It is widely accepted that a well-functioning financial system is crucial for the development of the economy. As such, it becomes crucial to investigate what the determinants of stock market development are. This would enable policy makers to take the necessary steps to enhance stock market development, which will in turn trigger a much-needed economic growth. This paper therefore aims at identifying the main macroeconomic determinants of stock market by using a dynamic Panel Vector Error Correction Model within a sample of Sub Saharan African countries. The results suggest that economic growth, banking development, stock market liquidity, investment and macroeconomic stability are key determinants of stock market development in the region. Interestingly, the study finds that savings ha ve a significant and detrimental impact on the growth of equity markets in the region. Moreover, results also indicate that economic growth indirectly stimulates stock market development in the short run.
The importance of financial development on growth has been stressed since very long back [
The determinants of stock market development can be broadly classified into two categories: macroeconomic determinants and institutional determinants. Yartey [
Despite the importance of identifying the determinants of stock market development, there is a surprisingly little number of theoretical and empirical work that has been carried out to answer the question, especially in developing countries. In an attempt to shed more light on the matter, this paper makes use of a dynamic Panel Vector Error Correction Model (PVECM) to find out what the potential determinants of stock market development are within a set of Sub Saharan African countries during the period 1989-2016. Although both institutional and macroeconomic determinants are important drivers of stock market development, this study follows Garcia and Liu [
In order to determine what the macroeconomic determinants of stock market development are in a set of 14 Sub-Saharan African countries under consideration, annual time series data spanning over a period of 28 years (1989-2016) are used. These 14 countries from the Sub-Saharan African region include Botswana, Cote d’Ivoire, Ghana, Kenya, Malawi, Mauritius, Namibia, South Africa, Swaziland, Nigeria, Tanzania, Uganda, Zimbabwe and Zambia. The countries were selected mostly based on availability of data. This section describes the model adopted and the empirical indicators of the variables of interest used so as to model the impact of macroeconomic factors on stock market development.
Drawing from models adopted by Gracia and Liu [
M C R i t = β 0 + β 1 G D P i t + β 2 T V T S R i t + β 3 D C T P S i t + β 4 G F C F i t + β 5 G D S g d s i t + β 6 I N F i t + ε i t (1)
The paper focuses on the determinants of stock market capitalization ratio (MCR), which is defined as the value of listed shares in the stock exchange divided by GDP. Drawing from Gracia and Liu [
The logic behind the inclusion of explanatory variables is discussed below:
Economic growth is often recognized as being a key determinant of equity growth. Indeed, according to the demand following hypothesis, the demand for financial services such as stock markets is amplified through an expansion of the economy. La Porta et al. [
Following Garcia and Liu [
Moreover, stock markets channel savings to investment projects. As such, a higher level of savings implies a higher level of capital flows through the stock markets. Drawing from Gracia and Liu [
Yet another potential determinant of stock market development is liquidity. Liquid stock markets facilitate long term, risky, and potentially more lucrative investments, which in turn improve capital allocation and long term growth. As such, the higher the liquidity of the stock markets, the higher the amount invested through them and the higher the amount channeled to investment projects. In other words, a liquid stock market will boost stock market development. In this study, stock market liquidity is captured through total value traded shares ratio, TVTSR, (which is the total value of shares traded on a country's stock exchanges expressed as a percentage of GDP). The same proxy has also been used by Gracia and Liu [
Finally, a high volatility in the economic environment is expected to act as a deterrent for savers participating in the stock markets. Moreover, fiscal, monetary and exchange rate policy changes, especially if unexpected, also negatively impact corporate profitability as pointed out by Gracia and Liu [
The time series data is extracted from the World Development Indicator database of the World Bank. The specification used in this model is a double log linear one for ease of interpretation, with the value of the estimates to be discussed in percentage terms. By taking log on all variables on both sides of the equation, Equation (1) from above results in the following:
m c r i t = β 0 + β 1 g d p i t + β 2 t v t s r i t + β 3 d c t p s i t + β 4 g f c f i t + β 5 g d s i t + β 6 i n f i t + ε i t (2)
where i denotes the different countries in the sample, t denotes the time dimension and E ε is the white noise disturbance term. The small letters above denote the natural logarithm of the variables.
