) reason that banks and stock markets can be substitutes in the short run, due to arbitrage between stock market returns and interest rates. Yartey’s (2008) argues that the banking sector is a complement to stock markets during early stages of its development. However, once they develop, stock markets and banks tend to compete with each other as a vehicle for financing investment.
Nevertheless, banks and stock markets are generally assumed to be complements of each other. Abundant empirical evidence confirms this positive linkage. Demirgüc-Kunt and Levine (1996) find that financial intermediary development is highly correlated with SMD.13 Countries with well-developed stock markets tend to have a well-developed banking sector. Chinn and Ito
13
(2006) argue that the development in the banking sector is a precondition for SMD. This view is further supported by findings of Cherif and Gazdar (2010), Naceur et al. (2007), Huang (2005), and Garcia and Liu (1999). Hence, in accordance with the other linkages within the financial system, this thesis hypothesizes that stock markets and the banking sector are complements, and that they are assumed to grow simultaneously. The level of banking sector development is most often measured by the ratio of credit to the private sector to GDP.14 The variable measures the activity of the banking system in channeling savings to investors, which is essentially the banking system’s main function.
H3: An economy’s level of banking sector development is positively related to its level of SMD in South Asia and the Asia-Pacific region.
3.3.5. Stock market liquidity
Liquidity is generally defined as the ease and speed at which agents can buy and sell securities (Garcia and Liu, 1999). Most high-return projects require a long-term commitment of capital, while at the same time investors bear higher default and liquidity risks. Investors are normally only willing to take on those risks if they can alter their portfolios quickly and cheaply, as is the case in liquid stock markets. Thus, the amount of savings that is channeled through to stock markets is higher in liquid stock markets. This sequentially improves the allocation of capital and enhances prospects for long-term growth (Yartey, 2008). Therefore, stock market liquidity is considered a positive contributor to SMD and economic growth.
The value traded ratio and the turnover ratio are normally used to measure stock market liquidity.15 The turnover ratio estimates the activeness of a stock market by dividing the value of equity transactions by the total equity value of listed firms, and is therefore used as an indicator of low transaction costs. The value traded ratio takes into account the value of total shares traded relatively to the size of a country’s economy, and therefore reflects liquidity on an economy-wide basis. Due to this dissimilarity, the two ratios complement each other, as well as complement the
14
The ratio of broad money supply M3 to GDP is also often used as an indicator of banking sector development. I am forced to exclude this variable however, as multiple observations are missing.
15
capitalization ratio (Demirgüc-Kunt and Levine, 1996). Hence, both ratios are used in this thesis. The corresponding hypotheses are expressed as follows.
H4a: An economy’s level of stock market liquidity, as measured by the value traded ratio, is positively related to its level of SMD in South Asia and the Asia-Pacific region.
H4b: An economy’s level of stock market liquidity, as measured by the turnover ratio, is positively related to its level of SMD in South Asia and the Asia-Pacific region.
3.3.6. Macroeconomic instability
Garcia and Liu (1998) predict a negative relationship between stock market capitalization and macroeconomic instability, as firms and investors have fewer incentives to participate in the stock market when the underlying economic situation is volatile. The inflation rate is normally used as an indicator of macroeconomic stability.16 Boyd et al. (2001) investigate the relationship between financial development and the inflation rate. They conclude that countries experiencing higher inflation rates are likely to have smaller, less efficient and less active stock markets and financial intermediaries. Nevertheless, evidence demonstrating a negative relationship between macroeconomic instability and SMD remains scarce.17 Besides the inflation rate, the real interest rate is also regularly used as a proxy for macroeconomic instability (Yartey, 2008; Cherif and Gazdar, 2010). Additionally, this thesis proposes the use of the GDP deflator as a measure of macroeconomic instability. The GDP deflator is also a measure of inflation. However, whereas the inflation rate is based on the CPI, which essentially uses a fixed basked of goods (a Laspeyres index), and therefore explicitly assumes that consumers never substitute, the GDP deflator uses a flexible basket of goods (a Paasche index), and therefore assumes an infinite amount of substitution. Due to this noteworthy distinction, the GDP deflator is better able to reflect up-to-date expenditure patterns caused by changing price levels than the inflation rate.
