The relationship between the stock market and economic growth

(1)

The Relationship Between the Stock Market and

Economic Growth

Bachelor Thesis

Student:

Lars Kulk

10275568

Supervisor:

Egle Jakucionyte

June 28, 2015

Bachelor Economics and Business

Economics and Finance

(2)

Abstract

This thesis examines how the stock market influences long-term economic growth of both the developed and developing countries. The relationship between economic growth and multiple stock markets variables is examined separately using multiple regressions. There is a clear link between the stock market and economic growth. The results from this thesis show that liquid stock markets promote economic growth. Volatility negatively influences economic growth. A small difference in the strength of the relationship between liquidity and economic growth for the developed and developing countries is found.

(3)

1. Introduction

The stock market is one of the most widely known parts of an economy. Almost everyone knows about the stock market, regardless of the economic background people may have. Many people own shares in one or more firms. For some, it is their pension, for others, trading is just recreational. While the stock market is one of the most known parts in an economy, is it also an important part? Does the stock market, for example, create economic growth? There are many determinants of long-term economic growth and the stock market is one of them. For example, the liquidity provided by the stock market (Levine & Zervos, 1998, p.554). Knowledge of how the relationship between the stock market and long-term economic growth works is very useful for many economic agents. Actions by economic agents influence the relationship between the stock market and long-term economic growth. When economic agents know how they can jointly influence the stock market and long-term economic growth, they can adjust their actions. This could create a more efficient economy with more economic growth and welfare.

This thesis focuses on the relationship between the stock market and long-term economic growth. There is substantial research about long-term economic growth, but very few studies include the stock market as a determinant of economic growth. Debate exists on the importance of the stock market. For example, Devereux and Smith (1994, p. 547) state that risk diversification through stock markets can reduce the savings rate, which negatively influences economic growth. On the other hand, Arestis, Demetriades, Luintel (2001, p.37) state that, while the stock market is important for economic growth, its influence is small compared to the influence of the banking system. These two articles show that there is no clear and consensual view on the relationship between the stock market and economic growth. This thesis aims to investigate if and how the stock market influences long-term economic growth and if there are differences in this relationship between the developed and developing countries. This thesis will research the possible differences between these two types of countries, because the developing economies grow at a higher pace than the developed economies and this could have an influence on the relationship between the stock market and economic growth. The research question of this thesis will be as follows: How does the stock market influence long-term economic growth in both developed and developing countries? Knowledge on this issue in relevant for the developing countries, because it can help their

(4)

development. The stock market could possibly be an important contribution to the economic development of those countries.

Our prediction is that the stock market will positively influence long-term economic growth, because the stock market increases investments. Firms can raise equity on the stock market, which can be invested into projects that benefit for the future of that firm. It is also expected that stock markets create extra control on firms’ actions, because the ownership and management of the firms are divided. The owners want to check how the management uses their capital. This control from shareholders will create a better performance of firms. The increase in investments and the control created by the stock market will create economic growth. We do expect differences in the relationship between the stock market and growth in the developed and developing countries. Firms in the latter countries have more growth opportunities and equity from the stock market can be very useful to accomplish growth.

The research question will be answered using a literature review and an empirical study. The literature review will discuss the existing literature on the topic of this thesis. The literature review will explain the relationship between the stock market and long-term

economic growth. In addition, the literature review will elaborate on the possible difference in the relationship between the stock market and economic growth in the developed and developing countries. The empirical study will show how strong the relationship between the stock market and long-term economic growth is and which variables are more important than others. For the stock market the three variables size, liquidity and volatility will be used. For economic growth the two variables output growth and savings will be used. A dummy

variable will be used to study the possible difference between the developed and developing countries. The regression analysis with these variables will show the strength of the

relationship and which variables are significant. A dataset including the developed and developing countries will be used to create the variables. The average values from the time series data will be used from 1994 until 2012 in a cross-section regression.

In what follows, an overview of relevant theory about the relationship between the stock market and economic growth will be given in part 2. Part 3 will provide a description of the research method. Part 4 will present and discuss the results of the empirical study. The results will be compared to the results of Levine and Zervos (1998). Finally, Part 5 will draw the conclusions and answer the research question.

(5)

2. Literature review

Economic growth is dependent on multiple determinants. Two examples of these

determinants are human capital and fiscal policy. Another source is financial development. Financial development includes stock market development, which is the topic of this thesis. Knowledge about the relationship between the stock market and economic growth is useful for many economic agents, because their actions influence this relationship. The difference in the relationship between the developed and developing countries can also be important knowledge. Economic growth is important for the growth of welfare for all people.

2.1 The main determinants of economic growth

Human capital is one of the basic determinants of economic growth. Human capital represents the knowledge, skills, and experience of the labour force. When human capital is integrated in the Solow model, it will significantly improve in terms of explaining growth and income differences (Gartner, 2009, p. 291).

Accumulation of physical capital is another basic determinant of economic growth. The accumulation of physical capital is an important part of the Solow model. The Solow model describes economic growth. Physical capital stock accumulates by investments and decreases by depreciation (Gartner, 2009, p. 249).

Research and development is another determinant of economic growth and stimulates the technological process. Research and development can be seen as an investment into a new and more efficient technology, leading to higher growth rates (Bassanini & Scarpetta, 2001, p. 15).

Macroeconomic policy is mainly focused on maintaining low inflation. Low inflation is important for economic growth, because it reduces uncertainty in the economy.

