Research proposal writer

CHAPTER FOUR

FINDINGS, PRESENTATIONS AND DISCUSSIONS

4.0 Introduction

This chapter presents the data analyzed from secondary data sources on the determinants of coffee exports in Uganda between 1991 and 2010. The data was tabulated to give a meaningful presentation and interpretation. Presentation and interpretation were based on the specific objectives to address the research problem.

This section further reports the estimates for Uganda’s coffee export function. In order to detect the long-run co-movement among the variables, the co integration procedure developed by Johansen (1991) and Juselius (1990) was employed. An error correlation model for the determinants of coffee exports was used.

4.1 Tests

The researcher had to first run some tests before analyzing his data and these included the following:

4.1.1 Test for normality

Figure 2: The histogram-normality test

A regression was run and on clicking on the view-residual test-histogram-normality test, the histogram is bell-shaped, suggesting a normal curve shape, and the jarque-bera statistics has high p-value of 0.628520 indicating that the errors in the regression are normal that is to say; the jarque-bera statistics probability of 0.628520 is greater than zero and it has a percentage of 63% greater than 10%(63%>10%) thus the errors in the regression are normal.

4.1.2 Test for the presence of heteroskedasticity

Figure 3:Residual plot

The residual plot clearly shows that the variance of the error term was not constant hence there is heteroskedasticity. The researcher therefore transformed the model by introducing logs so as to remove heteroskedasticity. This was done in Eviews and the logs for the dependent and independent variables were obtained.

4.1.3 Test for the constancy of parameters

Ho: There is perfect parameter stability

H_A: There is no perfect parameter stability

Figure 4:Recursive estimates to test for parameter stability

The above is a graph of recursive estimates to test for parameter stability

From the graph above, the null is not rejected since the CUSUM of squares plot strays within the band hence concluding that there is perfect parameter stability.

4.1.4 Test for omitted variables

Table 1: Chow Breakpoint Test

The null hypothesis is rejected and we conclude that there is a structural break inthe data.

4.2 Basic exploratory data analysis

This involved establishing the basic descriptive statistics and the correlation matrix. The descriptive statistics of all the variables in logarithms are displayed in table 4.2, while the correlation matrix in table 4.3 demonstrates the relationship between quantity of coffee exported (LQs) and the other variables used in the study.

Table 2: Descriptive statistics of the series, sample period 1991-2010

The Jarque-Bera tests the hypothesis that the series is normal. Since the probability value for price(LOGPRICE), exchange rate(LOGER),dummy for quality(DUMMYQ), and quantity of coffee exported (LOGQ) is greater than five percent significant level, the null cannot be rejected meaning the series is normal.

Table 3: Correlation matrix

From table 4.3, the correlation coefficient 0.217899 means quantity of coffee (LOGQ) is positively and weakly correlated with price (LOGPRICE). On the other hand the correlation coefficient -0.099234 implies that quantity of coffee (LOGQ) is negatively and weakly correlated with exchange rate (LOGER) and the correlation coefficient -0.763450 implies that quantity of coffee (LOGQ) is negatively and strongly correlated with quality of coffee (DUMMYQ).The correlation coefficients – 0.07130 and -0.131278 implies that Price (LOGPRICE) is negatively and weakly correlated with exchange rate (LOGER) and quality of coffee (DUMMYQ) respectively. Finally, the correlation coefficient 0.084663 implies that there is a weak positive relationship between exchange rate (LOGER) and quality of coffee (DUMMYQ)

4.2.1 Unit root test results

Unit root tests were carried out using the augmented Dickey-Fuller test statistic. This was carried out to check whether the series were stationary (integrated) or not. This is because standard inference procedures do not apply to regressions which contain an integrated dependent variable or integrated regressors. The test statistic tested the null hypothesis that the time series has a unit root against the alternative that there is no unit root. The test statistic values are compared to the critical values at five percent significant level. The test statistic values less than the critical values at five percent level of significance indicate that the series are non-stationary otherwise they are stationary.

