4.1 Summary Statistics
Table 1 indicates the descriptive statistics of variables used in the estimate analyses. Firstly, the real exchange rate movement (logRate) has a mean of 1.746 over the sampled period, which reveals the level of volatility with a minimum value of -2.61 and 9.828. Similarly, Table 1 reports the summary statistics of the underlying variables adopted for sanctions imposed by the United States, European Union, United Nations with economic sanctions, and the level of intensity with maximum and minimum values of 0 and 1 (refer to section 2.2.2). The mean and standard deviation of the sanctions imposed by the United States, European Union, United Nations with economic sanctions and intensity remain low, representing similar characteristics among the Sub-Saharan African countries.
Table 1: Descriptive Statistics
Variable
|
Description
|
N
|
Mean
|
Std.Dev.
|
Min
|
Max
|
|
|
|
|
|
|
|
logRate
|
The real effective exchange rate index is the nominal index adjusted for relative changes in consumer prices
|
627
|
1.746
|
1.291
|
-2.61
|
9.828
|
US
|
United States initiated economic sanctions
|
630
|
.273
|
.446
|
0
|
1
|
E.U.
|
European Union initiated economic sanctions
|
630
|
.211
|
.408
|
0
|
1
|
U.N.
|
United Nations initiated economic sanctions
|
630
|
.116
|
.32
|
0
|
1
|
Economic
|
Sanctions impact on the economy of the target economy
|
630
|
.376
|
.485
|
0
|
1
|
Intensity
|
The formal sanctions intensity
|
630
|
1.395
|
1.785
|
0
|
5
|
Current
|
Current account balance in % of GDP
|
435
|
-4.066
|
12.554
|
-147.997
|
42.227
|
Lending
|
The lending rate
|
367
|
27.236
|
74.462
|
4.737
|
1175
|
Inflation
|
Consumer price index
|
547
|
15.052
|
33.586
|
-16.117
|
448.5
|
GDS
|
Gross domestic saving in % of GDP
|
534
|
12.348
|
23.705
|
-141.974
|
83.287
|
In addition, Table 2 presents the correlation coefficients between variables. The table presents a low correlation between the variables. Thus, there exists no multicollinearity statistical problem in the models.
Table 2: Correlation Matrix
|
|
|
|
|
|
|
|
|
|
|
|
1
|
2
|
3
|
4
|
5
|
6
|
7
|
8
|
9
|
10
|
1.logRate
|
1
|
|
|
|
|
|
|
|
|
|
2.US
|
-0.275***
|
1
|
|
|
|
|
|
|
|
|
3.EU
|
-0.340***
|
0.701***
|
1
|
|
|
|
|
|
|
|
4.UN
|
-0.172*
|
0.244***
|
0.306***
|
1
|
|
|
|
|
|
|
5.Economic
|
-0.146*
|
0.787***
|
0.773***
|
0.506***
|
1
|
|
|
|
|
|
6.Intensity
|
-0.202**
|
0.854***
|
0.723***
|
0.575***
|
0.932***
|
1
|
|
|
|
|
7.Current
|
-0.295***
|
0.0640
|
0.0136
|
0.174*
|
0.0397
|
0.0917
|
1
|
|
|
|
8.Lending
|
0.156*
|
-0.0290
|
-0.0270
|
-0.293***
|
-0.129
|
-0.133
|
-0.261***
|
1
|
|
|
9.Inflation
|
0.0465
|
0.146*
|
0.0758
|
-0.194**
|
0.00596
|
0.0761
|
-0.122
|
0.441***
|
1
|
|
10.GDS
|
-0.124
|
0.116
|
0.197**
|
0.142*
|
0.132
|
0.179**
|
0.624***
|
-0.0888
|
0.0690
|
1
|
* p < 0.05, ** p < 0.01, *** p < 0.001
4.2 Baseline findings
Table 3 discloses the main findings of this study using the fixed effect regression. The information criteria used for the validity of the step-wise fixed effect regression are the F-test and the coefficient of determination. The F-test reflects the combined significance of the models' coefficients, which is significant with a p-value of 0.0000. The R2 shows the models' explanatory power with a minimum of 0.167. Based on the information criteria, the fixed effect regressions are devastatingly valid and reliable for empirical interference.