Furthermore, it enhances efficiency of the price mechanisms (Bassanini & Scarpetta, 2001, p. 16).

Fiscal policy is manipulating government spending and government revenues by taxes to achieve specific goals (Gartner, 2009, p. 125). Fiscal policy can influence economic

growth. Taxes negatively influence economic growth, because they worsen the efficient allocation of resources. Government spending positively influences growth. For example,

(6)

when a government finances education, human capital will increase (Bassanini & Scarpetta, 2001, pp. 18-20).

International trade is a determinant of economic growth. International trade can create growth by exploiting comparative advantages. Trade can also promote economies of scale and exposure to competition, as well as widen the scope of knowledge. This can result in higher efficiency and more investment (Bassanini & Scarpetta, 2001, p. 20).

Financial development is the last determinant of economic growth discussed and the topic of this thesis. Stock market development is a part of financial development. Another part is the banking system. The financial system provides funding for investments (Bassanini & Scarpetta, 2001, p. 20). A solid banking system is central for economic growth, more

important than a stock market. A banking system is of a great importance for the development of an economy, because banks can provide credit used for investments (King & Levine, 1993, p. 533). Stock markets do contribute to economic growth; however, their contribution is small compared to the contribution of the banking system, because corporate investments are not financed by issuing equity (Arestis, Demetriades, Luintel, 2001, p. 19).

2.2 The relationship of economic growth with the stock market

It is common knowledge that new flows of information influence stock markets. Positive economic news has a positive effect on the stock markets. This basically means the indices increase, because the prices of stock rise caused by an increase in the demand for stocks. The demand increases, because publicly owned firms can benefit from this positive news. Chen, Roll and Ross (1986, p. 384) also state that economic state variables influence the stock market aggregates. If this was not the case, all risk in the stock markets could be diversified. This is not the case, so there is an undiversifiable risk called systematic risk. In this case, the economy influences the stock markets, but it can also be the other way around.

There is extensive literature stating that the stock market influences the economy, specifically, economic growth. In their empirical paper on this relationship, Levine and Zervos (1996, p. 325-26) show that a discussion exists on whether the financial system is important for economic growth. Arestis et al. (2001) and Caporale, Howells, and Soliman (2004) state that the stock market influences economic growth. Some authors do not even mention the financial system in their research on economic growth. Levine and Zervos (1996, p. 326) performed an empirical study to research the importance of a stock market for

(7)

economic growth. Real per capita GDP is one of the variables used for long-term economic growth. This is regressed against multiple variables that describe the development of the stock market. The results suggest a positive effect of the stock market on long-term economic growth. The effect is also significant. This does not explain a causal relationship between the stock market and long-term economic growth, but only suggests a significant correlation between them. To establish a causal relationship, further empirical research is needed. The correlation is empirically studied with the channels of liquidity and risk diversification (Levine & Zervos, 1996, pp. 335-36). The theory suggests that there are more channels through which the stock market influences long-term economic growth. These include the liquidity function, risk diversification, acquisition of information about firms, corporate control, and savings mobilization (Levine & Zervos, 1996, p. 326).

Stock markets are a tool for investors that can provide liquidity, because investors can buy and sell shares at any moment in time. There is one requirement for a stock market in providing liquidity: there should be enough investors active in the stock market. This liquidity is useful for long-term economic growth, because it promotes investments. High-return investments need a long-term commitment from investors. Investors are reluctant to give up control over their capital for a long period. The liquidity of stock markets is useful, as investors can always sell their shares and firms can always raise equity needed for these investments (Bencivenga, Smith, & Starr 1996, p. 242). With more liquidity, more high-return investments are possible, as investors can easily step in and out of the investment at any moment. More high-return investments create long-term economic growth (Levine & Zervos, 1996, p. 327). On the other hand, Bencivenga and Smith (1991, p.204) state that, by reducing uncertainty, liquidity can decrease the savings rate. This can slow down economic growth.

Risk diversification can influence economic growth through internationally integrated stock markets. Stock markets are a tool in risk diversification. Investors can and will hold diversified portfolios to decrease systematic risk. Risk diversification causes investors to shift to high-return investments and, as with the liquidity channel, these investments promote long-term economic growth (Obstfeld, 1992, p. 35). In this channel, there is again the possibility that the savings rate decreases because of the risk in the stock markets that has declined. A decrease in the savings rate can slow down economic growth (Devereux & Smith, 1994, p. 547).

Stock markets promote the acquisition of information on publicly traded firms. Acquisition of information about firms influence economic growth by the possibility to make a return from information that an investor has. These returns are generated by trading. This

(8)

trade is based on the information the investor possesses about a firm before the information becomes widely known to other investors. The profitability of information stimulates investors to research and follow every move of these firms. This acquisition of information stimulates a more efficient allocation of capital, which creates long-term economic growth (Grossman & Stiglitz, 1980, p. 404). There is a debate on the influence of information, as the theory about efficient capital markets states that the available information is fully reflected by the prices. This makes it impossible to trade on information and make a return (Fama, 1970, p. 384).

Corporate control can influence long-term economic growth. First, the principal-agent problem is alleviated by efficient stock markets, as shareholders have more information about the activities of the manager and investments opportunities of the firm. This makes it easier to connect the compensation of the manager to stock performance (Jensen & Murphy, 1990, p. 226). Secondly, takeover threats stimulate managers to maximize their equity price, which is good for shareholders and alleviates the principal-agents problem. Both stimulate efficient capital allocation and economic growth (Scharfstein, 1988, p. 192). Again, there is a debate on this issue. Stiglitz (1985, p. 137) states that takeover threats would not improve corporate control, because outsiders are hesitant about a takeover, since they have less information than the insiders. Furthermore, the stock market development creates a more diffuse ownership and this impedes corporate control (Bhide, 1993, p. 43).