Table 4: Unit root tests of the series 1991-2010

Note: lag length for the Augmented Dickey-Fuller test statistic

In the table 4, the quantity of coffee (LOGQ), Exchange rate (LOGER) and Price (LOGPRICE) are stationary in the levels and after the first difference since there ADF statistic are greater than the critical values.

4.2.2Cointegration tests

Among the variables that are integrated of order 1(1), an attempt was made to check whether Cointegration holds. The purpose of the Cointegration tests was to determine whether a linear combination of a group of non-stationary series is stationary. Engle and Granger (1987) pointed out that a linear combination of two or more non- stationary series may be stationary. The linear combination of quantity of coffee, real exchange rate, and prices was checked to find out whether the residuals were stationary. The static equation, whose residuals were modeled, tested for stationarity and thereafter formed the error correction term after the first lagging as presented in table 4.5

Table 5: Cointegration tests output

The next attempt involved testing the residuals for the order of integration. The application of the Augmented Dickey Fuller test statistic revealed that the residuals are stationary in levels (table 4). This confirmed that the linear combination of quality of coffee, price, and exchange rate formation is indeed stationary.

4.3 Multivariate analysis

Table 6: Regression analysis

Dependent variable: LOGQUANTITY

The results indicate that the p-value(0.8636) is greater than 0.05 thus we accept the null and conclude that exchange rate has a significant influence on coffee exports. The p-value (0.4719) results into accepting the null hypothesis since it is greater than 0.05 and hence we conclude that prices have a significant influence on coffee exports. The p-value(0.0003) is less than 0.05 the level of significance, we reject the null hypothesis and conclude that quality of coffee is statistically insignificant. That is it has no significant influence on coffee exports.

The results indicate that the t-statistic │-0.174567│ is less than 2 hence do not reject the null hypothesis thus concluding that exchange rate has a significant influence on coffee exports.

The results also indicate that the t-statistic │0.736831│ is less than 2 hence do not reject the null hypothesis thus concluding that price has a significant influence on coffee exports.

The results further indicate that the t-statistic │-4.647847│ is greater than 2 hence the null hypothesis is rejected thus concluding that quality of coffee is statistically insignificant.

Basing on R-squared (0.597713), the model is a fairly good fit. This implies that 59.7% of the variations in quantity of coffee exported is explained by exchange rate, prices. The remaining 40.3% is explained by other variables.

The Durbin-Watson value (1.563469) implies that the time series data is free from autocorrelation since it is approximately 2 (1.5634692)

The coefficient of real exchange rate (-0.018595) implies that a one percent increase in exchange rate is likely to cause a decline in the volume of coffee exports by about 10%( in the short run). The coefficient of real effective exchange rate is negative and statistically significant both in the short run and in the long run. This agrees with the findings of Frauk and Yavuz (2007) that the real exchange rate is statistically significant and negative.

Looking at the coefficient of price (0.050706), results indicate that it is positive and statistically significant at 5% level. This means that a one percent increase in price of coffee leads to an increase in volume of coffee exports by about 0.05%. The results conform to that of Oyejide(1986) who contend that high and attractive prices are an incentive to producers and exporters to increase the volume of agricultural exports in Nigeria and Jebuni,et al.(1991) who found out that the elasticity of international price is positive and significant.

The coefficient of quality of coffee (-0.296218) is negative and statistically insignificant and agrees with the theory of apriori. This implies that a one percent increase in quality of coffee is likely to cause about 0.296 percent decline in the volume of coffee exports.

Substituted Coefficients:

=====================

LOGQUANTITY = 15.1445113 + 0.0507055676*LOGPRICE – 0.01859476562*LOGEXCHANGERATE -0.2962180921*DUMMYFORQUALITY

This implies that a unit increase in the log of price by 1% on average increases the quantity of coffee exported by approximately 0.0507055676 keeping other factors constant.

A unit increase in the log of exchange rate by 1% on average reduces the quantity of coffee exported by approximately 0.01859476562 keeping other factors constant.

