The effects of currency devaluation on Ethiopia’s major export commodities: The case of coffee and khat: Evidence from the vector error correction model and the Johansen co-integration test

Figures & data

Figure 1. Total coffee and khat export performance (in thousands of metric tone).

Source: NBE, 2020.

Figure 2. Share of coffee export (volume) by destination (2012/13-201,920).

Source: Ethiopian Coffee and Tea Authority, 2020

Figure 3. The values of khat export in million Birr.

Source: National Bank of Ethiopia and author’s computation, 2021.

Table 1. The volume of coffee and khat export (in metric ton) after devaluation in Ethiopia

Year	Coffee	Khat
The first devaluation and its effects on volume of commodity (in 1992 by 142 percent)
1991/92	32,249.00	251
1992/93	67,745.20	1,936.40
1993/94	69,160.00	2,808.48
1994/95	82,198.64	4,073.27
1995/96	97,578.82	3,698.29
The second devaluation and its effects on volume of commodity (in 2010)
2010/11	172,217.23	40,971.74
2011/12	169,408.06	41,052.87
2012/13	199,127.77	47,163.67
2013/14	189,669.31	51,689.76
2014/15	183,870.70	49,204.18
The last period of devaluation and its effects on volume of commodity (in 2017 by 15 percent)
2017/18	238,572.83	47,023.94
2018/19	230,931.42	53,565.13
2019/20	271,111.52	57,136.13

Source: National Bank of Ethiopia, 2020.

Figure 4. Trends of real effective exchange rate (REER).

Source: National Bank of Ethiopia and authors’ computation, 2021.

Table 2. Augmented Dickey–Fuller unit root test

	Augmented Dickey-Fuller test
	At level			At first difference			Order of integration
Variables	T-test	Crit. value	P-value	T-test	Crit. value	P-value
lnTEV	−1.22623	−2.95402	0.6511	−7.29019	−2.95711	0.0000	I(1)
lnREER	−1.85786	−2.95402	0.3472	−5.21319	−2.95711	0.0002	I(1)
lnRGDP	1.880279	−2.95402	0.9997	−4.67119	−2.95711	0.0007	I(1)
lnFRGDP	−2.43018	−2.95402	0.1416	−5.23389	−2.95711	0.0002	I(1)
lnFDI	−2.39337	−2.95402	0.1512	−4.80380	−2.95711	0.0005	I(1)

Source: The results are calculated by the author using Eviews 10.0 software, 2021.

Table 3. VAR optimal lag selection

Lag	LogL	LR	FPE	AIC	SC	HQ
0	−35.10675	NA	0.000159	2.607117	2.793943	2.666884
1	85.56044	201.1120	1.50e-07	−4.370696	−3.436565*	−4.071859
2	106.5806	29.42817*	1.15e-07	−4.705371	−3.023934	−4.167465
3	128.0060	24.28212	9.46e-08*	−5.067064*	−2.638322	−4.290088*
4	143.6500	13.55821	1.39e-07	−5.043336	−1.867289	−4.027292

* indicates lag order selected by the criterion.

LR: sequential modified LR test statistic (each test at 5 percent level); FPE: Final prediction error; AIC: Akaike information criterion; SC: Schwarz information criterion; HQ: Hannan-Quinn information criterion.

Table 4. Johansen co-integration test (trace and eigen value test)

Max rank	Eigen values	Trace test			Eigen value
		Trace statistic	Critical values (0.05)	Prob.	Max-eigen statistic	Critical values (0.05)	Prob.
0	0.93	188.95	69.81	0.00	84.31	33.87	0.00
1	0.83	104.63	47.85	0.00	54.56	27.58	0.00
2	0.61	50.07	29.79	0.00	28.43	21.13	0.00
3	0.48	21.64	15.49	0.00	19.77	14.26	0.00
4	0.06	1.86	3.84	0.17	1.86	3.84	0.17

Source: The results are calculated by the author using Eviews 10 software, 2021.