(From here onwards, the small letters denote the natural logarithm of the variables).
mcr | gdp | dctps | gfcf | gds | inf | tvtsr | |
---|---|---|---|---|---|---|---|
Mean | 2.875603 | 23.18013 | 2.957132 | 2.471197 | 2.473652 | 2.313639 | −0.241574 |
Median | 2.949541 | 22.90693 | 2.765186 | 2.545703 | 2.728064 | 2.190492 | −0.267170 |
Maximum | 5.776586 | 27.06627 | 5.075953 | 3.338344 | 3.895823 | 10.10279 | 4.916278 |
Minimum | −4.679185 | 20.36217 | 1.129337 | 0.850424 | −1.648623 | −2.676491 | −7.793365 |
Std. Dev. | 1.582391 | 1.337051 | 0.879682 | 0.463912 | 0.924003 | 1.107060 | 2.263081 |
Skewness | −1.162012 | 0.917442 | 0.533176 | −1.012236 | −1.534631 | 1.122417 | −0.191992 |
Kurtosis | 7.230912 | 3.545697 | 2.830290 | 4.351985 | 5.898708 | 11.28452 | 3.761406 |
Jarque-Bera | 320.232 | 1110.3 | 8615.0 | 1.4801 | 5428.2 | 88.787 | 725.57 |
Probability | 0.1520 | 0.2560 | 0. 2525 | 0.4770 | 0.4313 | 0.7629 | 0.5196 |
Sum | 557.8671 | 9086.612 | 1108.924 | 711.7047 | 843.5152 | 858.3602 | −49.52274 |
Sum Sq. Dev. | 483.2648 | 69−8.9933 | 289.4161 | 61.76642 | 290.2858 | 453.4651 | 1044.793 |
Observations | 194 | 392 | 375 | 288 | 341 | 371 | 205 |
The mean and the median of all the variables are relatively close together, indicating that the variables are potentially normally distributed. An analysis of the skewness of the variables confirms that all the series are almost symmetrical (except market capitalization ratio, investment, saving, and inflation whose values are closer to 1), with their skewness values revolving around 0. When the kurtosis of the data is analyzed, it can be seen that again, most of the series have values relatively close to 3 (except from market capitalization, savings and inflation), which mean that most of the distributions are believed to be rather normal. The Jarque-Bera tests conducted also confirm that the series follow a normal distribution.
Before proceeding with the estimation of the model to identify the significant determinants of stock market development in the sample of Sub-Saharan African countries under consideration, a few preliminary tests are essential. First, it is important to determine whether the time series under investigation are stationary. To this end, panel unit root tests are used to find the order of integration of the various variables under consideration. ADF-Fisher, PP-Fisher and Levin, as well as Levine, Lin & Chu and Im, Pesaran and Shin W-stat panel unit root tests are applied on the panel series in levels. While Levine, Lin & Chu assume a common unit root, the other unit root tests assume an individual unit root. Overall, the results of all the different unit root tests reject stationarity in favor of a unit root for all the variables1. This implies that the variables are integrated of order one, that is, they are non-stationary in levels but achieve stationarity after being differenced once.
This being the case, an interesting question arises: Is there a long run equilibrium relationship among the underlying variables. In other words, although non-stationary variables may deviate from each other in the short run, economic forces may act in response to the deviations from equilibrium, thus bringing back their association in the long run. This implies that even though each variable is integrated, there exists a linear combination of the variables that is stationary. In this study, both Johansen Fisher Panel Cointegration Test and Kao Residual Cointegration Test are resorted to in order to verify the presence of a long run relationship among the variables. The results confirm that a cointegrating relationship exists among the variables. Thus, having established the presence of a long run relationship, the study opts for a panel vector error correction model, and proceeds with its estimation.