Table 3: Sanctions Imposed and Real Exchange Rate via Fixed Effect Regression
|
|
|
|
|
|
|
Model 1
|
Model 2
|
Model 3
|
Model 4
|
Model 5
|
|
|
|
|
|
|
Current
|
0.00245
|
-0.00115
|
0.00396
|
0.00197
|
0.00142
|
|
(0.00306)
|
(0.00324)
|
(0.00323)
|
(0.00318)
|
(0.00315)
|
Lending
|
0.0154***
|
0.0153***
|
0.0159***
|
0.0160***
|
0.0159***
|
|
(0.00377)
|
(0.00371)
|
(0.00389)
|
(0.00381)
|
(0.00378)
|
Inflation
|
-0.00805***
|
-0.00864***
|
-0.00947***
|
-0.00897***
|
-0.00824***
|
|
(0.00190)
|
(0.00184)
|
(0.00194)
|
(0.00189)
|
(0.00189)
|
GDS
|
0.0000724
|
-0.000148
|
-0.000201
|
0.00000674
|
0.000168
|
|
(0.00199)
|
(0.00196)
|
(0.00208)
|
(0.00201)
|
(0.00199)
|
|
|
|
|
|
|
U.S.
|
-0.217***
|
|
|
|
|
|
(0.0625)
|
|
|
|
|
E.U.
|
|
-0.337***
|
|
|
|
|
|
(0.0773)
|
|
|
|
U.N.
|
|
|
-0.157
|
|
|
|
|
|
(0.169)
|
|
|
Economic
|
|
|
|
-0.186***
|
|
|
|
|
|
(0.0650)
|
|
Intensity
|
|
|
|
|
-0.0646***
|
|
|
|
|
|
(0.0187)
|
|
|
|
|
|
|
Constant
|
1.643***
|
1.650***
|
1.626***
|
1.640***
|
1.642***
|
|
(0.0742)
|
(0.0730)
|
(0.0767)
|
(0.0750)
|
(0.0742)
|
|
|
|
|
|
|
Observations
|
209
|
209
|
209
|
209
|
209
|
R-squared
|
0.214
|
0.241
|
0.167
|
0.198
|
0.213
|
RMSE
|
0.281
|
0.277
|
0.290
|
0.284
|
0.282
|
F-test
|
10.15
|
11.80
|
7.463
|
9.206
|
10.10
|
Prob > F
|
0.0000
|
0.0000
|
0.0000
|
0.0000
|
0.0000
|
Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1
From Table 3, the dependent variable is proxied by real exchange rate movement, and the interesting explanatory variables include the United States sanctions, European Union sanctions, United Nations sanctions, the economic sanctions, and the level of sanctions intensity. Under Table 3, the model I & 2 show the extent to which the United States and European Union sanctions respectively weaken the real exchange rates, and the results disclose that the coefficients of the United States and European Union sanctions are significantly negative at 1% level of significance. This infers that with the increase of U.S. and E.U. sanctions on Africa, the target region's exchange rates weaken. The results are in accordance with the study of Wang, Wang, & Chang (2019), which asserts that the sanctions are more strenuous and short-term. In the short term, the target countries might not seek countermeasure to reduce the negative influence of the sanctions.
Moreover, Model 3 in Table 3 shows the estimate of U.N. sanctions. The result reports show that the coefficient of U.N. sanctions is insignificant. This infers that irrespective of the number of sanctions by the U.N., there remains frail effects on the exchange rate in the region. The insignificant impact of the U.N. sanctions could be traced to the fact that U.N. rarely imposes sanctions on the target region. The insignificance could be traced to the assertive sanctions imposed by the United Nations in the short-term period.
