**4.1 Results**

A stationery time series is defined as a series with constant mean, constant variance and constant auto-variance for each lag. A unit root test examines whether the time series variables are nonstationary. If there are unit roots (nonstationary) in the series, a series of successive differences can change it into a stationary one. **Figures 1** and **2** depict the exchange rate processes. This is to show how nonstationary and stationary processes look like.

After differencing, there exists a stationary linear combination of nonstationary random variables; the variables combined are said to be cointegrated. When two or more variables move together for a long period of time, we say they are cointegrated. The ADF test was used to investigate the order of integration. This test is important in determining whether cointegration methods can be applied to study the long-run relationship between the variables. The variables in this study are integrated to the same order. In order to examine the impact of oil price and the exchange rate on FDI, the VAR model was used. A selection is done using a maximum of four lags.

**Figure 1.** *Exchange rate.*

**Figure 2.**

*Differenced exchange rate.*


#### **Table 2.**

*Pair-wise correlation results.*

Correlation in statistics is defined as the relationship between variables. It is explained by numbers ranging between �1 and 1, called correlation coefficients. The pare-wise correlation is shown in **Table 2**, which shows that FDI was correlated with all the variables and that there was no specific variable that was correlated with all the variables. With this information, we can conclude that there is a less chance of multicollinearity being a problem.

According to the trace and maximum eigenvalue test for cointegration, the null hypothesis for at least one cointegration equation was accepted, concluding that there is one cointegration vector and that there is one significant long-run relationship between variables. From both the Johansen's trace and maximum eigenvalue tests, the results showed that the variables (EXRATE, OIL and FDI) could either have a short- or long-run relationship.

If the variables are cointegrated, the residuals from estimated regression can be used to estimate error correction model so that the short- and long-run effects of the variables can be analysed; VECM was used to analyse these results. **Tables 3** and **4** depict these analyses.

The long-run relationship is shown in the equation below:

$$FDI = 888851.1 - 1491.34 \text{EXRATE} - 10923.00 \text{OIL} \dots \tag{2}$$

**Table 3.**

**21**

*Vector error correction estimates.*

*Trade and Investment in South Africa*

*DOI: http://dx.doi.org/10.5772/intechopen.87186*

The equation above shows that EXRATE and OIL have a negative long-run relationship with FDI, which can be interpreted by saying that a unit increase in EXRATE and OIL results in a decrease in FDI.


**Table 3.**

*Vector error correction estimates.*

Correlation in statistics is defined as the relationship between variables. It is explained by numbers ranging between �1 and 1, called correlation coefficients. The pare-wise correlation is shown in **Table 2**, which shows that FDI was correlated with all the variables and that there was no specific variable that was correlated with all the variables. With this information, we can conclude that there is a less chance

FDI 1 0.808648 0.718984 EXRATE 0.808648 1 0.378294 OIL 0.718984 0.378294 1

**FDI EXRATE OIL**

According to the trace and maximum eigenvalue test for cointegration, the null hypothesis for at least one cointegration equation was accepted, concluding that there is one cointegration vector and that there is one significant long-run relationship between variables. From both the Johansen's trace and maximum eigenvalue tests, the results showed that the variables (EXRATE, OIL and FDI) could either

If the variables are cointegrated, the residuals from estimated regression can be used to estimate error correction model so that the short- and long-run effects of the variables can be analysed; VECM was used to analyse these results. **Tables 3** and **4**

The equation above shows that EXRATE and OIL have a negative long-run relationship with FDI, which can be interpreted by saying that a unit increase in

*FDI* ¼ 888851*:*1 � 1491*:*34*EXRATE* � 10923*:*00*OIL*… (2)

The long-run relationship is shown in the equation below:

of multicollinearity being a problem.

have a short- or long-run relationship.

EXRATE and OIL results in a decrease in FDI.

depict these analyses.