The PVAR is an econometric model that can be viewed as a hybrid of the traditional VAR approach and panel data approach. Panel data VAR thus interestingly combines the traditional VAR approach in a time series, which treats all the variables in the system as endogenous, with the panel data approach, which allows for unobserved individual heterogeneity. If only a panel data framework is used, there is a risk of loss of dynamic information as stock market development may have something to do in explaining itself as well [
The pth order PVECM is specified as follows:
Δ y i t = Π i y i , t − 1 + Γ i 1 Δ y i , t − 1 + ⋯ + Γ i , p − 1 Δ y i , t − p + 1 + u i t (3)
where yit is a vector comprising of all the variables used in the model, i denotes the different countries in the sample, t denotes the time dimension, and uit is a standard white noise process. In this study, an optimal lag length of 2 is chosen based on the Akaike Information Criterion (AIC), Schwarz Information Criterion (SC), and Hannan-Quinn Information Criterion (HIC).
Drawing from Garcia and Liu [
The Long Run Equation
Model 1 | Model 2 | Model 3 | |
---|---|---|---|
Stock Market Development, mcr | 1 | 1 | 1 |
Economic Growth, gdp | 0.14507* [1.65445] | 0.846471*** [2.84109] | 0.02741* [1.79501] |
Banking Development, dctps | 0.37744* [1.65363] | 0.557385* [1.75811] | 0.89635** [2.39218] |
Stock Market Liquidity, tvtsr | 0.45551*** [2.71946] | 1.4171*** [7.15448] | 0.60543*** [3.84115] |
Savings, gds | −1.72887*** [4.86793] | −1.85013*** [−5.71500] | |
Investment, gfcf | 2.14311* [1.69928] | ||
Inflation, inf | −0.94963*** [−3.62064] | ||
C | −2.78045 | −30.0119 | −9.07428 |
***indicates significance at 1% level, **at 5% and ***at 10% respectively. The small letters denotes variables in natural logarithmic and t-statistics are in parentheses.
The first column considers the determinants of stock market development if only economic growth (gdp), banking development (dctps), stock market liquidity (tvtsr) and savings (gds) are included in the regression model. This model is used as the basic regression. Zooming in on the model, it is found that all the variables are significant. Model 1 shows that income level has a significant role in generating gains in terms of stock market development within the sample of African regions under scrutiny. Indeed, a 1% rise in income level causes a 0.145% rise in stock market development. This is in line with previous studies [
On the other hand, to test whether investment rate is also a determinant of stock market development, the second regression includes an investment proxy instead of saving rate. Results are displayed in Model 2 of the table above. All the variables are seen to have positive and significant impacts on stock market development. This is consistent with results of model 1. However, the results indicate that, unlike savings, investment is an important determinant of stock market. Indeed, it has a positive and significant coefficient of 2.14, which implies that a 1% increase in investment would simulate a 2.14% rise in stock market development. This is in line with the results of Gracia and Liu [
Finally, in Model 3 of the
The Short Run Equation.
The table above shows the short run coefficient estimates of Model 1, along with the t statistics in brackets. From Model 1, we find that the only short run determinant of stock market development is the second lag of economic growth: It has a positive and significant coefficient of 0.635. Results support the fact that it takes time for the impact of the determinants to be felt. Interestingly, a bidirectional causality is detected between stock market development and economic growth. Indeed, both lags of market capitalization ratio are also seen to promote economic growth, with a positive and significant coefficient of about 0.05 in both cases. Additionally, an indirect determinant of stock market development can be extracted from the short run results of Model 1. In fact, stock market liquidity indirectly promotes stock market development through economic growth. Indeed, if last year’s stock market liquidity rises by 1%, then, this leads to a 0.033% rise in economic growth. As mentioned previously, increasing economic growth by 1% would cause stock market development to shoot up by 0.635%. Combining these two results together, a 1% rise in liquidity leads to a 0.033*0.635 percentage point increase in stock market development, through economic growth.