H5a: An economy’s level of macroeconomic instability, as measured by the inflation rate, is negatively related to its level of SMD in South Asia and the Asia-Pacific region.
16
This can be either by considering the normal inflation rate, by changes in the inflation rate and by the standard deviation of the inflation rate.
17
H5b: An economy’s level of macroeconomic instability, as measured by the GDP deflator, is negatively related to its level of SMD in South Asia and the Asia-Pacific region.
H5c: An economy’s level of macroeconomic instability, as measured by the real interest rate, is negatively related to its level of SMD in South Asia and the Asia-Pacific region.
3.3.7. Private capital flows from abroad
As capital markets have become more globally integrated in the last decades, foreign investors have emerged as major contributors to the development of emerging stock markets. Foreign investment is normally associated with regulatory and institutional development, adequate disclosure and listing requirements and fair trading practices. In turn, an increase in operational and informational efficiency is likely to lead to greater confidence in the market (Yartey, 2008). Rajan and Zingales (1996) additionally stress that in the absence of trade and financial openness, incumbent firms have strong incentives to block the development of a more competitive and transparent financial sector. Hence, several papers support the view that policies, which encourage openness to external trade, have the ability to boost financial development (e.g. Do and Levchenko, 2004). In line with those arguments, Yartey (2008) and Huang (2005) find positive and statistically significant relationships between the FDI to GDP ratio and financial development.
H6: An economy’s level of private capital flows from abroad is positively related to its level of SMD in South Asia and the Asia-Pacific region.
3.3.8. Institutional quality
Several scholars have indeed identified a significant relationship between SMD and institutional quality. Levine (1999) finds that the legal system and the regulatory environment both have an important impact on financial development. Huang (2005) identifies a system involving contract enforcement, good accounting practices, and property rights as essential for an economy’s financial development. Yartey (2008) argues that low political risk demonstrates the existence of good institutional quality. He identifies three measures of political risk: the quality of governance (including political rights, public sector efficiency, corruption, and regulatory burdens), law and order (including legal code origin, property rights, and accounting standards), and accountability and limits placed on executive and political leaders. Creane et al. (2004) and Billmeier and Massa (2007) estimate institutional quality by means of the Index of Economic Freedom (Heritage Foundation, 2010), and validate the significance of good institutions on levels of financial development in the Middle East, North Africa, and Central Asia. Even though economic freedom is not a synonym for institutional quality, several aspects of the index can directly be related to institutions. The index takes into account ten factors, each with equal weights. This thesis uses the overall institutional index and seven components of the index, each corresponding to one of the following hypotheses.18
H7: An economy’s level of overall institutional quality is positively related to its level of SMD in South Asia and the Asia-Pacific region.
H7a: An economy’s level of institutional quality, as measured by the property rights index, is positively related to its level of SMD in South Asia and the Asia-Pacific region.
H7b: An economy’s level of institutional quality, as measured by the freedom from corruption index, is positively related to its level of SMD in South Asia and the Asia-Pacific region.
H7c: An economy’s level of institutional quality, as measured by the business freedom index, is positively related to its level of SMD in South Asia and the Asia-Pacific region.
H7d: An economy’s level of institutional quality, as measured by the investment freedom index, is positively related to its level of SMD in South Asia and the Asia-Pacific region.
18
H7e: An economy’s level of institutional quality, as measured by the trade freedom index, is positively related to its level of SMD in South Asia and the Asia-Pacific region.
H7f: An economy’s level of institutional quality, as measured by the monetary freedom index, is positively related to its level of SMD in South Asia and the Asia-Pacific region.
H7g: An economy’s level of institutional quality, as measured by the financial freedom index, is positively related to its level of SMD in South Asia and the Asia-Pacific region.