Also a unit increase in the log of dummy for quality by 1% on average reduces the quantity of coffee exported by approximately 0.2962180921 keeping other factors constant.

CHAPTER FIVE

SUMMARY, CONCLUSIONS AND POLICY RECOMMENDATIONS

5.1 Introduction

The chapter presents the discussion of findings, conclusions and recommendations for solutions to be undertaken as a result of the study and the areas seemed important for further study. The discussions, conclusions and recommendations are presented in reference to the study objectives.

5.2 Summary of findings

Based on the econometric analysis, the results show that depreciation of the exchange rate reduces coffee export volumes.

Increase in coffee price leads to a great increase in coffee export volumes. This is because increase in price acts as an incentive to producers and exporters to increase production and exports of coffee.

A decrease in the quality of coffee leads to a great depreciation of coffee exports.

The model was a fairly good fit for the data as shown by the coefficient of determination. Some tests were run on the data and it was found that the data was normal.

There was autocorrelation and heteroskedasticity and thus the model as transformed by introducing logs to remove heteroskedasticity. The parameters were also found to be perfectly stable.

5.3 Conclusions

In conclusion, the results from the study indicated that exchange rate had a significant effect on the quantity of coffee exported. An increase in the volume of coffee exports can therefore be attained by reducing the exchange rate since quantity of coffee exported has an inverse relationship with exchange rate.

Since its inception as a commercial crop, coffee has dominated and continues to dominate as the most single important cash/export crop. Its contribution stands at around 20% of total exports, but is threatened by a number of issues; Lack of a national coffee policy, Lack of farmer ownership in the coffee value chain, Lack of a national law that addresses all industry issues from research, production, marketing, processing and exporting, Lack of a law on coffee research and its funding, Deteriorating funding for research, Domination of coffee exports by poor grade coffees, Absence of an industry specific advisory services to farmers, Public and private institutions in the industry are not streamlined on their respective responsibilities and coordination of such responsibilities.

The empirical results based on Cointegration analysis show that the coffee export volumes have a long-run relationship with exchange rate, coffee prices and quality ofcoffee.

From the results, it can be concluded that real effective exchange rate depreciation leads to a big reduction in the coffee export volumes while an increase in coffee price greatly increases coffee export volumes. On the other hand, the decrease in coffee quality leads to areduction in the coffee export volumes. In order to increase the coffee export earnings, the government should embark on exporting high quality coffee which can fetch high prices on the international market.

5.4 Policy recommendations

The findings of the study led to the following policy recommendations necessary to ensure steady and sustainable increase in the coffee export volumes. The policy proposals are as follows:

In view of the statistical significance of coffee price, the exporters should initiate the establishment of agreements with international coffee buyers. This will help in increasing coffee prices thereby encouraging coffee production and increase in coffee export volumes. Also, the exchange rate should not be allowed to depreciate to avoid reduction in coffee export volumes.

The government should sensitize people on better farming methods, avail them with fertilizers, disease resistant coffee tress and better farming tools to help in improving coffee bean quality before export.

For the government to reap from its coffee exports, it should regulate its exports on grounds of quality. Once quality coffee is put on market, it fetches good prices whereas poor quality coffee may spoil the country’s marketability in regards to coffee output.

Open policies on export activities that is to say; free trade and free foreign exchange regimes to maximize economies of scale.

There should be stronger oversight mechanisms by MAAIF and MOFPED with regard to how UCDA deploys its revenues so that the bulk of the resources are spent on developing the sector,

Review the regulatory framework for support institutions such as NARS/COREC, NAADS to make them responsive to the research and extension needs of the sector. Research should be demand driven.

5.5 Areas for further study

The government needs to embark on more research within these areas to help in granting more opportunities to get ideas and insight on quality improvement among others.

Effect of trade liberalization on coffee exports in Uganda.

The impact of the government fiscal policy on coffee exports in Uganda.

The researcher only looked at the coffee export supply function. The coffee demand function also need be tackled and then the equilibrium position established by use of simultaneous equations. It is therefore suggested that a future study focuses on this kind of analysis so as to drive improved results.