Error Correction | D(mcr) | D(gdp) | D(dctps) | D(tvtsr) | D(gds) |
---|---|---|---|---|---|
Coint Equation (1) | 0.090704*** | −0.009454 | 0.018275 | 0.076892 | −0.17242*** |
[2.54608] | [−0.81558] | [1.57525] | [1.27238] | [−5.69064] | |
D (mcr(−1)) | −0.029332 | 0.058502* | 0.008927 | 0.231839 | 0.27202*** |
[−0.29848] | [1.82964] | [0.27894] | [1.39075] | [3.25458] | |
D (mcr(−2)) | −0.012779 | 0.050093* | 0.021313 | −0.029424 | 0.23068*** |
[−0.13980] | [1.68438] | [0.71605] | [−0.18977] | [2.96731] | |
D (gdp(−1)) | −0.064038 | 0.211972** | 0.29610*** | 0.367151 | −0.71001*** |
[−0.22429] | [2.28173] | [3.18462] | [0.75805] | [−2.92374] | |
D (gdp(−2)) | 0.634815** | −0.009958 | −0.038754 | 0.005886 | −0.378442 |
[2.18681] | [−0.10543] | [−0.40996] | [0.01195] | [−1.53278] | |
D (dctps(−1)) | −0.347401 | 0.054504 | 0.228257** | 0.108716 | −0.473367* |
[−1.15712] | [0.55795] | [2.33468] | [0.21347] | [−1.85379] | |
D (dctps(−2)) | −0.026354 | 0.014583 | −0.101075 | −0.729681 | −0.032979 |
[−0.09210] | [0.15664] | [−1.08477] | [−1.50336] | [−0.13552] | |
D (tvtsr(−1)) | −0.040333 | 0.032932** | 0.029620** | −0.091239 | −0.18131*** |
[−0.93737] | [2.35228] | [2.11395] | [−1.25003] | [−4.95418] | |
D (tvtsr(−2)) | 0.034453 | 0.001922 | 0.015166 | −0.22019*** | −0.12830*** |
[0.80346] | [0.13778] | [1.08610] | [−3.02711] | [−3.51772] | |
D (gds(−1)) | 0.337085 | 0.002201 | 0.009062 | 0.186828 | −0.42289*** |
[1.17366] | [0.06369] | [0.26199] | [1.03695] | [−4.68125] | |
D (gds(−2)) | 0.086964 | 0.036158 | −0.024033 | −0.67013*** | 0.055103 |
[0.81226] | [1.03796] | [−0.68932] | [−3.68988] | [0.60512] | |
C | −0.075010 | 0.0609*** | 2.43E−05 | 0.035747 | 0.092381** |
[−1.43977] | [3.59235] | [0.00143] | [0.40448] | [2.08484] |
***indicates significance at 1% level, **at 5% and ***at 10% respectively. The small letters denotes variables in natural logarithmic and t-statistics are in parentheses.
Error Correction | D(mcr) | D(gdp) | D(tvtsr) | D(dctps) | D(gfcf) |
---|---|---|---|---|---|
Coint Equation (1) | 0.000615 | 0.012985 | 0.28990*** | 0.004529 | −0.02401** |
[0.01668] | [1.45181] | [3.99465] | [0.57026] | [−2.11950] | |
D (mcr(−1)) | 0.036991 | 0.053858* | −0.175590 | 0.041663* | 0.066552* |
[0.33301] | [1.99978] | [−0.80351] | [1.74234] | [1.95141] | |
D (mcr(−2)) | 0.199906 | −0.044030 | 0.251514 | −0.008295 | 0.068566* |
[1.54554] | [−1.40403] | [0.98843] | [−0.29791] | [1.72661] | |
D (gdp(−1)) | −0.204474 | 0.391257 | 0.253840 | 0.230526** | −0.233041* |
[−0.46455] | [3.66629] | [0.29315] | [2.43297] | [−1.72448] |
D (gdp(−2)) | 0.582206 | −0.117172 | −0.257797 | −0.030260 | 0.30509** |
---|---|---|---|---|---|
[1.26544] | [−1.05041] | [−0.28482] | [−0.30553] | [2.15988] | |
D (tvtsr(−1)) | −0.076564 | −0.015325 | 0.161283 | 0.018375 | −0.04449*** |
[−1.38942] | [−1.14701] | [1.48774] | [1.54905] | [−2.62988] | |
D (tvtsr(−2)) | −0.035132 | 0.015447 | −0.067288 | 0.015113 | −0.027082 |
[−0.64733] | [1.17388] | [−0.63022] | [1.29357] | [−1.62530] | |
D (dctps(−1)) | 0.954952* | 0.051083 | −1.311182 | 0.007705 | 0.032698 |
[1.88682] | [0.41629] | [−1.31686] | [0.07072] | [0.21042] | |
D (dctps(−2)) | −0.353492 | −0.286275** | −0.514425 | 0.125461 | −0.091890 |
[−0.73322] | [−2.44910] | [−0.54238] | [1.20889] | [−0.62080] | |
D (gfcf(−1)) | −0.146210 | 0.152304* | −0.157167 | −0.110330 | −0.112532 |
[−0.40522] | [1.74097] | [−0.22141] | [−1.42046] | [−1.01583] | |
D (gfcf(−2)) | −0.166917 | 0.186164** | −0.886642 | 0.023852 | 0.014269 |
[−0.45453] | [2.09088] | [−1.22728] | [0.30173] | [0.12655] | |
C | −0.027235 | 0.05587*** | 0.045162 | 0.015005 | 0.008015 |
[−0.41859] | [3.54164] | [0.35282] | [1.07133] | [0.40122] |
***indicates significance at 1% level, **at 5% and ***at 10% respectively. The small letters denotes variables in natural logarithmic and t-statistics are in parentheses.