Table 5. Long-run relationship of co-integrated vector

Regressors	Coefficient	Standard deviation	t-statistics
lnREER	4.620439	1.43315	−3.22397*
lnRGDP	−0.558702	0.27724	2.01525*
lnFRGDP	0.007214	0.05042	−0.14308
LnFDI	0.288340	0.07894	−3.65247*
C	−2.747368

Source: Author’s computation using EVIEWS 10 software, 2021.

*indicates the significance of the coefficients (the rejection of the null hypothesis) at 5%.

Table 6. Short-run dynamics of VECM

Model, VECM	Coefficients	Standard error	t-Statistic	P-value
cointEq1	−0.308818*	0.101640	−3.038364	0.0068*
D(LTEV(−1))	−0.164813	0.198783	−0.829109	0.4173
D(LTEV(−2))	0.202650	0.213683	0.948368	0.3548
D(LREER(−1))	−0.604585	0.480026	−1.259482	0.2231
D(LREER(−2))	−0.176349	0.332505	−0.530367	0.6020
D(LRGDP(−1))	0.188690	0.464295	0.406402	0.6890
D(LRGDP(−2))	0.487889	0.491414	0.992827	0.3333
D(LFRGDP(−1))	0.052183*	0.019279	2.706727	0.0140*
D(LFRGDP(−2))	0.0074226	0.026398	0.281290	0.7815
D(LFDI(−1))	0.091497*	0.040163	−2.278151	0.0345*
D(LFDI(−2))	−0.014983	0.046140	−0.324720	0.7489
C	0.006749	0.046555	0.144972	0.8863
R-squared	0.632171	Mean dependent var		0.021339
Adjusted R-squared	0.419217	S.D. dependent var		0.138280
S.E. of regression	0.105382	Akaike info criterion		−1.377811
Sum squared resid	0.211000	Schwarz criterion		−0.822720
Log likelihood	33.35608	Hannan-Quinn criter.		−1.196865
F-statistic	2.968581	Durbin-Watson stat		1.873040
Prob(F-statistic)	0.018123

Source: Author’s computation, 2021. *indicates the significance of the coefficients (the rejection of the null hypothesis) at 5%.

Figure 5. The stability of VEC model.

Source: Own computation, 2021.

Table 7. Pairwise Granger causality test

Null hypothesis	Observation	F-statistic	P-value
LREER does not Granger Cause LTEV	32	8.78624	0.0012*
LTEV does not Granger Cause LREER		0.38439	0.6845
LRGDP does not Granger Cause LTEV	32	1.78930	0.1863
LTEV does not Granger Cause LRGDP		4.28973	0.0241*
LFRGDP does not Granger Cause LTEV	32	5.26309	0.0117*
LTEV does not Granger Cause LFRGDP		0.10127	0.9040
LFDI does not Granger Cause LREER	32	16.7888	2.E-05*
LREER does not Granger Cause LFDI		0.88193	0.4256
LFDI does not Granger Cause LRGDP	32	3.83526	0.0342*
LRGDP does not Granger Cause LFDI		0.96470	0.3938

Source: Own computation using EVIEWS 10 software, 2021. *indicates the significance of the coefficients (the rejection of the null hypothesis) at 5%

Table 8. Variance decomposition of total export value

Period	S.e	LTEV	LREER	LRGDP	LFRGDP	LFDI
1	0.105382	100.0000	0.000000	0.000000	0.000000	0.000000
2	0.138768	82.61380	5.525700	0.940336	10.90801	0.012155
3	0.193070	77.29081	6.153133	1.240613	10.14045	5.174996
4	0.230832	71.07031	5.586979	1.754849	12.41516	9.172696
5	0.280758	68.36372	4.019843	1.200397	15.91898	10.49705
6	0.319491	66.33599	3.632886	1.005472	17.40309	11.62256
7	0.358304	67.31343	3.028538	0.941199	17.40962	11.30721
8	0.390854	67.12782	2.716545	0.942781	18.13774	11.07511
9	0.422307	67.32937	2.566819	0.888909	18.01373	11.20117
10	0.449410	66.96576	2.492646	1.003907	18.15241	11.38527

Source: Own computation using EVIEWs software, 2021.

Figure 6. Impulse response of total export value.

Source: Author’s computation using STATA software, 2021.