Savings mobilization can influence long-term economic growth. When looking at raising capital at large, liquid and efficient stock markets can help savings mobilization. The accumulated savings can be invested in investments projects, creating economic growth. There is some debate on this issue, because the fraction of newly raised capital used for investments is small. As a result, the effect on long-term economic growth is small (Levine & Zervos, 1996, p. 328).

Price volatility is another channel, which is not discussed by Levine and Zervos (1996), but which does influence economic growth. Price volatility in a stock market negatively influences long-term economic growth. As discussed above, stock markets have the ability to promote efficient allocation of capital. This efficient allocation creates economic growth. Some volatility on the stock market is not bad, as it can reflect the flows of new information in an efficient market. However, excess volatility on the stock market has

negative effects. These effects are inefficient allocation of capital and increasing interest rates caused by uncertainty. Both effects reduce productivity and investments. Therefore, they negatively influence economic growth (Arestis et al., 2001, p. 18).

(9)

There are no differences between the developed and developing countries in the way the stock market influences economic growth. The five channels described by Levine and Zervos (1996) and the volatility channel work for both the developed and developing countries. In the dataset of 41 countries, Levine and Zervos (1996) use both the developed and developing countries. There could be a difference in the strength of the relationship between the stock market and long-term economic growth. While the developed and most advanced developing countries have developed stock markets, poorly developed countries have undeveloped stock markets or not even a stock market (Singh, 1991, p. 64). This difference in the development of the stock markets could influence the strength of the

relationship, because more developed stock markets can more efficiently raise equity. On the other hand, the growth rates in the developing countries with less developed stock markets are higher and this could also influence the strength of the relationship between the stock market and economic growth. The empirical part of this thesis will investigate this issue.

This thesis will reexamine the relationship between the stock market and economic growth in the same empirical way as Levine and Zervos (1998) did. In this empirical study, however, a different dataset will be used. Specifically, while Levine and Zervos (1998) did not consider the differences in the strength of the relationship for the developed and

(10)

3. Method

The research question will be answered by using the literature review and an empirical study. The literature is the basis of the empirical part of this thesis.

3.1 Dataset

For the dataset developed and developing countries are used in this thesis. Using these two types of countries makes it possible to look if there is a difference in the influence of the stock market. The Worldbank ranks countries. While the developing countries are middle- and low-income countries, the developed countries are high-low-income countries. There are 75 developed countries in the world and 139 developing countries, from which 34 are classified as low-income (‘’Country and Lending Groups’’, 2015). The dataset consist of 30 developed

countries and 30 developing countries. These countries are selected on the availability of data. Countries with most data available are used in the dataset. The reason for this criterion is that the data availability of the developing countries is low and using this criterion assures that sufficient data are available for our empirical research. From these countries time-series data will be used from 1994 until 2012 in a cross-section regression. A cross-country regression instead of a time-series regression is used because this is best suited for growth analyses and there are fewer econometric problems, such as autocorrelation. For the growth variables, the averages of each country are used and, for the stock market variables and control variables, the data from 1994 are used, as, if averages are used for the stock market and control

variables, then a common shock in these variables can influence the results. Second, using the averages for the stock market and control variables will not account for an endogenous

influence on the variables in the regression (Levine & Zervos, 1998, p. 544).

3.2 Variables

The first variable used to describe the stock market development is size. This variable shows the market capitalization of the listed companies in a specific country as a percentage of that country’s GDP in 1994. While many stock markets are not perfectly efficient due to the distortion of taxes, many see capitalization as a good indicator for the stock market

(11)

development (Levine & Zervos, 1998, p. 540). There is one observation for each country and the data are retrieved from the databank of the Worldbank.

Liquidity is the second variable used to describe the development of the stock market. Two ratios will be used to measure liquidity. The first one is the turnover ratio. The turnover ratio is the value traded divided by the market capitalization. The second ratio is the value traded ratio in which value traded is divided by GDP. Data for both ratios are from 1994. Each country has one observation for each ratio. The theory suggests a positive relationship between liquidity and economic growth (Bencivenga, Smith & Starr 1996, p. 242). The data are retrieved from the databank of the Worldbank.

Volatility is the third and last variable used to describe the stock market development. The theory suggests a negative relationship between volatility and long-term economic growth (Arestis, Demetriades, Luintel, 2001, p. 18). Volatility is calculated by considering the standard deviation of the returns. These are the returns from the domestic stock market indices, such as, for example, the S&P500 for the United States. The volatility is calculated over the year 1994. There is one observation for each country in the dataset.

A dummy variable for the developed/developing countries will show if there is a difference in the strength of the relationship between the stock market and economic growth. In the results, the dummy variable is called developed. The binary variable is 1 for the developed countries and 0 for the developing countries. Multiplying the binary variable with the stock market variable generates the dummy variable. The multiplying is done, as then the dummy variable will show the difference in the strength of the relationship between the stock market variable with the growth variable. The relationship for the developed countries is stronger than that for the developing countries when the dummy variable is positive. If the dummy variable is negative, the relationship of developing countries is stronger.