of a bidirectional relationship between stock market development and banking development. Additionally, this model also reveals that economic growth is an indirect determinant of stock market development. Indeed, economic growth indirectly promotes stock market development through banking development. Results indicate that a 1% rise in economic growth triggers a 0.231*0.955 percentage point rise in stock market development.
As for Model 3 (
Error Correction | D(mcr) | D(gdp) | D(tvtsr) | D(dctps) | D(gds) | D(inf) |
---|---|---|---|---|---|---|
Coint Equation (1) | 0.062443* | −0.007074 | 0.078121 | 0.012620 | −0.1627*** | −0.1265*** |
[1.78912] | [−0.62431] | [1.38096] | [1.08520] | [−5.75975] | [−3.31911] | |
D (mcr(−1)) | −0.005381 | 0.061006* | 0.278592* | 0.027661 | 0.23738*** | 0.24874** |
[−0.05251] | [1.83387] | [1.67734] | [0.81014] | [2.86159] | [2.22250] | |
D (mcr(−2)) | −0.000739 | −0.040809 | −0.054508 | 0.030906 | 0.25355*** | 0.202185* |
[−0.00743] | [−1.26371] | [−0.33807] | [0.93247] | [3.14866] | [1.86097] | |
D (gdp(−1)) | −0.213811 | 0.182298 | 0.009763 | 0.29308** | −0.75853** | −1.0896*** |
[−0.68320] | [1.79434] | [0.01925] | [2.81068] | [−2.99410] | [−3.18790] | |
D (gdp(−2)) | 0.563822* | 0.011569 | 0.283398 | −0.049400 | −0.409551 | 0.610573* |
[1.80916] | [0.11435] | [0.56103] | [−0.47573] | [−1.62337] | [1.79380] | |
D (tvtsr(−1)) | −0.062063 | 0.03156** | −0.060431 | 0.03210** | −0.1962*** | −0.048727 |
[−1.36156] | [2.13278] | [−0.81795] | [2.11368] | [−5.31632] | [−0.97876] | |
D (tvtsr(−2)) | 0.043302 | 0.008118 | −0.1892*** | 0.017901 | −0.1494*** | 0.013369 |
[0.95367] | [0.55074] | [−2.57111] | [1.18326] | [−4.06361] | [0.26958] | |
D (dctps(−1)) | −0.343173 | 0.025284 | −0.100835 | 0.217634** | −0.52990** | −0.237086 |
[−1.04;993] | [0.23828] | [−0.19033] | [1.99835] | [−2.00269] | [−0.66413] | |
D (dctps(−2)) | −0.071762 | 0.019505 | −0.528772 | −0.117655 | −0.040677 | 0.511571 |
[−0.23530] | [0.19700] | [−1.06967] | [−1.15781] | [−0.16476] | [1.53579] | |
D (gds(−1)) | 0.395145 | 0.000532 | 0.251967 | 0.014099 | −0.4611*** | 0.201967* |
[1.56033] | [0.01478] | [1.40074] | [0.38128] | [−5.13224] | [1.66625] | |
D (gds(−2)) | 0.139004 | 0.026662 | −0.7080*** | −0.030902 | 0.068200 | 0.150497 |
[1.24911] | [0.73802] | [−3.92489] | [−0.83342] | [0.75707] | [1.23823] | |
D (inf(−1)) | −0.031659 | −0.002954 | −0.102816 | −0.015198 | 0.140274** | −0.3968*** |
[−0.41928] | [−0.12049] | [−0.84009] | [−0.60408] | [2.29488] | [−4.81165] | |
D (inf(−2)) | −0.036271 | 0.011733 | −0.008054 | −0.019086 | 0.084621 | −0.3241*** |
[−0.53643] | [0.53450] | [−0.07348] | [−0.84717] | [1.54597] | [−4.38933] | |
C | −0.045728 | 0.06220*** | 0.070154 | −0.000563 | 0.105339** | −0.018167 |
[−0.81139] | [3.39997] | [0.76798] | [−0.02998] | [2.30890] | [−0.29514] |
***indicates significance at 1% level, **at 5% and ***at 10% respectively. The small letters denotes variables in natural logarithmic and t-statistics are in parentheses.
