A project Undertaken as Part of bsc (Hons) Economics with Finance degree, Westminster International University in Tashkent

Table 2. Unit root test results

Table 2. Unit root test results	Augmented Dickey-Fuller (ADF)		Philips-Perron (PP)
Variables	At level	First difference	At level	First difference
new_businesses	-2.501 *	-3.584*	-2.342	-5.086**
gdp	-5.798**	-4.980**	-5.779 **	-11.820**
tf_ratio	-3.486*	-3.803**	-3.463*	-5.187**
inflation	-8.015**	-3.946**	-7.379 **	-11.268**
rate	-7.735**	-3.836**	-7.459 **	-18.694**
Roi	-1.293	-1.282	-1.134	-5.375
Notes: ** rejects null hypothesis (H0=variable has a unit root) at the level of 5 percent significance level. model with a constant and trend

From the table 2 we can see that GDP inflation and interest rate variables are stationary with constant and trend at level, while number registered businesses, total financing ratio and return on investment have unit root according to both Augmented Dickey-Fuller and Philips-Perron tests. On the other hand taking first difference from the variables makes TF ratio becomes stationary. RGDP_growth, inflation and FV are stationary at level both with constant, and with constant and trend.
Skewness-Kurtosis test

Variable	Observations	Skewness	Kurtosis	adj chi2 (2)	Prob>chi2
Myresiduals	26	0.8496	0.1597	2.21	0.3309

This test show if the data is normally distributed. Normally distribution is the crucial parameter of the data in order to regress OLS. The Null hypothesis of normality test is that data is normally distributed and alternative hypothesis is non-normal distribution. In order to find out the normality of the data, we have to do Skewness-Kurtosis test. The table above shows the results of this test, where three probability values are presented. According to Skewness and Kurtosis tests probabilities we fail to reject the null hypothesis and hence assume that our data is normally distributed. However, the test results contain more accurate and joint value for these tests. chi2 probability of 0.33 is greater than 5%, based on this value also we fail to reject null hypothesis. Consequently, we assume that data is normally distributed.

Correlation matrix
The table illustrates the correlation of all variables. For instance, the correlation between regbus and lnroi is 0.7833. the correlation of the variables with their own is always equal to one because it describes itself perfectly, but you can see that with other variables it is different. For example dgdp and lnCPI are -0.1315 correlated. Consequently, we do not face multi-colinearity problems. You can see all the correlations from the table below.

	regbus	lnroi	dgdp	lnCPI	drate	tf_ratio
regbus	1.0000
lnroi	0.7833	1.0000
dgdp	-0.0132	0.1344	1.0000
lnCPI	-0.3967	-0.3487	-0.1351	1.0000
drate	-0.0556	-0.1257	-0.2254	-0.3270	1.0000
tf_ratio	0.5399	0.3985	0.0164	-0.3194	0.0570	1.0000

Granger causality test
A variable tf_ratio (Islamic banking) is said to Granger-cause a variable regbus (motivation) if given past values of tf_ratio and past values of regbus, it is possible to predict regbus. The following table represents Granger causality test results.

Equation Excluded	Chi2 df prob>chi2
regbus lnroi regbus dgdp regbus lnCPI regbus drate regbus tf_ratio regbus ALL	.93851 2 0.625 1.2622 2 0.532 5.2259 2 0.073 1.6816 2 0.431 3.3254 2 0.190 11.977 10 0.287
lnroi regbus lnroi dgdp lnroi lnCPI lnroi drate lnroi tf_ratio lnroi ALL	20.542 2 0.000 4.8402 2 0.089 22.712 2 0.000 14.133 2 0.001 26.838 2 0.000 91.065 10 0.000
dgdp regbus dgdp lnroi dgdp lnCPI dgdp drate dgdp tf_ratio dgdp ALL	14.795 2 0.001 2.4308 2 0.297 6.843 2 0.033 .73768 2 0.692 13.23 2 0.001 43.741 10 0.000
lnCPI regbus lnCPI lnroi lnCPI dgdp lnCPI drate lnCPI tf_ratio lnCPI ALL	.64352 2 0.752 3.9771 2 0.137 .06624 2 0.967 4.3449 2 0.114 1.7635 2 0.414 21.35 10 0.019
drate regbus drate lnroi drate dgdp drate lnCPI drate tf_ratio drate ALL	.43748 2 0.804 1.9994 2 0.368 8.8717 2 0.012 12.394 2 0.002 1.3648 2 0.505 29.194 10 0.001
tf_ratio regbus tf_ratio lnroi tf_ratio dgdp tf_ratio lnCPI tf_ratio drate tf_ratio ALL	.48269 2 0.786 1.6913 2 0.429 6.1409 2 0.046 1.6932 2 0.429 3.5151 2 0.172 17.199 10 0.070

The null hypothesis for this test is one variable does not granger causes the other variable. We focus on the table of regbus variable results only for this particular case; however, the table shows granger causality test results for all variables. According to the results, probability of chi2 does not fall into 5 percent confidence level. Therefore, the null hypothesis that independent variables granger cause dependent variable cannot be rejected.

Conclusion

Regression results

After conducting tests and analyzing the data graphs and descriptive statistics we adopt our regression model for purpose of accurate results.
Modified regression model:
regbus_t= β₀ + β₁ lnreturn_on_investment_t + β₂ dgdp_growth_t+ β₃ lnCPI_t + β₄ dbase_rate_t + β₅TF_ratio_t + ε
In the next step, OLS model is used to set the model estimation and specification. It illustrates the relationship amongst Islamic banking and entrepreneur motivations. The most important specification is to include correct variables in the model. The regression analysis is carried out using STATA software. You can see the results in the table below.