The first growth variable used in the regressions is output growth. For long-term economic growth, the variable output growth is used. Output growth equals real per capita GDP growth. This is given in a percentage. The average of output growth is calculated over the period from 1994 to 2012. There is one observation for each country in the dataset. The data is retrieved from the databank of the Worldbank.

The second growth variable used is savings. The variable savings equals gross private savings. The gross private savings are given as a percentage of the GDP. This makes the variable comparable across different countries. The average over the period 1994 -2012 is calculated creating one observation per country. The data are retrieved from the databank of the Worldbank.

(12)

The control variables used in the regressions are initial output, government

expenditure, enrollment rate, inflation, and international trade. Initial output is the logarithm of real output per capita in 1994. The control variable government is the government

expenditure as a percentage of the GDP. For enrollment rate, the logarithm of the secondary school enrollment rate in 1994 is used. For inflation, the GDP deflator from 1994 is used. International trade is the last control variable. Like the other control variables, international trade is a determinant of economic growth. For international trade, we use the sum of export and import of goods and services as a percentage of GDP. There is one observation per country for each control variable. The data for all control variables are retrieved from the Worldbank database.

3.3 Regression analyses

In the regression, the empirical relationship between the stock market and long-term

economic growth will be investigated. The growth variables, output growth and savings, will be regressed on the stock market variables and the dummy variable. In the cross-county regression, output growth and savings are used as dependent variables and the stock market variables size, liquidity, and volatility and the dummy variable as independent variables. Ten regressions will be performed with one of the growth variables on each stock market variables separately. The control variables specified above will be added in each of the ten regressions. Control variables are used to correctly assess the empirical relationship between the stock market variable and the growth variable (Levine & Zervos, 1998, p. 544). For the stock market variable liquidity, there are six regressions: two regressions, including the turnover ratio, and four with the value traded ratio. In the ratio value traded, there is the problem with the price-effect. The price-effect is that when investors anticipate profits, then the stock price will increase. This causes an increase in the value traded ratio without a change in the number of transactions. Regressing a growth variable against value traded can yield bad results, because it is not the increase in liquidity influencing the growth variable, but the stock price rise. The solution for this problem is that in this regression the stock market variable size is added in the regression (Levine & Zervos, 1998, p. 540). These ten regressions will show how each stock market variable influences long-term economic growth. In each regression, a dummy variable is included to assess the difference between the developed and developing countries. In the ten regressions, the option robust is used. This means that heteroskedastic

(13)

standard errors are calculated. Heteroskedastic standard errors are used because they ensure the validity of the results. Heteroskedastic standard errors are valid whether or not the

standard errors are heteroskedastic (Stock & Watson, 2012, p. 201). Heteroskedastic standard errors are also useful, because then the result can be compared to the results van Levine and Zervos (1998).

A possible IV regression to estimate the relationship between economic growth and the stock market, which is not performed in this thesis, is to use an investment variable as an instrument. The stock market variable used in the cross-country regression in this thesis should be regressed on the investment variable. In the second stage regression, the growth variables used in this thesis should be regressed on the investment variable and control variables should be included. In most of the channels described in the literature review, the stock market influences economic growth through investments. Stock market development promotes investments and these investments create economic growth. This IV regression could give an insightful overview on how strong the relationship between investments and stock market development is and on the relationship between investments and economic growth. The included control variables are other determinants of economic growth. The instrument variable investment is exogenous, because the correlation with the residual is zero.

(14)

4 Results

Each table represents two regressions with one of the stock market variables. A summary of the statistics used is included in the appendix.

Table 1

Dependent variables Independent

variables

Output growth Savings

Size -.0003145 (.0052768) .0351804 (.0246378) Developed .0038417 (.0108326) .0079638 (.0647641) Enrollment 1.544517 (1.26055) -5.793164 (6.668485) Government .0109674 (.0677891) -.1970833 (.3500731) Inflation .0011899 (.0029739) .0271146 (.0220135) Trade -.0000822 (.0060377) .0347889 (.0252712) Initial output -.9886619 ** (.4151179) 2.295701 (2.697864) R-squared 0.2999 0.1471 Observations 45 45

Notes: Heteroskedastic standard errors in parentheses. Significant at 10%=*, 5%=** and 1%=***.

Table 1 shows the results of the regressions in which the stock market variable size is included. The stock market variable size does not significantly enter the regression with output growth, meaning that no relationship is found between the market capitalization of the stock market and the growth rate of GDP. This is consistent with the results reported by Levine and Zervos (1998) where no reliable link between size and output growth was found. In their results, a significant coefficient for size was found, but this largely depended on some

(15)

countries in their dataset. A possible explanation for this finding can be that there is a measurement error, especially in the developing countries, where the stock markets are not very developed yet and where the information about market capitalization could be imprecise. When these countries are removed, the stock market variable becomes insignificant. Only the control variable initial output enters the regression significant at a 5% level. The control variable initial output negatively influences output growth. This means that the countries with a higher level of GDP per capita in 1994 have a lower GDP growth rate over the period 1994 till 2012. This result is logical considering the used dataset: the developing countries have higher growth rates and a lower level of GDP per capita in 1994 as compared to the

developed countries. All other control variables do not significantly enter the regression. A possible explanation for these insignificant control variables can be related to the used dataset, where, some of the control variables have multiple large outliers. The summary of statistics in Appendix 1 shows the minimum and maximum of the variables and for the control variables inflation, school enrollment and trade, they are far apart. The maximum inflation is from Brazil and is more than five standard deviations away from the mean. There is no evidence suggesting that there is a difference in the strength of relationship between the stock market variable size and output growth for the developed and developing countries because, the dummy variable does not significantly enter the regression. A possible reason is that the stock market variable did not significantly enter the regression and the dummy variable is

calculated using the stock market variable size.