This study tries to investigate what the potential macroeconomic determinants of stock market development are in a sample of Sub Saharan African countries for the years 1989-2016. The empirical analysis depicts five interesting findings. First, the Panel VECM indicates that the main long run drivers of stock market development in the region are economic growth, banking development, stock market liquidity, and investment. Secondly, the study supports the belief that banking development actually complements stock market development in the long run. Thirdly, the study finds that savings has a significant and detrimental impact on stock market development. This unexpected link might be due to the fact that a rise in savings leads to a fall in economic growth, which in turn translates into a fall in stock market development due to the latter’s bi-directional link with economic growth. Fourthly, in the short run, only economic growth and banking development are seen to be significant determinants of stock market development. Moreover, the study also reveals a bidirectional relationship between stock market development and economic growth, as well as between stock market development and banking development in the short run. Finally, in the short run, indirect determinants are also detected. Indeed, stock market liquidity and economic growth are two different channels through which stock market development is triggered indirectly.
The study also unleashes interesting ideas to promote equity growth in the region. Firstly, the positive impact of stock market liquidity on equity market growth that has been detected in the African region suggests that measures have to be taken for the markets in the African region to become more liquid. Unfortunately, African stock markets are notoriously known for having a low market liquidity compared to international norms. There are several avenues that can be explored to spark off the much-needed liquidity. It is believed that the introduction of a wider range of products in the different market segments should be encouraged. Moreover, most foreign investors have been reluctant to invest in the region for fear of the operational risk that relates to the lack of standardization. It is as such important to standardize products. Additionally, the restrictive limits on short selling, which are poorly regarded in Africa, should be eliminated. Indeed, such a step can promote liquidity if it is properly managed and effectively conducted. More retail investors must also be attracted into the market. It is additionally recommended that the number of counters in the exchanges be increased. Unfortunately most counters in Africa are now dominated by listings of brewery, banks and telecom, as opposed to natural resources which are much sought after. Furthermore, the benefits of listing privately held companies should be more aggressively marketed.
Moreover, it is also important to initiate policies to foster economic development and investment in the region so that they can both reach that strategic point where they can significantly direct growth towards equity development. Such measures would in turn fuel economic growth in the region. Additionally, measures should also be taken to promote the banking sector as the development of the latter will simultaneously enhance the development of stock markets in the region. Although competitiveness and innovation have both improved in the region, more effort has to be made to promote financial depth and penetration, which are still low. Moreover, major impediments to financial inclusion have to be curbed to further boost banking development.
On the other hand, domestic savings was seen to be detrimental to equity growth. To switch this relationship, a better promotion to retail investors about the benefits and returns of investing in stock markets is required. Besides, steps must also be taken to stabilize the political environment. Such measures would decrease the inherent volatility in the stock markets, and as such, the latter will become a more appealing alternative to investing in banks. As for the saving barriers in Africa, they have to be destroyed so that savings can be channeled productively into stock markets and not shallow financial systems.
Matadeen, S.J. (2017) The Macroeconomic Determinants of Stock Market Development from an African Perspective. Theoretical Economics Letters, 7, 1950-1964. https://doi.org/10.4236/tel.2017.77132