Since the test statistic for the variables in their level form do not exceed the critical value at * level of significance therefore the null hypothesis of a unit root is accepted, indicating that the variables of macroeconomic variables and stock prices are non-stationary.
According to Granger, a non-stationary series must be differenced to achieve the stationary series. Stationarity of time series is of cardinal importance because the analysis of non-stationary variables produces spurious results.
It can be seen from the table that the first difference form of the variables are stationary since the test statistic is greater than the critical values.
The result indicates that all the data were non stationary at levels but become stationary after first differencing. The LNSEDX and other macroeconomic indicators-LM2, LEXC, LTB, LROI are found to be stationary at first difference at 5% significant level and LCPI at 10% significant level. Therefore, all the macroeconomic variables and LNSEDX are regarded to be integrated of order one i.e I(1) in the subsequent tests.
5.1.1Phillips Perron Test
Variables
Phillips-Perron
Test statistics
Level
First
Order of Integration
Difference
LNSEDX
-0.186
-9.15
I(1)
LM2
0.088
-15.422
I(1)
LROI
-1.026
-10.142
I(1)
LCPI
-1.216
-11.131
I(1)
LEXC
-2.628
-6.398
I(1)
LTB
-1.176
-9.385
I(1)
5% Level
-2.888
10% Level
-2.578
Table:5
PP test provides evidence more towards stationary of all the variables in their first difference.
Both results indicate that all the data are non stationary at levels but first differences are stationary at *significant level. We conclude that all the variables are I(1).
As the unit root test results reported in Table 2 satisfy this condition, it is possible to proceed to the 2nd step of the analysis, testing for cointegration to examine whether there exist long run equilibrium between stock price and macroeconomic variables.
5.2 Cointegration
The next step involves estimating the model and determining the rank, r to find the cointegrating relations in our model. Concerning the optimal lagged length, we use lag 2 submitted by AIC,HQIC and FPE criteria.
The trace test is employed to find the number of cointegrating vectors.
Johansen Cointegrating Test
EIGEN VALUES
HO
H1
STATISTIC
5 % CRITICAL VALUE
TRACE
TEST
r=0
r=1
110.632
94.15
0.36664
r<=1
r=2
51.717*
68.52
0.1653
r<=2
r=3
28.409
47.21
0.13491
r<=3
r=4
9.715
29.68
0.03288
r<=4
r=5
5.402
15.41
0.3126
r<=5
r=6
1.305
3.76
0.01007
r<=6
r=7
According to the result from Johansen Maximun likelihood procedure in table 6, the trace test find one cointegrating relationship among the stock price and the five macroeconomic variables. The test suggests cointegrating rank 1.That is, the series are cointegrated . Therefore, we conclude that there is a cointegrating relationship between the stock price and macroeconomic variables.
Since there is one cointegrating rank in the system, the relationship between the variables cannot be applied in a VAR. Therefore a VECM is the most suitable to be used.
Since the trace statistic take into account all of the smallest eigenvalues,it possesses more power than the maximum eigen statistic(Kasa,1992;Serletis and King,1997)
5.2.1 COINTEGRATING VECTOR
The normalized cointegrating relation between the macroeconomic variables and the stock price normalized on LNSEDX is given by:
The cointegration vector
VARIABLES
LNSEDX
LM2
LCPI
LROI
LEXC
LTB
INTERCEPT
VECTOR
1
-5.941
-4.326
8.504
20.47
-7.303
6.456
(NONE)
(-0.639)
(-0.929)
(-1.4932)
(-2.693)
(-1.078)
Yt= (LNSEDX LM2 LROI LEXC LCPI LTB C ) 2
β = 1.000 -5.941 8.504 20.47 -4.326 -7.303 6.456
These values represent long term elasticity measures due to logarithmic transformation of LNSEDX, LM2, LROI, LEXC, LCPI, LTB. Thus the cointegrating relationship can be re-expressed as:
LNSEDXt =5.941 LM2t -8.504 LROIt -20.47 LEXCt + 4.326 LCPI + 7.303 LTBt -6.456
The coefficient for Lm2, Lcpi, Ltb are positive while the signs between Lecx and Lroi are both negative. The intercept term is negative. As the variables are in logarithm form, the coefficient shows the elasticity of the variables. Thus, a rise of 1% in LM2, LCPI AND LTB will result in a 5.9%, 4.3% and 7.3% increase in stock price respectively. Nevertheless, a 1% rise in LEXC and LROI will result in a 20.4% and 8.5% decrease in stock price respectively. As the elasticity of LM2, LCPI, LTB is less than one; we can conclude that they are inelastic.