The variable size does not significantly enter the regression with savings either. This result is consistent with Levine and Zervos (1998) who found no significant relationship between size and savings. These results suggest that market capitalization is not a determinant of economic growth. In the regression with saving, none of the control variables entered significantly. A possible explanation for these insignificant results could again be the presence of large outliers in the dataset. There is no evidence to suggest that there is a difference in the strength of relationship between the stock market variable size and savings for the developed and developing countries, because the dummy variable does not enter significantly. Again, this could be caused by the insignificant stock market variable.

Table 2

Dependent variables

(16)

variables Turnover .0296921 *** (.0048694) .1497878 *** (.0226372) Developed -.0205089 *** (.0068717) -.0868949 * (.0447483) Enrollment 2.23455 * (1.225972) -2.109696 (6.126193) Government -.0088705 (.0457538) -.3580515 (.28142) Inflation -.0068702 (.0041666) -.0224221 (.0394607) Trade .0031683 (.0046848) .0632611 ** (.0276646) Initial output -.8349913 *** (.2450945) 3.065349 (2.175367) R-squared 0.6956 0.4756 Observations 42 42

Table 2 shows the results of the regressions in which the stock market variable turnover is included. The stock market development variable turnover significantly enters the regression with output growth at a 1% level. The relationship between the variable turnover and output growth is positive, though small. If turnover increased by 1%, then output growth would increase by 0.03 percentage points per year. Over 18 years, from 1994 till 2012, this would be 0.54%. This result is consistent with the theory from Bencivenga, Smith and Starr (1996) stating that there should be a positive relationship between the liquidity function of a stock market and economic growth. This result is consistent with the results of Levine and Zervos (1998). The control variables enrollment and initial output significantly enter the regression at a 10% and 1% significance level, respectively. Enrollment positively influences output

growth. This is consistent with economic literature on the issue. Gartner (2009, p. 291) states that human capital is one of the basic determinants of economic growth. Initial output has a negative relationship with output growth. This result is consistent with the results reported in Table 1. The results suggest that there is a difference in the strength of the relationship

(17)

between the stock market variable turnover and output growth for the developed and

developing countries. The dummy variable is significant at a 1% level. The coefficient for the dummy variable is very small. This means that the difference in the strength of the

relationship with the stock market variable is small, but the relationship is stronger for the developing countries. This shows that liquidity has a greater contribution to economic growth in the developing countries as compared to the developed countries.

The stock market development variable turnover significantly enters the regression with savings at a 1% level. The turnover variable has a positive relationship with savings. If turnover increased by 1%, then savings would increase by 0.15 percentage points. In 18 years, this would amount to an increase by 2.74%. A possible explanation for this result can be that liquidity causes economic growth and, with economic growth, savings will increase. On the other hand, Bencivenga and Smith (1991, p.204) state that more liquidity can decrease the savings rate. This result is not consistent with the results of Levine and Zervos (1998), because these authors find no significant relationship between turnover and savings. A possible explanation these results differ with the results from Levine and Zervos can be that they added the variable bank credit to their regression. Only the control variable trade enters the regression significantly at a 5% level. This is consistent with the economic theory.

International trade is a determinant of economic growth by exploiting comparative advantages (Bassanini & Scarpetta, 2001, p. 20). The other control variables are insignificant. There is a significant difference in how turnover influences savings in the developed and developing countries. The dummy variable significantly entered the regression with savings at a 10% level. The dummy variable shows that the relationship between turnover and savings is stronger in the developing countries.

Table 3

variables

Value traded .009395 (.0094212) .138912 *** (.0419603) Developed .0007356 (.0136237) -.0638647 (.0765072)

(18)

Enrollment 1.501329 (1.344951) -2.028971 (6.678112) Government .0145427 (.0618483) -.0659017 (.3257589) Inflation .001348 (.002942) .0219932 (.0217586) Trade -.0015792 (.0081706) .0031319 (.0418267) Initial output -1.003137 ** (.3922493) 1.434464 (2.535389) R-squared 0.3413 0.2407 Observations 44 44

Table 3 shows the results on the regressions in which the stock market variable value traded is included. Value traded does not significantly enter the regression with output growth,

meaning that no relationship is found between the liquidity variable value traded and output growth. A possible reason for the insignificant result could be the price effect. On the other hand, Levine and Zervos (1998) do find a significant relationship between value traded and output growth. Table 4 shows the regression in which the price-effect is controlled for. Only the control variable initial output significantly enters the regression with output growth at a 5% level. Initial output negatively influences output growth. This is consistent with previous results in the previous regressions. The results do not suggest a difference in the strength of the relationship between the stock market variable and output growth because the dummy variable is not significant. The dummy variable could be insignificant because the stock market variable, used to calculate the dummy variable, is insignificant.