4. Data and methodology
The aim of this thesis is to explain deviations in the levels of SMD observed for economies in Asia-Pacific by considering macroeconomic and institutional variables. Due to the multi-dimensional nature of the data required, a panel dataset is created. The panel dataset is unbalanced, due to several missing observations. The market capitalization ratio will be used as the dependent variable, while the explanatory variables consist out of the most widely used macroeconomic and institutional indicators of SMD. These indicators are selected based on data availability and on findings in existing literature, with the only exception being the GDP deflator. Linear regressions (OLS) will be run in Eviews in order to model the relationship between relative stock market size and the appropriate indicators of SMD. Table 5 gives an overview of the variables used in this thesis. Hypotheses 1 to 6 reflect the macroeconomic indicators and hypotheses 7 and 7a to 7g are related to measures of institutional quality.
4.1. Methodology
The model that this thesis intends to estimate is based on the methodological approach suggested by Cherif and Gazdar (2010), who in turn derived their model from Calderon-Rossell (1990). it it it it MACRO INSTI MCR =
α
+β
+θ
+µ
for i = 1,2,…N, t = 1,2,…Ti (1)where MCR is the dependent variable, defined as the ratio of total market capitalization to GDP. The MACRO variable is made up of income level, saving rate, investment rate, banking sector development, stock market liquidity, macroeconomic stability, and private capital flows. The INSTI variable is an indicator of institutional quality,
α
is the intercept term, andµ
it is the disturbance term, which can be divided into:it t i
it
µ
λ
υ
µ
= + + (2)H Indicator Measured by Lagged Abbr. A priori sign exp.
Source
DV Stock market size
Log of ratio of stock market cap. to GDP
- MCR + World Bank
1a Income level Log of real GDP in USD Yes GDP + World Bank 1b Income level Log of per capita real GDP
in USD
Yes GDPCAP + World Bank 2a Saving rate Ratio of gross savings to
GDP
Yes SAV + World Bank
2b Investment rate Ratio of gross fixed capital formulation to GDP
Yes INV + World Bank
3 Banking sector development
Ratio of domestic credit to private sector to GDP
No DCPS + World Bank
4a Stock market liquidity
Ratio of total value of shares traded to GDP
Yes VTR + World Bank
4b Stock market liquidity
Ratio of total value of shares traded to market capitalization
Yes TOR + World Bank
5a Macroeconomic instability
Inflation rate (CPI) No INFL - World Bank 5b Macroeconomic
instability
GDP deflator (ratio of nominal to real GDP)
No DEFL - World Bank
5c Macroeconomic instability
Real interest rate No RIR - World Bank
6 Private capital flows
Ratio of FDI net inflows to GDP No PCF + World Bank 7 Institutional quality Overall index of institutional quality No INSTI + Heritage foundation 7a Institutional quality
Freedom from corruption index
No FFCI + Heritage
foundation 7b Institutional
quality
Property rights index No PRI + Heritage foundation 7c Institutional
quality
Business freedom index No BFI + Heritage foundation 7d Institutional
quality
Trade freedom index No TFI + Heritage
foundation 7e Institutional
quality
Investment freedom index No IFI + Heritage foundation 7f Institutional
quality
Monetary freedom index No MFI + Heritage foundation 7g Institutional
quality
Financial freedom index No FFI + Heritage foundation
4.2. Sample
Of all economies located in Asia-Pacific, eighteen are found to be suitable for the purpose of this research (see Table A2 in the appendix for an overview of the economies). The sample represents all economies in the Asia-Pacific region that had established a stock exchange at the beginning of the sample period.19 As explained in chapter 2, a distinction will be made between advanced and emerging countries, as substantial differences can exist between the economic and financial development of those groups. Six of the economies included in the sample are classified as advanced, while the remaining twelve are classified as emerging.
The sample period consists out of the years 1995 to 2008. While 2008 reflects the most recent year for which all data observations are available, the year 1995 is chosen as a starting point as institutional data is only available from this year and onwards. Additionally, the starting year coincides roughly with the opening of the Shenzhen and Shanghai stock exchanges in the first half of the nineties. Yearly observations of eighteen economies yield 252 observations in total.
4.3. Data sources
Data concerning the dependent and the macroeconomic variables is derived from World Bank sources (World Bank, 2010). The institutional indicators in this thesis are acquired through the 2010 Index of Economic Freedom, which is published by the Heritage Foundation (Heritage Foundation, 2010).