Money supply has a positive long run effect on semdex and one unit increase in LM2 brings about a 5.914 increase in LNSEDX. Therefore in respect of the long run coefficient of LM2 obtained in the VECM, it can be said that monetary policy measured by LM2 in this dissertation has a positive effect on stock price. The findings are consistent with the findings of Mukerjee and Naka(1995) who examine the relationship between money supply and stock price, found that money supply appear as a significant variable in the long run.
Rate of interest has a negative effect on semdex, a 1% increase in LROI result in a 8.504 % decrease in stock price. The negative relationship LROI and LSEDX is expected because treasury bills acts as the rate of return offered by the risk free asset. The shifting of funds between risky equity and risk free assets by portfolio managers is expected in consequence. When treasury bills rate is high , rational investors would tend to invest in less risky assets with high returns. The Treasury bill positively affects the stock price.
A 1% increase in exchange rate is expected to reduce stock price by 20.47%. As the Mauritian economy depends on imported goods and that the depreciation of the Mauritian rupees makes increases prices of goods which push inflation up. This depreciation is also disincentive to non-resident investors in terms of repatriation of profit and earnings. In the long run, currency depreciation seems to be associated with the stock market decline, as evidenced by negative exchange rate coefficient.
Semdex is positively related to CPI where a 1% rise in Lcpi results in a 4.326% in Lnsedx. With a mild inflation in Mauritius, this positive association between semdex and CPI seem to support the view that stock price are a good hedge against inflation. It is in accordance with the Fisher(1930) hypothesis which implies a positive one-to-one relationship between stock price and inflation. It also implies that investors are compensated for inflationary increase in prices.
It is to be pointed out that LNSEDX and LM2,LROI and LTB and LM2 and LEXC are highly collinear whereas that of the other variables are less collinear.
5.3 GRANGER CAUSALITY
Since the variables are cointegrated in the Johansen technique therefore a VECM is the suitable model used to conduct the Granger causality test.
The table report the chi square statistics obtained when causality was tested in a VECM:
After testing the long run equilibrium for the stock price and macroeconomic variables, we investigate the dynamic interactions between these variables.
The presence of cointegration also rules out non-causality among the variables. In other words, there must be at least a unidirectional causality from one variable to the other.
As seen from the granger causality test it is suggested that there exists no granger causality between stock price and treasury bills and consumer price index as it is insignificant at 10% level of significance, the null of no causality from these variables is rejected but unidirectional causality is observed in the case of money supply, rate of interest, exchange rate and stock price, the statistics are significant at 10% level of significance.
Stock market index is not a leading indicator for the economic variable of the change in interest rate, which shows the evidence of informationally inefficient market. LROI is exogenous. Interest rate cause stock price, it can be explained that as loan rate decrease, companies may take more loan, cost of borrowing will be low, and they will invest which tend to increase stock price. This finding is consistent with Ahmed et al (2007) who found a unidirectional causality of interest rate causing stock price.
As interest increase it will attract more foreign investors in the country. Thus an increase in capital inflow will increase share price as shown in table
As interest rate in economy is determined by monetary policy of the country, policy makers can concentrate on it for adjusting the volume of stock market and overall investment in economy.
On the other hand it can have a negative impact, if interest rate increases, people will invest in banks. This will cause a shift away from semdex to banks, hence stock price will fall.
Semdex Granger causes money supply implying that all money present in the economy is determined to a great extent by Semdex.LM2 is exogenous. It can be explained as when Semdex increase, people's expectations of prosperous future in economic terms also rise. Due to boom in capital market, people start investing their money in more liquid assets which increases money supply. This is in line with Chakravaty(2007) who found a unidirectional causality between money supply and stock price.
Semdex cause exchange rate. This case supports the portfolio balance approach. As Mauritius is a well developed financial market, it has an interaction with the foreign exchange market. This result indicate that stock price are the indicator to exchange rate for Mauritian case, as stock price are found to Granger-cause exchange rates, exchange rate conditions may be strengthen via improving the stock market's fundamentals. The result is consistent with the findings of Ai-Yee Ooi (2009) in Thailand who found a unidirectional causality between stock price and exchange rate.