Under Table 3, Model 4 estimates the economic sanction. It is deduced that the estimate of economic sanction is negatively significant at a 5% level of significant, which shows that economic sanction negatively impacts the real exchange rate of the target region. This result conforms with the economic sanction definition, which implies that any method of sanctions that relate to the region's economy will influence the region's real exchange rate. This conforms with the perception that economic sanctions influence such a country's economy (Dylan, 2017). In addition, Model 5 shows the Sanction Intensity estimate. The coefficient of sanction intensity is negatively significant at a 1% level of significance. It infers that such intensity has a negative impact on the real exchange rate of the target region. This suggests that the real exchange rate is weakened with sanction intensity, which identifies a lesser relationship between two economies in finance, trade, and personal exchange.
In addition, a significant control variable (i.e., inflation) has the expected outcomes. From Table 3, we establish a significantly negative effect of the average rate of price changes in the goods and services index (i.e., inflation rate) on currency performance, which in agreement with the study of Ito & Sato (2008). This result implies that inflation increases the domestic currency decline, which tends to change the trade structure while the exchange rate depreciates alongside. However, the result for lending supports that there is a significantly positive effect on currency performance. Against the conventional agreement that asserts the effect of lending rate on currency depreciation, our result shows otherwise.
As identified in the earlier section, the Quantile regression approach is employed to consider the initial levels of the exchange rate movement. Quantile regression is useful to identify the conditional determinants of estimates. Hence, to investigate how the initial level of the real exchange rate plays out when developed economies impose various sanctions, we adopt Quantile Regression. With Quantile regression, exchange rates are distinguished in terms of weak, static, and strong rates. Table 4-7 show Quantile 0.25, 0.50, Quantile 0.75 and Quantile 0.90 respectively.
Table 4: Sanctions Imposed and Real Exchange Rate via Quantile Regression
|
|
|
|
|
|
Panel A: Quantile 0.25
|
Model 1
|
Model 2
|
Model 3
|
Model 4
|
Model 5
|
|
|
|
|
|
|
U.S.
|
-0.942***
|
|
|
|
|
|
(0.312)
|
|
|
|
|
E.U.
|
|
-0.948***
|
|
|
|
|
|
(0.362)
|
|
|
|
U.N.
|
|
|
-1.003
|
|
|
|
|
|
(0.619)
|
|
|
Economic
|
|
|
|
-0.835**
|
|
|
|
|
|
(0.335)
|
|
Intensity
|
|
|
|
|
-0.237***
|
|
|
|
|
|
(0.0881)
|
|
|
|
|
|
|
Control Variables
|
Yes
|
Yes
|
Yes
|
Yes
|
Yes
|
|
|
|
|
|
|
Constant
|
0.997***
|
0.992***
|
0.953***
|
0.979***
|
0.992***
|
|
(0.243)
|
(0.247)
|
(0.307)
|
(0.284)
|
(0.266)
|
|
|
|
|
|
|
Observations
|
209
|
209
|
209
|
209
|
209
|
Pseudo R-squared
|
0.257
|
0.251
|
0.212
|
0.226
|
0.242
|
Standard errors in parentheses
*** p<0.01, ** p<0.05, * p<0.1
While Table 4 (i.e., the sample that falls within the 25th percentile) reiterates the results of the baseline finding, the results in Table 5 show that U.S., E.U., U.N., Economic and Intensity sanctions are negative and significant at 1% level, which implies the effects of such sanctions contribute significantly on weakening real exchange rate for economies in the 50th percentile of the sample. This suggests that such economies lack the strategies to countermeasure the negative effect of the imposed sanctions.