4.4. Missing data
Missing data reduces the total sample of 252 potential observations into 241 observations.20 Missing values for the turnover ratio and real interest rate could not be removed from the whole sample as this would mean that observations for Pakistan and Mongolia were removed from the sample practically altogether. Hence, regressions containing one of those variables consist out of a smaller sample than the standard 241 observations, and do therefore not fully represent Pakistan and Mongolia.
19
In this way, the Vietnamese and Cambodian economies are excluded, as they did not have established a stock exchange at the beginning of the sample period, while a smaller economy as Fiji is included. Taiwan is the only exception to this rule, as the World Bank does not publish data for this economy.
20
4.5. Normality
In order to test whether the dependent variable is normally distributed, the Jarque–Bera test is executed.21 This is a goodness-of-fit measure of departure from normality, and is based on skewness and kurtosis.22 The distribution of the variable observed in Figure A1 in the appendix is not normal. The variable has a high Jarque-Bera value, with a low probability value, which provides clear evidence that the alternative hypothesis of a non-normal distribution is accepted.23 Figure A2 in the appendix takes the natural logarithm of the variable. In this way the variable has a skewness of -0.25, a kurtosis of 2.53, and a Jarque-Bera of 5.01. With a probability of 0.08, the variable is normally distributed at the 5% level. Hence, the natural logarithm of the market capitalization ratio is used as the dependent variable in this thesis, which is in line with the approach taken by Billmeier and Massa (2007).
4.6. Extreme outliers
Due to the limited number of observations, extreme outliers of both the dependent variable and the independent variables can not simply be deleted. Hence, these values are adjusted by means of the Winsorization method, which implies that extreme figures are scaled back so that they do not deviate more than three times the standard deviation from the mean in the initial sample (Ruan et al., 2005).24 In this way, outliers are prevented from heavily influencing the distribution of statistics.
4.7. Random and fixed effects
Given the cross-sectional and time series characteristics of the data, the average values of the variables and the relationships between them are likely to vary over time and across all of the cross-sectional units in the sample (Brooks, 2008). Indeed, the data used in this thesis shows significant differences across countries (due to political, social, and economic factors), as well as across time (e.g. due to the severe economic downturn in the years 1997 and 2008).
21
The selection of the independent variables is based on (i) their use in related papers, and (ii) data availability. These variables are not tested for normality.
22
Samples from a normal distribution have an expected skewness of 0 and an expected excess kurtosis of 0, which is equal to a kurtosis of 3.
23
The null hypothesis assumes that the variable is normally distributed.
24
Hence, a random or fixed effects model is desired. In this way intercepts in the regression model can vary cross-sectionally and over time, as statistical outcomes are corrected for country-specific and time-country-specific heterogeneity.25 In general, fixed effects models are preferred when the entities in the sample effectively represent the entire population, whereas random effects models are chosen when entities in the sample are randomly selected from the population.
Due to the unbalanced nature of the dataset, it is not possible to apply two-way random effects or mixed fixed and random effects. Therefore, the only options are to use one-way random or fixed effects or two-way fixed effects. I first test whether a pooled regression will suffice by applying a two-way redundant fixed effects test. This test evaluates the joint significance of the fixed effects, where a zero p-value indicates that the fixed effects are significant and that they should not be restricted to zero. Next, I apply a one-way Hausman test to investigate whether random effects are uncorrelated with the explanatory variables.26 When the p-value of the test is zero or close to zero, a random effects model is not appropriate, and a fixed effect model is preferred.
4.8. Causality
It is important to point out that the relationship between financial sector development and economic growth is assumed to be bi-directional. Hence, financial deepening promotes economic growth, while economic growth stimulates financial development (Huang, 2005; Naceur et al., 2008). Table A3 in the appendix illustrates the high interdependence between GDP growth and market capitalization growth in the economies of Asia-Pacific. As the focus of this thesis is on identifying relationships between variables, and not as such on the causality of relationships, I correct for the causality problem by using last years measures for the value traded, turnover ratio, income level, saving and investment rate variables. This method is proposed by Cherif and Gazdar (2010) and Garcia and Liu (1999).
25
This results in the same statistical outcome as a simple pooled regression with country and time dummies.