The explanation of semdex causing exchange rate can be due to foreign direct investment to Mauritius during the ten years as shown below:
The exchange has enabled a net inflow of Rs 3.71 billion during the period 2000-2009. It has benefitted the country with foreign exchange reserves and local companies with foreign capital. Foreign investments have undoubtedly had a positive effect on market growth and efficiency. An increase in capital flows will appreciate the exchange rate thus this rise in demand will push up the stock market level. The appreciation of the rupees will bring good effect for stock price. The foreign investor lead to appreciation of the currency, they lead to the exports industry becoming uncompetitive due to the appreciation of the rupees. Thus excess fund inflow can also make a negative impact on the economy of the industry.
Another dynamics of the causality are given by the impulse response function and the forecast error variance decomposition.
5.4 Impulse Response Functions
The IRF traces out the impact if shocks in Macroeconomic variables on the Semdex. In a VAR if the system of equations is stable any shock should decline zero whereas in a VECM it converge to a non zero value. As the variables are found to be cointegrated in a long run relationship it is conducted in a VECM, the effect of a unit shock in one variable will have a permanent effect on the other variable. Moreover, when the impact of the shock is completely absorbed by the response variable, the correlation will become stable and the IRF will balance at a specific value.
As it has three unidirectional causality, we will use the causality to continue with IRF.
Figure 4 shows the response of lm2 to a lnsedx shock.
The IRF graph traces out the effect of lm2 following a shock in lnsedx. Lnsedx has a negative causal influence on lm2 as shown by the impulse response, there is volatility till about the 8th horizon and the IRF converges to -0.054 as from10 years.
The negative impact of a shock of lnsedx on the lm2 can be explained by the following factors: the money supply in Mauritius is influenced mainly by foreign investors. If the interest rate is high relative to other countries , the foreign investors are likely to leave their money in the bank rather than invest in the risky stock market. If the interest rate is low the investors might prefer to invest in other markets. Hence the shock of lnsedx on the lm2 results in a negative impact.
The table illustrates the IRF function to a disturbance in rate of interest. A positive impact is observed on rate of interest and as time increase, the shock on lnsedx is permanent.
There is thus a positive causal relationship from lroi to lnsedx and the shock stabilizes at 0.314 in 11 years.
As rate of interest has a positive innovation, the cost of capital cause stock price to go down. Since the stock price is the present value of all the future cash flows, stock price consequently goes down.
The exchange rate responds negatively at all horizon to standard shock given to the equation for the stock price. There is volatility in the exchange rate till about tenth horizon. Then a shock to the equation for the stock price produces a stable impact on the exchange rate at -0.0058 in 13years.
The negative effect of lnsedx on Rs/dollar is viewed interms of the portfolio balance approach. Increasing lnsedx attract investors who sell their dollar in exchange for Rs. Increasing demand for Rs will result in its appreciation which indicates that less Rs can be bought per dollar and hence the exchange rate Rs/dollar will fall. This analyzes the negative relationship between lecx and lnsedx.
Nevertheless, to hold the portfolio balance approach, it is important that investors invest in Mauritius. It is to be pointed out that following the abolition of exchange rate controls in 1994, investments by foreigners was made possible. Source: The Stock Exchange of Mauritius Ltd
The impulse response analysis has brought ample awareness for the causality of the variables between the macroeconomic variables and semdex. Further approach of these relationships are obtained by the forecast error variance decomposition .
5.5 VARIANCE DECOMPOSITION
Having deciding the sign of the causality in the impulse response analysis, the variance decomposition analysis is used to determine the relationship of stock price to macroeconomic variables.
The table below presents the percentage of forecast variance in lnsedx in a VECM. The values have been tabulated for the period 1, 5, 10, 15, and 20-step ahead forecast.
VARIANCE DECOMPOSITON FOR STOCK PRICE
In view of the table, the following observations can be made:
The reported figures indicate the percentage of movement in the ith variable that can be attributed to its own shock and the shocks to the other variables in the system. The result tends to support the argument that the movements in the lnsedx can be explained by some of the macroeconomic analyzed.