Value traded does significantly enter the regression with savings at a 1% level. The results show a positive relationship between value traded and savings. If value traded increases by 1%, then savings would increase by 0.14 percentage points. In 18 years, this would be 2.55%. This is not consistent with the economic literature, as Bencivenga and Smith (1991, p.204) state that more liquidity can decrease the savings rate. This is also not

consistent with Levine and Zervos (1998), who find no significant relationship between value traded and savings. A possible reason could be that Levine and Zervos (1998) included the

(19)

variable bank credit into their regression. In this thesis, we are solely interested in the effect of the stock market, not the banking system. None of the control variables significantly entered the regression with savings. A possible explanation for these insignificant results can be the presence of large outliers in the dataset. These outliers can be found the variables inflation, school enrollment, and trade. The results do not suggest a difference in the strength of the relationship between the stock market variable and savings, because the dummy variable is not significant.

Table 4

variables

Value traded .040489 * (.0212218) .2495829 ** (.1142292) Size -.0194276 * (.0101791) -.0691474 (.0603316) Developed -.006766 (.0141019) -.0905644 (.0787558) Enrollment 2.108067 (1.469753) .13055 (6.801156) Government .0288855 (.0606474) -.0148526 (.329807) Inflation -.001154 (.003172) .013088 (.0247596) Trade -.0026912 (.0082865) -.0008261 (.0430981) Initial output -1.1624 *** (.4221294) .8676106 (2.525657) R-squared 0.4304 0.2805 Observations 44 44

(20)

Table 4 shows the regression in which the price-effect is controlled for. Both stock market variables value traded and size significantly enters the regression with output growth at 10%. If the price-effect influences the relationship between the stock market variable value traded and output growth; then, by including the stock market variable size in the regression, value traded should not remain significant (Levine & Zervos, 1998, p. 549). In Table 3, value traded was not significant meaning that the price-effect is not driving the relationship, because no relationship is found between value traded and economic growth. This result is consistent with the results of Levine and Zervos (1998), as they find no evidence for the price-effect either. Only the control variable initial output significantly enters the regression with output growth at a 1% level. This is consistent with the previous results. The dummy variable is insignificant meaning that there is no evidence to suggest a difference in the strength of the relationship between the stock market variables and output growth for developed and developing countries.

In the regression with savings, only the stock market variable value traded significantly enters at a 5%. The variable size is insignificant. This is not consistent with economic theory: as stated by Bencivenga and Smith (1991, p.204), more liquidity can decrease the savings rate and thereby savings. This result is not consistent with the results of Levine and Zervos (1998) either, because these authors find no significant relationship with both stock market variables. None of the control variables is significant. Again, a possible explanation for these insignificant results can be the presence of large outliers in the dataset. The results do not suggest a difference in the strength of the relationship between the stock market variable and savings, because the dummy variable is insignificant.

Table 5

variables

Volatility -.226616 ** (.0904803) .7469213 (.7051931) Developed .1108317 (.0747234) -.5312558 (.9074311) Enrollment -.2595969 (1.336863) -.0417897 (8.934776)

(21)

Government .012156 (.0322925) -.3889957 (.2669924) Inflation -.0984509 ** (.0423387) -1.162596 *** (.1767344) Trade .0051786 (.0035877) .1070462 *** (.0173176) Initial output -1.118292 *** (.3272843) -.1681981 (2.282082) R-squared 0.8300 0.8629 Observations 16 16

Table 5 shows the results on the regressions in which the stock market variable volatility is included. Volatility significantly enters the regression with output growth at 5%. The results show that volatility negatively influences output growth. If volatility increased by 1%, then output growth would decrease by 0.23 percentage point. In 18 years, this would amount to a decrease by 4.06%. This is in line the economic theory on this issue. Arestis et al., (2001, p. 18) state that volatility can negatively influence economic growth. This result is consistent with the results of Levine and Zervos (1998), who find the same relationship. Inflation enters the regression with output growth significantly at a 5% level. There is a negative relationship between inflation and economic growth. Some inflation is acceptable, but a high inflation will slow down investment, which, in turn, will slow down economic growth (Barro, 2013, p. 107). Initial output significantly enters the regression at 1% and this is consistent with the results of the previous regressions. The dummy variable is not significant. This suggests that there is no difference in the strength of the relationship between volatility and output growth for the developed and developing countries.

The stock market variable volatility does not significantly enter the regression with savings. The low number of observations could underlie this outcome. On the other hand, there is also no economic literature connecting stock market volatility with savings. This result is consistent with the results of Levine and Zervos (1998). The control variables inflation and trade significantly enter the regression with savings at 1%. The relationship between inflation and savings is negative, meaning that an increase in inflation causes a decrease in savings. There is a logical explanation for this relationship. Inflation causes the

(22)

real value of the savings to fall (Von Ungern-Sternberg, 1981, p. 974). This causes that fewer people to save money. The coefficient for trade is positive, meaning that trade positively influences savings. This is in line with the economic theory: specifically, Agénor &

Aizenman (2004, p. 336) show that the terms of trade indeed do positively influence savings. Their results also show a small, but positive relationship. There is no evidence to suggest a difference in the strength of the relationship between volatility and saving, because the dummy variable is not significant in the regression.