Table 5: Sanctions Imposed and Real Exchange Rate via Quantile Regression
|
|
|
|
|
|
Panel A: Quantile 0.50
|
Model 1
|
Model 2
|
Model 3
|
Model 4
|
Model 5
|
|
|
|
|
|
|
U.S.
|
-0.979***
|
|
|
|
|
|
(0.202)
|
|
|
|
|
E.U.
|
|
-1.156***
|
|
|
|
|
|
(0.234)
|
|
|
|
U.N.
|
|
|
-1.658***
|
|
|
|
|
|
(0.426)
|
|
|
Economic
|
|
|
|
-0.637***
|
|
|
|
|
|
(0.238)
|
|
Intensity
|
|
|
|
|
-0.225***
|
|
|
|
|
|
(0.0633)
|
|
|
|
|
|
|
Control Variables
|
Yes
|
Yes
|
Yes
|
Yes
|
Yes
|
|
|
|
|
|
|
Constant
|
2.240***
|
2.213***
|
1.956***
|
2.079***
|
2.163***
|
|
(0.158)
|
(0.160)
|
(0.211)
|
(0.202)
|
(0.191)
|
|
|
|
|
|
|
Observations
|
209
|
209
|
209
|
209
|
209
|
Pseudo R-squared
|
0.459
|
0.448
|
0.356
|
0.367
|
0.389
|
Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1
In addition, results in Table 6 reveal that U.S. and E.U. sanctions are negative and significant at a 1% level, which indicates that such sanctions result in the weakened real exchange rate for countries in the 75th percentile of the sample. Finally, for countries in the 90th percentile of the sample, results in Table 7 show that only significant with the negative contribution to the weakened real exchange rate of the target region. The discrepancies in results could be traceable to the strategies implemented by the target country to counterfeit the impact of such sanctions on its economy. The reasons behind such significant influence of E.U. sanctions over other sanctions could be explained by the speculations highlighted by Wang, Wang, & Chang (2019) that compare U.S. and E.U. sanctions' relevance on exchange rate volatility. The authors emphasize that (1) the E.U. sanction tends to be more concentrated as it focuses on five specific instruments of sanction; (2) the approach of E.U. sanction is mostly short term and (3) the proximity of E.U. to the target countries tends to have the stronger influence of the target economy.
Table 6: Sanctions Imposed and Real Exchange Rate via Quantile Regression
|
|
|
|
|
|
Panel A: Quantile 0.75
|
Model 1
|
Model 2
|
Model 3
|
Model 4
|
Model 5
|
|
|
|
|
|
|
U.S.
|
-0.970***
|
|
|
|
|
|
(0.235)
|
|
|
|
|
E.U.
|
|
-1.358***
|
|
|
|
|
|
(0.177)
|
|
|
|
UN
|
|
|
0.188
|
|
|
|
|
|
(0.283)
|
|
|
Economic
|
|
|
|
0.225
|
|
|
|
|
|
(0.163)
|
|
Intensity
|
|
|
|
|
0.0331
|
|
|
|
|
|
(0.0595)
|
|
|
|
|
|
|
Control Variables
|
Yes
|
Yes
|
Yes
|
Yes
|
Yes
|
|
|
|
|
|
|
Constant
|
2.475***
|
2.481***
|
2.432***
|
2.433***
|
2.477***
|
|
(0.183)
|
(0.121)
|
(0.140)
|
(0.138)
|
(0.179)
|
|
|
|
|
|
|
Observations
|
209
|
209
|
209
|
209
|
209
|
Pseudo R-squared
|
0.341
|
0.514
|
0.464
|
0.465
|
0.359
|
Standard errors in parentheses*** p<0.01, ** p<0.05, * p<0.1
Table 7: Sanctions Imposed and Real Exchange Rate via Quantile Regression
|
|
|
|
|
|
Panel A: Quantile 0.90
|
Model 1
|
Model 2
|
Model 3
|
Model 4
|
Model 5
|
|
|
|
|
|
|
US
|
0.202
|
|
|
|
|
|
(0.137)
|
|
|
|
|
EU
|
|
-0.852***
|
|
|
|
|
|
(0.210)
|
|
|
|
U.N.