In the first month ,100% of the variability in the lnsedx is explained by its own shocks that is at horizon one all the variance in lnsedx is explained by itself while after five months 56.43 (79.39%) of the variability is explained by its own innovations-this indicate that at longer horizon ,variance in the lnsedx is caused by the variance in other variables, in other words there are causal relationship among the lnsedx and the other variables at longer horizons,0.74 (0.31%) by the shocks of lm2, 13.03 (4.23%) by lroi,0.0541 ( 0.0324%) by lcpi, 0.102 (0.187%) by lecx, 28.63(15.862%) by ltb. The lm2, lcpi and lexc plays little role in explaining the variance of the lnsedx. The foregoing discussion indicates that the most influential macroeconomic determinants of the stock prices in Mauritius are ltb and lroi.
The total contribution of the shocks, of each variable at each step, including the shock of lnsedx, should be equal to 100% which has been calculated and it all adds up to 100%.
5.6 DISCUSSION OF RESULTS
A rise in money supply has a positive effect on the stock market which is in line with Sellin(2001) which explain that if money increase ,means demand will increase which will lead to a rise in economic activity, hence higher cash flow and an increase in stock price. Growth in money supply appears to exert a positive impact on stock price.
Positive and significant long run relationship between increases in money supply in Malaysia may be explained by the findings of Maysami and Koh (2000) who find money supply has a direct positive liquidity impact on the stock market. Moreover, Mukherjee and Naka (1995) pointed out that injection of money supply leads to boost in corporate earnings.
Money supply and stock price in Mauritius are positively correlated and this is also consistent with the findings for the US and Japan. The increase in money supply has a direct positive liquity effect on the stock market.
Another explanation follows from Fama's (1981) explanation for inflation. Increases in real activity which drive stock returns also stimulates the demand for money via the simple quantity theory thus creating the positive relation between money supply and stock price which is spurious rather than causal.
As for the expected sign we have got a reverse sign for inflation where we should have got a negative relation.The positive relationship between LCPI and LNSEDX is associated with the mild inflation rate in Mauritius. The Fisherian relation between nominal interest rates and expected inflation leads us to expect that one of the reasons people hold various assets is to hedge against inflation and so stock returns should be positively correlated with inflation. However, according to the Fisher effect, it indicates a positive one for one relationship between stock price and inflation. It means that investors are compensated for inflationary increase in prices.
The positive relationship between LCPI and LNSEDX is inconsistent with the findings of Fama(1981), Modigliani and Cohn (1979), Fieldstein(1980), Mukeerjee and Naka's(1995) findings for Japan and Fama and Schwert's(1977) and Chen, Roll and Ross'(1986) findings for the US where they all found a negative relationship between inflation and stock price.
The relationship between stock price and exchange rate is negative as hypothesized and is in line with the findings of Aggarwal (1981), Smith(1992), Solnik(1987). As for the exchange rate it shows currency depreciation seem to be associated with a fall in stock price. In fact, currency effect may have a positive or a negative with stock price depending on the nature of the economy. For net exporting economies , currency depreciation leads to an increase in net exports as domestic products becomes cheaper in the world market. Hence the increase in firm's profitability will be reflected in the value of the stocks. However, for economies that depend heavily on imports, currency depreciation may lead to higher import prices causing a fall in firm's profit and in turn price of stocks.
The negative relationship between stock price and exchange rate is in accordance with the portfolio balance models of exchange rate determination, it postulates a negative relationship stock price and exchange rate and the causation runs from stock price to exchange rate.
A stronger currency drives the Mauritian market higher, as such an appreciation of the currency would make exports less competitive thus lowering export earnings. As Mauritian has a strong import and export content , a strong Mauritian rupees lowers the cost of imported inputs, thus allowing local producers to be more competitive internationally.
As for the expected sign we have got a reverse sign for treasury bills. The positive relationship between treasury bills and stock price implies that as the yield on treasury bills increase, there is a rise in stock price. This is in line with Ratanapakorn and Sharma (2007) which reported positive relationship between S&P 500 and Treasury bill rate in US.
The negative relationship between rate of interest and stock price can be explained as follows:
A stock's required rate of return is made up of two parts: the risk-free rate and the risk premium. As the government adjusts key interest rates, the risk-free rate will change. If the government raises rates, the risk-free rate will rise also. If nothing else changes, the stock's target price should drop because the required return is higher. The reverse is true. If interest rates fall, then the stock's target price should rise because the required return has dropped.
The interest rate is negative which is consistent with Mukherjee and Naka's(1995) findings for Japan, he suggest that this is because interest rates serves as a better proxy for the nominal risk-free component used in the discount rate in stock valuation models.
An increase in the interest rate raises the opportunity costs of holding cash and likely to lead to a substitution effect between stocks and other interest instruments.