26
4.9. Multicollinearity
To test for multicollinearity, the correlations between the dependent and independent variables are estimated (see Table A4 and A5 in the appendix). Several of the indicators indeed show high correlations with each other.27 Care should be taken when using those variables for regression analyses, as the use of highly correlated variables in the same regression might bias results. I correct for the multicollinearity problem by not using highly correlated variables in the same regression. For instance, as the GDP per capita variable shows high correlations with several of the institutional quality factors, the real GDP variable is used as our default measure of income. It should be noted additionally that related papers also experience similar levels of correlations between variables, which implies that high correlations are inevitably related to research on the topic of SMD (e.g. Garcia and Liu, 1999). Nevertheless, we test for multicollinearity by implementing the variance inflation factor (VIF) in each model. I use a cut-off value of five (Stundenmund, 2001) indicating that a figure larger than five is a sign that multicollinearity is a severe problem.
4.10. Simple bivariate regression
The correlation matrix can additionally be interpreted as a simple linear regression estimation. Simple bivariate regression is a regression method where only one single independent variable is used to explain changes in the dependent variable. The slope of a fitted line in such a regression is equal to the correlation between the two variables involved corrected by the ratio of standard deviations. As the correlation matrix in Table A4 includes the dependent variable, it can be observed whether the market capitalization ratio correlates positively or negatively to each independent variable. All relationships are assumed to be positive in theory, except for the macroeconomic instability predictors. The signs of the correlations described in Table A4 all confirm the hypothesized relationships of the independent variables to the market capitalization ratio. The next chapter will take a closer look at those relationships and their significant levels.
5. Panel regression results and discussion
Table 6 provides a summary of the variables used in this thesis. The full sample consist out of 241 observations, but the sample size for the turnover ratio and real interest rate variables are smaller, due to several missing values for those variables (as explained in paragraph 4.4). Measures of central tendency and dispersion are included.
Variable Abbr. N Mean Median SD Min. Max.
Market capitalization ratio MCR 241 69.8 39.8 81.6 1.4 349.8
Income (GDP) GDP 241 25.3 25.4 2.0 20.7 29.3
Income (GDP per capita) GDPCAP 241 7.8 7.4 1.7 5.3 10.6
Saving rate SAV 241 26.8 25.0 11.3 -5.7 51.8
Investment rate INV 241 25.3 23.6 6.3 13.9 43.6
Domestic credit DCPS 241 78.7 72.7 53.3 5.8 231.1
Value traded VTR 241 45.4 26.0 53.9 0.0 223.2
Turnover ratio TOR 222 74.1 50.9 75.0 0.3 326.7
Inflation rate INFL 241 4.9 3.9 4.7 -4.0 24.4
GDP deflator DEFL 241 5.2 4.2 5.5 -6.2 28.7
Real interest rate RIR 230 5.5 5.0 4.9 -24.6 23.8
Private capital flows PCF 241 3.5 1.9 4.8 -5.1 19.7
Institutional quality INSTI 241 64.7 61.9 12.5 40.9 90.5
Table 6: Descriptive statistics (adjusted sample). MCR, SAV, INV, DCPS, VTR, TOR, INFL, DEFL, RIR, and PCF are in percentages, GDP and GDPCAP are natural logarithms, and INSTI is an index, where zero represents the weakest institutions and 100 embodies the best and most free institutions. A negative saving rate indicates that gross savings are negative, which occurs when total consumption is larger than gross national income. A negative inflation rate and negative GDP deflator imply deflation, while a negative real interest rate indicates that the inflation rate is greater than the nominal interest rate. Negative PCF entails that disinvestment in an economy is larger than new investment inflows.