|
|
|
-0.143
|
|
|
|
|
|
(0.266)
|
|
|
Economic
|
|
|
|
0.145
|
|
|
|
|
|
(0.155)
|
|
Intensity
|
|
|
|
|
0.0332
|
|
|
|
|
|
(0.0398)
|
|
|
|
|
|
|
Control Variables
|
Yes
|
Yes
|
Yes
|
Yes
|
Yes
|
|
|
|
|
|
|
Constant
|
2.614***
|
2.596***
|
2.759***
|
2.506***
|
2.519***
|
|
(0.107)
|
(0.143)
|
(0.132)
|
(0.131)
|
(0.120)
|
|
|
|
|
|
|
Observations
|
209
|
209
|
209
|
209
|
209
|
Pseudo R-squared
|
0.406
|
0.299
|
0.342
|
0.339
|
0.371
|
Standard errors in parentheses
*** p<0.01, ** p<0.05, * p<0.1
Table 8: Sanctions Imposed and Real Exchange Rate (excluding 2008-2010)
|
|
|
|
|
|
|
Model 1
|
Model 2
|
Model 3
|
Model 4
|
Model 5
|
|
|
|
|
|
|
Current
|
0.00873**
|
0.00550
|
0.0100**
|
0.00904**
|
0.00832**
|
|
(0.00397)
|
(0.00408)
|
(0.00402)
|
(0.00405)
|
(0.00405)
|
Lending
|
0.0225***
|
0.0206***
|
0.0237***
|
0.0229***
|
0.0226***
|
|
(0.00443)
|
(0.00439)
|
(0.00443)
|
(0.00444)
|
(0.00442)
|
Inflation
|
-0.00765***
|
-0.00758***
|
-0.00856***
|
-0.00818***
|
-0.00772***
|
|
(0.00208)
|
(0.00200)
|
(0.00207)
|
(0.00205)
|
(0.00208)
|
GDS
|
-0.00269
|
-0.00338
|
-0.00304
|
-0.00294
|
-0.00270
|
|
(0.00289)
|
(0.00283)
|
(0.00294)
|
(0.00291)
|
(0.00289)
|
|
|
|
|
|
|
U.S.
|
-0.121*
|
|
|
|
|
|
(0.0696)
|
|
|
|
|
E.U.
|
|
-0.293***
|
|
|
|
|
|
(0.0940)
|
|
|
|
U.N.
|
|
|
-0.0809
|
|
|
|
|
|
(0.173)
|
|
|
Economic
|
|
|
|
-0.0907
|
|
|
|
|
|
(0.0760)
|
|
Intensity
|
|
|
|
|
-0.0356*
|
|
|
|
|
|
(0.0212)
|
|
|
|
|
|
|
Constant
|
1.402***
|
1.455***
|
1.372***
|
1.395***
|
1.403***
|
|
(0.0940)
|
(0.0942)
|
(0.0935)
|
(0.0954)
|
(0.0944)
|
|
|
|
|
|
|
Observations
|
163
|
163
|
163
|
163
|
163
|
R-squared
|
0.269
|
0.302
|
0.255
|
0.261
|
0.268
|
RMSE
|
0.285
|
0.279
|
0.288
|
0.287
|
0.285
|
F-test
|
10.40
|
12.20
|
9.651
|
9.974
|
10.35
|
Prob > F
|
1.66e-08
|
7.86e-10
|
6.08e-08
|
3.46e-08
|
1.81e-08
|
Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1
4.3 Robustness Analysis
To further estimate the resulting robustness, we eliminate the year 2008-2010. It is believed that the exchange rate of countries changed significantly after the year 2008 financial downturn, which could distort the results. Thus, the data eliminating the financial crisis tends to reflect the impact of sanctions on the exchange rate. From Table 8, we deduce that except for economic sanctions, other sanctions, together with other explanatory variables, are consistent with the result of the baseline results in Table 2. The result shows that U.S., E.U., and Intensity sanctions significantly contribute to the weakened real exchange rate in the target region during the sample period without a financial crisis. It can be inferred that the insignificance effect of the economic sanctions measure remains ineffective as the target countries could navigate through the strategy implemented to counter the impact of economic sanctions on their economy before the 2008 financial crisis.