According to the results presented above, some stock market variables influence economic growth, but others do not. Specifically, the results suggest that the stock market variables size and value traded do not influence economic growth. To check if all stock

market variables together influence economic growth, this thesis will perform four regressions with all the stock market variables included on output growth and savings. An F-test can conclude if the stock market variables together are significantly different than zero. H0: β1 =β2 = β3 =β4 =0 and H1: one or more of these 4 restrictions are not true. Each beta is for one of the stock market variables. The regression on output growth and the F-test result in F(4, 6)= 6.25 and p-value is 0.0248. This means that all the stock market variables together at a 5% level are significantly different from zero and indeed influence output growth. The regression on savings and the F-test result in F(4, 6)= 0.73 and p-value is 0.6011. This means that all the stock market variables together are not significantly different from zero at even a 10% level, suggesting that savings are not influenced by all stock market variables together. The same is done, but then without volatility, because, due to the lack of the data on volatility, the observations were drastically decreased. Without the variable volatility the observations increased from 16 to 42. In this case, H0: β1 =β2 = β3 =0 and H1: one or more of these 3 restrictions are not true. The regression on output growth and the F-test result in F(3, 33)= 5.66 and p-value is 0.0030. This means that all the stock market variables together without volatility are significantly different from zero at a 1% level. This suggests that the three stock market variables influence output growth. The regression on output growth and the F-test result in F(3, 33)= 5.59 and p-value is 0.0033. This means that all the stock market variables together without volatility are significantly different from zero at a 1% level. This suggests savings are influenced by the three stock market variables. The table with the coefficient of these four regressions can be found in Appendix 2 and 3.

(23)

5. Conclusion

Economic growth has many determinants; one of them is financial development. Stock market development belongs to financial development. This thesis studied the relationship between stock market development and long-term economic growth. There is a discussion in relevant literature on the importance of the stock market for economic growth. Within financial development, the banking system is more important for economic growth than the stock market. This thesis reexamined the relationship between the stock market and economic growth, but also looked into the differences in the strength of this relationship between the developed and developing countries. This can be useful knowledge for both the developed and developing countries in the ways they can promote economic growth by emphasizing specific features of the stock market.

The theory shows that there are different channels through which the stock market influences long-term economic growth, such as the liquidity function, risk diversification, acquisition of information about firms, corporate control, savings mobilization, and volatility. Among these channels, only volatility negatively influences economic growth. The

overviewed literature suggests that there are no differences in how the channels work in the developed and developing countries.

Our results convincingly show a significant relationship between economic growth and the stock market through multiple channels. The channels turnover, volatility, and value traded, when the price-effect is controlled for, show a significant relationship with output growth. The liquidity variables turnover and value traded have a positive relationship with output growth. This means more liquid stock markets promote output growth. Volatility has a negative relationship with output growth. The channels turnover and value traded, also when the price-effect is controlled for, have a positive and significant relationship with savings. This shows that there is a clear relationship between liquidity and savings. There is some evidence that the strength of the relationship with the stock market is different for the developed and developing countries. A difference is only found in the relationship between turnover and economic growth. The relationship is stronger for the developing countries. This shows that liquidity has a greater contribution to economic growth in the developing countries as compared to the developed countries. Among the control variables, one relationship with economic growth stands out: the negative relationship between initial output and output growth. This means that the countries with a higher GDP in 1994 have a lower growth rate. This makes sense, because, by definition, the developed countries have lower growth rates.

(24)

The F-test showed that all stock market variables together have a significant relationship with output growth. This means that, overall, the stock market does influence output growth. Between the stock market variables together and savings only, a significant relationship can be found when volatility is not included.

Further research is needed in order to study the differences in the strength of the relationship between the stock market and economic growth in the developed and developing countries. This can be useful knowledge for the developing countries, because the stock market influences economic growth. Knowledge on this issue could help the economic development of these countries.

(25)

6. Appendices

Appendix 1. Summary of statistics

Appendix 2. Regression results for the F-test Dependent variables Independent

variables

Size .0033676 (.0148212) -.0780034 (.1457252) Turnover .0085928 (.0096348) .0103919 (.0796156)

(26)

Value traded .000903 (.0240773) .0958255 (.243681) Volatility -.2711271 ** (.081006) .3080565 (1.315464) Enrollment .7143629 (1.065372) 1.084348 (9.447088) Initial output -1.333133 (.4246387) -2.495243 (4.08122) Government .0388083 (.0269011) -.4925807 (.3753281) Trade -.0003824 ** (.0044014) .1171402 * (.0531758) Inflation -.0526057 (.0438736) -1.406529 ** (.3867887) R-squared 0.8876 0.9005 Observations 16 16

Appendix 3. Regression results for the F-test (Volatility excluded) Dependent variables

Independent variables

Size -.0054927 (.0044219) .0079441 (.0359942) Turnover .0184478 ** (.0072294) .0976735 ** (.0386887) Value traded .0072003 (.0104754) .0405909 (.0715827) Enrollment 1.663657 (1.26268) -3.535065 (6.239022) Initial output -1.01309 *** 1.828329

(27)

(.2820225) (2.108347) Government -.0289785 (.0575537) -.3703343 (.3167708) Trade .0041321 (.005421) .0470401 (.0292336) Inflation -.0154711 ** (.0068405) -.0510688 (.0330103) R-squared 0.5982 0.4301 Observations 42 42

(28)

7. Bibliography

- Agénor P. R., & Aizenman J. (2004). Savings and the Terms of Trade under the Borrowing Constraints, Journal of international economics, 63, pp. 321-340. Retrieved from

http://ac.els-cdn.com/S0022199603000692/1-s2.0-S0022199603000692-

main.pdf?_tid=4ec4c856-0d1b-11e5-b003-00000aab0f6b&acdnat=1433684844_24239d67833ca6417c3bea4104222aee - Arestis P., Demetriades P. O., & Luintel K. B. (2001). Financial Development and