5.1. Macroeconomic determinants of SMD
Model 1 2 3 4 5 6 7 8
Hypothesis 1a, 2a, 3, 4a
1b 2b 4b 5a 5b 5c 6
Income (GDP) 1.064 0.893 1.407 1.064 1.239 1.135 1.142
2.53 *** 2.25 *** 3.11 *** 2.52 *** 2.97 *** 2.71 *** 2.65 *** Income (GDP per cap) 1.125
2.69 *** Saving rate -0.006 -0.006 -0.008 -0.006 -0.007 -0.005 -0.006 -0.83 -0.79 -0.96 -0.77 -0.96 -0.73 -0.82 Investment rate 0.002 0.28 Domestic credit 0.004 0.004 0.004 0.005 0.004 0.004 0.004 0.004 2.02 ** 2.05 ** 1.65 * 2.22 ** 1.99 ** 1.99 ** 2.11 ** 1.85 * Value traded 0.001 0.001 0.001 0.001 0.001 0.001 0.001 1.53 1.24 1.50 1.50 1.39 0.93 1.40 Turnover ratio 0.000 0.64 Inflation rate -0.005 -0.50 Deflator -0.023 -3.01 ***
Real interest rate 0.008
1.03
Private capital flows 0.009
0.73 Adj. R-squared 0.60 0.63 0.60 0.46 0.61 0.61 0.59 0.63 Observations 241 241 241 222 241 241 230 241 Redundant test 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 Hausman test 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 VIF 2.54 2.73 2.56 1.90 2.59 2.61 2.51 2.77
The statistical outcomes of the regressions show some significant deviations from existing literature. Model 1 represents the default regression.28 It can be observed from the output that the coefficients and significance levels of the variables included are fairly constant across all models. The data provides sound evidence for hypotheses 1a and 3, as last year’s real income level and the domestic credit to GDP ratio both have a significant and positive impact on market capitalization ratios of Asian-Pacific economies. The coefficient of the income variable is considerable larger though than the domestic credit coefficient, which implies that income levels have a far larger impact on levels of SMD than the other indicators. The significant impact of the domestic credit variable provides additional evidence that banks and stock markets are each others complements rather than substitutes.
Model 1 does not provide evidence that the value traded ratio and the domestic saving rate are significant predictors of levels of SMD in Asia-Pacific, which is in strong contrast with previous finding in general, but in particularly with the results found by Garcia and Liu (1999) who focus on the same region. While the value traded variable has the expected sign, the coefficient is close to zero and the relationship is not found to be significant at the 10% level.29 The saving rate shows a negative sign, whereas a positive relationship is hypothesized. This implies that saving rates have a negative impact on levels of SMD. This counterintuitive finding can however be explained when you consider various declining saving rates for the period and economies involved.30 As a consequence, hypotheses 2a and 4a are not accepted.
To test the effect of an alternative measure of income, Model 2 includes the real GDP per capita variable. The table shows that GDP per capita has a positive impact on levels of SMD in Asia-Pacific, which is in line with the real GDP variable. The coefficient is again large and significant, which provides sufficient evidence for the rejection of the null hypothesis that income levels have no impact on levels of SMD in Asia-Pacific. Nonetheless, it should be highlighted again that this variable is highly correlated with the institutional quality indicators. The inclusion of the institutional measures must therefore prove whether this effect runs through the GDP per capita or institutional variables.
Model 3 replaces the saving rate with the investment rate. Compared to the saving rate, the impact of the investment rate variable is statistically less significant. However, as hypothesized the investment rate is positively related to levels of SMD, which is in contrast to the saving rate.
28
The default regression consists out of the four variables that are most often found to have a significant impact on levels of SMD in related studies (see Table A1 in the appendix).
29
The effect is statistically significant at the 15% level in Model 1, 3, and 5.
But given the low significance level of this relationship, it is unlikely that the investment rate is a more accurate indicator of SMD than the saving rate.
To test the effect of another measure of stock market liquidity, Model 4 replaces last year’s value traded ratio with last year’s turnover ratio. As the coefficient is zero, changes in the turnover ratio do not have an impact on levels of SMD. This observation provides further evidence that stock market liquidity on an economy-wide basis is a more accurate predictor of SMD than an indicator of low transaction costs. Additionally, the relatively low value of the coefficient of determination (the adjusted R-squared value) indicates that the inclusion of the turnover ratio has a negative effect on the model’s ability to explain future outcomes.