Variables	Coefficient	t value	p>[t]
dgdp	9.650401**	4.53	0.000
TF ratio	.173059**	2.5	0.0307
lnCPI	-0.646324	-0.87	0.392
lnROI	.0688852	2.23	0.0469
drate	-3.056949	-1.78	0.090
_cons	23.03119	4.41	0.000
No Obs: 26 Prob > F = 0.0001 R-squared = 0.6988 Adjusted R-squared = 0.6235 * p<0.05; ** p<0.01

According to the regression results, the choosen model explains changes in depended variable for 62 per cent. The rest of 38 per cent change come from unexplained variables which are not included in the model. Probability of F value implies that all coefficients jointly are not equal to zero. Before exploring the relationship of variables, we analyze and discuss signs of coefficients. First differential of GDP growth has strong positive effect on business people to start business. The coefficient is statistically significant at 5 percent confidence level. The same is true for TF ratio and lnROI. All these variables have positive slope and are statistically significant 95 percent of times. Variables with negative coefficient are logarithm of inflation and first difference of interest rate. However, they are not statistically significant according to t test. The outcome in terms of relationship direction, corresponds to general expectations and findings of another researchers. For example, economics growth in the country measured by GDP growth has strong positive effect on entrepreneurial motivation. When the economy of the country is growing most people start thinking to open businesses. This is rational finding. On the other hand when price level also increases people may wait until prices adjust and they can take less risk. This is shown by negative slope for lnCPI variable. Similar impact is observed with interest rate variable. Since interest rate illustrates the cost of the borrowing financial resources the higher cost that is interest rate the less people will go into business. Return on investment has vice versa effect. The more return on investment is the higher motivation to start business. Speaking about volume of the affect all of the variables have different effect in absolute value. For example, 1 percent change in GDP growth results 9.66 unit increase in startups’ growth. This figure is 170 thousand and 70 thousand for TF ratio and return on investment variables correspondingly. Meanwhile 1 percent change in inflation results slowing down in business opening for 660 000. But this is true only in 40 cases out of 100. When interest rate rises for 1 % new businesses slow down opening for 69 000.

All in all, Islamic banking highlights the importance of the profit and loss-sharing finance, which can lead to the development of the economy. Risk sharing and promoting enterprises are two main features in Islamic banking accordingly small and medium businesses require both encouragement to risk sharing and entrepreneurship. Small or extremely talented enterprises may suffer from imperfections in the financial market. For instance, transaction costs and information asymmetries. In addition, without inclusive financial system, small enterprises have to take all the costs on their own and invest from their own wealth or internal resources in order to become an entrepreneur. Islamic Banking in Malaysia has shown an incredible growth over the past decades. After all the studies and researches from different directions the main aim was to put some light on tools of Islamic financing system that can be the reason to accelerate the entrepreneurship in Malaysia. By using Islamic financial system, the country can bring the confidence, feasible and easy way of financing for the startups and small businesses, which can change the picture of Malaysian market. Compared to traditional banks the Islamic Banking is on its early stages, by cooperating with relevant national and international institutions The Islamic bank in Malaysia can contribute not only to the country’s welfare but also to the whole world.

Limitations and room for the future researches

Choosing the right topic to conduct the research is the main aim of the writer thus; it took some time to choose an effective and interesting topic. Time management was also very crucial while conducting this research.. In this section of the paper limitations of the research have been provided besides, some recommendations have been made for future researches. The research has been done with high rate of accuracy as much as possible in order to make it more reliable and valid. Despite of that there are some ways that could have improved the research.
Firstly, the data limitation was the main obstacle that we have faced. The number of observation could have been larger if we had collected data for more years. Comparisons of Islamic Banking in different countries, which use Islamic Banking such as Iran and Indonesia could have make this research more complete. Secondly, we could use different method of estimation other than OLS to make this paper more accurate. For instance, implementing simultaneous equations and using 3OLS technique could give better results.

Resources

Academic publications were studied in order to complete and write this paper. Emerald Insight and Science Direct were the main domain and sources for the following paper. Annually time series data were used and the domain is WorldBank.

Appendix

Do file commands

clear all
use "C:\Users\thispc\Downloads\Telegram Desktop\diplom.dta"
tsset year
gen roi=gross_net_saving/gross_investment
gen dtf_ratio=tf_ratio-tf_ratio[_n-1]
dfuller dtf_ratio, trend regress lags(0)
dfuller lnCPI, trend regress lags(0)
histogram new_businesses
histogram lnCPI
gen lnCPI=ln(inflation)
gen drate=rate-rate[_n-1]
histogram drate
dfuller drate, trend regress lags(0)
pperron new_businesses, trend
pperron gdp, trend
pperron tf_ratio, trend
pperron inflation, trend
pperron rate, trend
pperron roi, trend
gen dbus=new_businesses-new_businesses[_n-1]
gen dtf=tf_ratio-tf_ratio[_n-1]
gen dCPI=inflation-inflation[_n-1]
gen droi=roi-roi[_n-1]
pperron dbus, trend regress
pperron dtf, trend regress
pperron dCPI, trend regress
pperron droi, trend regress
pperron dgdp, trend regress
pperron drate, trend regress
gen lnbus=ln(new_businesses)
gen motivation_growth=lnbus-lnbus[_n-1]
reg new_businesses lnroi dgdp lnCPI drate tf_ratio
sktest myresiduals
cor regbus lnroi dgdp lnCPI drate tf_ratio
var regbus lnroi dgdp lnCPI drate tf_ratio
vargranger

1^ http://econ.worldbank.org/research/entrepreneurship