Economic Growth: The Role of Stock Markets. Journal of money, credit and banking, 33 (1), pp. 16-41. Retrieved from

http://www.jstor.org/stable/pdfplus/2673870.pdf?acceptTC=true

- Barro R. J. (2013). Inflation and Economic growth. Annals of economics and finance, 14 (1), pp. 85-109. Retrieved from http://aeconf.com/Articles/May2013/aef140105.pdf - Bassanini A., & Scarpetta S. (2001). The Driving Forces of Economic growth: Panel

Data evidence for the OECD countries. OECD economic studies, 33 (2), pp. 10-53. Retrieved from http://www.oecd.org/economy/growth/18450995.pdf

- Bencivenga V. R., & Smith, B. D. (1991). Financial intermediation and Endogenous Growth. Review of economic studies, 58 (2), pp. 195-209. Retrieved from

http://www.jstor.org/stable/pdfplus/2297964.pdf

- Bencivenga V. R., Smith, B. D., & Starr, R. M. (1996). Equity Markets, Transaction Costs, and Capital Accumulation: An Illustration. World bank economic review, 10 (2), pp 241-265. Retrieved from

http://wber.oxfordjournals.org/content/10/2/241.full.pdf

- Bhide A. (1993). The Hidden Cost of Stock Market Liquidity. Journal of financial economics, 34, p. 31-51. Retrieved from http://ac.els-cdn.com/0304405X9390039E/1-

s2.0-0304405X9390039E-main.pdf?_tid=42c91d9e-a953-11e4-b529-00000aab0f27&acdnat=1422713759_dbba176964f0a9369b083a9d59bab91c - Caporale G. M., Howells P. G. A., & Soliman A. M. (2004). Stock Market

Development and Economic Growth: the Causal Linkage. Journal of economic development, 29 (1), pp. 33-50. Retrieved from http://www.jed.or.kr/full-text/29-1/02_J665_.PDF

(29)

- Chen N., Roll R., & Ross S. A. (1986). Economic forces and the stock market. The journal of business, 59 (3), pp 383-403. Retrieved from

http://www.jstor.org/stable/pdfplus/2352710.pdf?acceptTC=true&jpdConfirm=true - Devereux M. B., & Smith G. W. (1994). International risk sharing and economic

growth. International economic review, 35 (3), pp. 535-550. Retrieved from http://www.jstor.org/stable/pdfplus/2527072.pdf

- Fama E. F. (1970). Efficient Capital Markets: a review of theory and empirical work. The journal of finance, 25 (2), pp. 383-417. Retrieved from

http://onlinelibrary.wiley.com/store/10.1111/j.1540-6261.1970.tb00518.x/asset/j.1540-6261.1970.tb00518.x.pdf?v=1&t=i473sis6&s=c2cce7b8763902ee64b622d8994a611c e8a9bc3a

- Gartner M. (2009). Macroeconomics (3th ed.). Essex: Pearson Education Limited. - Grossman S. J., & Stiglitz J. E. (1980). On the impossibility of informationally

efficient markets. The American review, 70 (3), pp. 393-408. Retrieved from http://www.jstor.org/stable/pdfplus/1805228.pdf

- Jensen M. C., & Murphy K. J. (1990). Performance pay and top-management incentives. Journal of political economy 98 (2), pp. 225-264. Retrieved from http://www.jstor.org/stable/pdfplus/2937665.pdf

- King, R. G., & Levine, R. (1993). Finance, Entrepreneurship, and Growth. Journal of monetary economics, 32, pp. 513-542. Retrieved from

http://ac.els- cdn.com/030439329390028E/1-s2.0-030439329390028E-main.pdf?_tid=4d2891ea-

a7f8-11e4-993f-00000aacb362&acdnat=1422564741_b6e99c0515cd27fb9fb784912bbac701

- Levine, R., & Zervos, S. (1996). The stock market development and long term growth. World bank economic review, 10 (2), pp 329-339. Retrieved from

http://wber.oxfordjournals.org/content/10/2/323.full.pdf+html

- Obstfeld, M. (1992). Risk-taking, Global diversification, and Growth. NBER Working Paper No. 4093, pp. 1-39. Retrieved from http://www.nber.org/papers/w4093.pdf - Scharfstein, D. (1988). The Disciplinary role of Takovers. The review of economic

studies, 55 (2), pp. 185-199. Retrieved from

http://cms.kdis.edu.cn/cms/economics_whu/achievements/cases/resource/69721c91e9 5274a003762741b6344ad8.pdf

(30)

- Singh A. (1991). The stock market and economic development: should developing countries encourage stock markets? Retrieved from

http://mpra.ub.uni-muenchen.de/54927/1/MPRA_paper_54927.pdf

- Stiglitz J. E. (1985). Credit Markets and the Control of Capital. Journal of Money, Credit and Banking, 17 (2), p. 133-152. Retrieved from

http://www.jstor.org/stable/pdfplus/1992329.pdf

- Stock J. H., & Watson M. M. (2012). Introduction to Econometrics (3th ed.). Essex: Pearson Education Limited.

- The World Bank. (2015). Country and Lending Groups. Retrieved from http://data.Worldbank.org/about/country-and-lending-groups

- Von Ungern-Sternberg T. (1981). Inflation and Savings: International Evidence on Inflation-Induced Income Losses. The economic journal, 91 (364), pp. 961-976. Retrieved from http://www.jstor.org/stable/pdfplus/2232502.pdf