Model 5, 6, and 7 each adds a measure of macroeconomic instability to the default regression. No evidence is provided for hypotheses 5a and 5c, as the inflation rate and real interest rate variables are statistically insignificant. Besides, the real interest rate variable shows a positive sign. On the contrary, the GDP deflator is found to have a statistically significant impact on levels of SMD. When an economy’s GDP deflator decreases by one percentage point, the market capitalization ratio is expected to increase by 0.023 percentage point. This observation implies that the GDP deflator is a more accurate indicator of SMD than the inflation rate in Asia-Pacific, even though the correlation between the inflation rate and the GDP deflator is high and both variables are accepted measures of inflation.31 The reason why the GDP deflator is a more accurate estimator is likely to be related to the variable’s ability to quickly reflect new expenditure patterns, caused by consumers’ reactions to changing prices.
Lastly, the effect of private capital flows on SMD is tested in Model 8. The results show that the private capital flows variable is positively related to SMD, but this relationship is insignificant. Still, the inclusion of the private capital flows variable leads to a relatively high coefficient of determination value. It appears that the inclusion of the fixed effects model filters out some of the explanatory power of the private capital flow variable.
5.2. Institutional determinants of SMD
The next section reports the impact of institutional variables on levels of SMD. Institutional quality is estimated by considering several components of the Index of Economic Freedom (Heritage Foundation, 2010). Table 8 presents the outcome of the regressions.
Model 1 2 3 4 5 6 7 8 Hypothesis 7 7a 7b 7c 7d 7e 7f 7g Income (GDP) 1.067 1.033 1.133 1.498 0.851 1.076 1.004 1.109 2.53 *** 2.43 *** 2.66 *** 3.32 *** 1.80 * 2.49 *** 2.35 ** 2.57 ** Savings rate -0.006 -0.006 -0.006 -0.010 -0.006 -0.006 -0.005 -0.006 -0.87 -0.75 -0.84 -1.30 -0.77 -0.84 -0.73 -0.85 Domestic credit 0.004 0.004 0.005 0.001 0.004 0.004 0.004 0.004 1.65 * 2.02 ** 2.24 ** 0.45 2.07 ** 1.97 ** 1.88 * 1.92 * Value traded 0.001 0.001 0.001 0.001 0.002 0.001 0.001 0.001 1.39 1.50 1.60 0.92 1.78 * 1.46 1.53 1.53
Overall inst. index 0.009 0.76 Corruption index 0.002 0.57 Property rights -0.004 -1.09 Business freedom 0.014 2.48 ** Trade freedom 0.004 0.89 Investment freedom 0.000 0.14 Monetary freedom 0.004 0.89 Financial freedom 0.002 0.50 Adj. R-squared 0.69 0.65 0.65 0.68 0.67 0.64 0.62 0.62 Observations 241 241 241 241 241 241 241 241 Redundant test 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 Hausman test 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 VIF 3.10 2.77 2.70 3.12 2.73 2.58 2.70 2.70
The indicators included in Model 1 show roughly the same coefficients and significance levels as in the default regression in Table 7. The saving rate and value traded ratio are statistically insignificant, while income and domestic credit are significantly related to levels of SMD. Model 1 adds the overall institutional index to the default regression. Even though the coefficient is relatively large, the relationship is statistically insignificant. Hence, the model proves that institutional quality in general is not a considerable determinant of SMD in Asia-Pacific.
Model 2 to 8 represent hypotheses 7a to 7h, which examine the individual components of the overall institutional quality index. The table demonstrates that out of the seven variables included, only the business freedom index is found to have a significant impact on SMD. Even though institutional quality in general is no significant predictor of SMD in Asia-Pacific, making it easier to do business in an economy can still have a positive effect on the relative market capitalization of that economy. While the business freedom index is statistically significant, the domestic credit variable becomes insignificant in model 4. The reason for this can be found in the rather high correlation between the business freedom index and the domestic credit variable (see Table A5 in the appendix).
The table furthermore shows that the property rights index is negatively related to the dependent variable, while the investment freedom index has no impact on an economy’s level of SMD. Model 2, 6, 7, and 8 illustrate that changes in the freedom from corruption, trade freedom, monetary freedom, and financial freedom index have a positive effect on levels of SMD, albeit statistically insignificant. Interestingly, the value traded variable becomes significant at the 10% level when the trade freedom index is included in the model.