**20**

**Figure 2.**

**Table 2.**

*Differenced exchange rate.*

*Regional Development in Africa*

*Pair-wise correlation results.*


For stationarity tests, the ADF and PP test were used. The PP test showed that variables used are I(1). The pair-wise correlation showed that FDI was correlated with all the variables and that there were no specific variables that were correlated with all the variables, concluding that there is less chance of multicollinearity. The vector error correction estimates showed that the variables ð Þ *EXRATE* and *OIL* have a negative long-run relationship with FDI. All explanatory variables were statistically significant in explaining FDI since they have absolute *t*-values above 2. According to the trace and maximum eigenvalue test for cointegration, the null hypothesis for at least one cointegration equation was accepted, concluding that there is one cointegration vector and that there is one significant long-run relation-

The equation above shows that EXCHANGE and OIL have a negative long-run relationship with FDI, which can be interpreted by saying that a unit increase in

The results show that exchange rate is an important variable in determining FDI inflows into South Africa as it was highly correlated with FDI. Among other things, this suggests that South Africa needs to reduce inflation rates because if inflation is lower than competing countries, South Africa's goods will become attractive and the demand will rise. Also, the country needs to increase interest rates as this would attract hot money flow, which occurs when banks and financial institutions move money to other countries to get a better rate of return on savings. The empirical results also suggested that oil is another important factor in determining FDI inflows as it was also highly correlated with FDI. The study showed that the country should continue focusing on policies aimed at strengthening its exchange rate and

What is pertinent to the picture painted above is that economic growth in the country will continue to be stifled if the urgent need to enhance economic growth is not taken into consideration. We have to come up with the methods on how to save

Department of Decision Sciences, University of South Africa, Pretoria, South Africa

© 2019 The Author(s). Licensee IntechOpen. This chapter is distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/ by/3.0), which permits unrestricted use, distribution, and reproduction in any medium,

ship between variables.

*Trade and Investment in South Africa*

*DOI: http://dx.doi.org/10.5772/intechopen.87186*

stabilizing oil prices.

our economy.

**Author details**

**23**

Garebangwe Victoria Mabe-Madisa

provided the original work is properly cited.

\*Address all correspondence to: mabemgv@unisa.ac.za

The long-run relationship is shown in the Eq. (2).

EXRATE and OIL results in a decrease in FDI.

**5. Conclusion and recommendations**

#### **Table 4.** *Vector error correction model coefficients.*

The vector error correction equation is given as

$$\begin{aligned} \Delta FDI &= -0.0225 \times (FDI\_{t-1} - 1491.3433 \times EXRATE\_{t-1} - 10923.0039 \times OLL\_{t-1} \\ &+ \\$88851.0584) + 0.8878 \times \Delta FDI\_{t-1} + 1.0943 \times \Delta EXRATE\_{t-1} \\ &+ 208.6206 \times \Delta OIL\_{t-1} + 2744.589 \end{aligned} \tag{3}$$

where ð*FDIt*�<sup>1</sup> � 1491*:*3433 � *EXRATEt*�<sup>1</sup> � 10923*:*0039 � *OILt*�<sup>1</sup>þ 888851*:*0584Þ is the error correction term.

## **4.2 Summary**

To reiterate, South Africa's appeal has improved; however, in contrast with other developing countries, South Africa has not been attracting FDI. The question at hand was "what method can South Africa use to draw more FDI?" This is a very important question, in light of the fact that FDI can play a major role in the continent's development. FDI can fortify household speculation, encourage innovation exchange, make business, advance fares and, most importantly, produce financial growth. The part of FDI as a wellspring of wealth is especially critical in South Africa, in a setting where net official development assistance to the country declines.

This section analysed FDI and its determinants. The descriptive statistics showed that the mean values of the variables are all positive, ranging from 50.762 to 783,438.3, with oil price being the lowest and FDI being the highest. This can be interpreted by saying that South Africa has experienced more FDI inflows over the years understudy.

#### *Trade and Investment in South Africa DOI: http://dx.doi.org/10.5772/intechopen.87186*

For stationarity tests, the ADF and PP test were used. The PP test showed that variables used are I(1). The pair-wise correlation showed that FDI was correlated with all the variables and that there were no specific variables that were correlated with all the variables, concluding that there is less chance of multicollinearity. The vector error correction estimates showed that the variables ð Þ *EXRATE* and *OIL* have a negative long-run relationship with FDI. All explanatory variables were statistically significant in explaining FDI since they have absolute *t*-values above 2.

According to the trace and maximum eigenvalue test for cointegration, the null hypothesis for at least one cointegration equation was accepted, concluding that there is one cointegration vector and that there is one significant long-run relationship between variables.

The long-run relationship is shown in the Eq. (2).

The equation above shows that EXCHANGE and OIL have a negative long-run relationship with FDI, which can be interpreted by saying that a unit increase in EXRATE and OIL results in a decrease in FDI.