5.3. Implications for advanced and emerging Asia
This section makes use of a development dummy to make a distinction between advanced and emerging economies in Asia-Pacific (see Table A2 in the appendix), and examines whether SMD levels for those two groups of economies are influenced by different factors.32 Table 9 presents the coefficients and significance levels of the indicators for the advanced and emerging economies. When the coefficients for a specific indicator are found to be statistically different, the output in Model 7 and 8 might be biased towards the group of emerging economies, as the sample consists out of twelve emerging economies and only six advanced economies.
Model 1 in Table 9, the default regression, shows that income (as measured by GDP) is again an important indicator of SMD, as the coefficients for both groups of economies are relatively large. The outcomes are statistically significant, which demonstrates that the coefficients for emerging and advanced economies are indeed statistically different for this indicator. Interestingly, income levels are positively related to SMD in emerging economies, whereas this relationship is negative for advanced economies. Hence, real GDP growth has a positive impact on levels of SMD in emerging countries, but once they turn into developed economies, the impact of income growth on SMD becomes negative. This observation can partly be explained by considering the structure of the dependent variable, as the market capitalization ratio corrects total market capitalization of an economy by its income level. The growth rate of total market capitalization must therefore be larger in absolute values than the income growth rate in emerging economies.33 When economies reach a certain development threshold level, market capitalization growth can no longer keep up with GDP growth, and hence the ratio of market capitalization to GDP declines.
32
Billmeier and Massa (2007) use a comparable approach, as they implement a dummy which takes into account the resource endowment levels of economies. They find significant differences between resource-rich and resource-poor economies in the Middle East and Central Asia.
33
Model 1 2 3 4 5 6 7 8
Hypothesis 1a, 2a,
3, 4a 1b 2b 5b 7 7c 7d 7e Income (GDP) E 0.750 0.581 1.006 0.713 1.271 0.713 0.854 1.73 * 1.33 2.35 ** 1.64 2.75 *** 1.49 1.95 * A -0.953 -1.832 -0.903 -0.946 -0.991 -0.493 -0.855 -2.39 ** -3.54 *** -2.32 ** -2.36 ** -2.55 ** -1.04 -2.05 **
Inc. (GDP per cap) E 0.872
2.01 ** A -1.237 -2.29 ** Saving rate E -0.004 -0.004 -0.002 -0.005 -0.004 -0.009 -0.004 -0.004 -0.52 -0.57 -0.31 -0.69 -0.50 -1.23 -0.49 -0.61 A -0.035 -0.032 -0.011 -0.034 -0.028 -0.032 -0.017 -0.028 -1.49 -1.37 -0.43 -1.50 -1.14 -1.42 -0.67 -1.19 Investment rate E -0.002 -0.19 A -0.054 -2.24 ** Domestic credit E 0.008 0.007 0.008 0.007 0.006 0.004 0.008 0.007 2.71 *** 2.55 ** 2.38 ** 2.59 *** 1.95 * 1.41 2.61 *** 2.56 ** A -0.005 -0.004 0.001 -0.004 -0.003 0.000 -0.006 -0.005 -1.16 -1.08 0.24 -1.06 -0.58 -0.09 -1.37 -1.19 Value traded E 0.003 0.003 0.003 0.003 0.003 0.002 0.003 0.003 2.33 ** 2.19 ** 2.24 ** 2.12 ** 2.09 ** 1.13 2.43 ** 2.25 ** A -0.002 -0.002 -0.002 -0.002 -0.002 0.000 -0.002 -0.003 -1.03 -1.04 -1.13 -0.86 -0.89 -0.06 -1.28 -1.40 Deflator E -0.019 -2.42 ** A -0.038 -1.83 *
Overall inst. Index E 0.016
1.22 A -0.045 -1.55 Business freedom E 0.023 3.45 *** A -0.027 -2.47 ** Trade freedom E 0.001 0.22 A -0.055 -2.15 ** Investment freedom E 0.000 -0.10 A 0.018 2.10 ** Adj R-Squared 0.62 0.67 0.65 0.63 0.72 0.71 0.71 0.68 Observations 241 241 241 241 241 241 241 241 VIF 2.73 3.09 2.93 2.79 3.77 3.61 3.66 3.28
