Do remittances affect labour participation decisions and hours worked? Evidence from Ethiopia

Figures & data

Table 1. Variable name, measurement and summary statistics of dependent and independent variables

Variable Name	Measurement	Mean	Std. Dev.		Mean		Std. Dev.	Min		Max
		Age 7–14			Age 15–64
Dependent Variables:
Labor participation (LPd)	=1 if yes,0 = other wise	.39		.49	.55	.5		0	1
LPd is in HH agriculture (HHAgr)	=1 if yes,0 = otherwise	.37		.48	.37	.48		0	1
LPd is in HH non-agriculture activities (HHnonAgr)	=1 if yes,0 = otherwise	.02		.14	.12	.32		0	1
LPd is in temporary paid activities (tempaid)	1 if yes,0 = otherwise	.01		.08	.04	.2		0	1
LPd is in non-temporary paid activities (nontempaid)	1 if yes,0 = otherwise	0		.04	.08	.28		0	1
Hours worked (Hours_worked)	Hours worked in a week	9.57		16.56	17.41	22.81		0	130
Independent Variables:
Foreign remittances	Continuous (LCU)	512.38		4444.3	869.13	9164.3		0	130,000
Domestic remittances	Continuous (LCU)	373.7		4825.2	540.23	6270.8		0	400,000
Receiving foreign remittances	=1 if yes,0 = otherwise	.05		.22	.07	.25		0	1
Receiving domestic remittances	=1 if yes,0 = otherwise	.11		.32	.13	.34		0	1
sex	=1 if male,0 = otherwise	.51		.5	.47	.5		0	1
Age	Year	10.41		2.24	31.83	13.02		7	14
Years ofschooling	Years ofschooling	2.08		2.16	4.48	4.73		0	18
Head Sex	=1 if male,0 = otherwise	.79		.41	.76	.43		0	1
Head Age	Year	45.95		12.21	45.94	13.71		8	98
Household Size	count	6.45		2.13	5.56	2.4		1	16
Working male	Total working age males	1.61		1.1	1.85	1.29		0	10
Working Female	Total working age females	1.53		.88	1.74	1.02		0	8
Paeq.consumption expenditure	Continuous (LCU)	5503.89		5243.57	6802.87	6356.31		810.65	347,501.4
Wealth	Index	−2.95		2.33	−2.2	2.86		−6.03	9.71
Shock occurrence	=1 if yes,0 = otherwise	.51		.5	.48	.5		0	1
Rural dummy	=1 if rural,0 = otherwise	.79		.41	.65	.48		0	1
Year dummy	= 1 if 2013,0 = otherwise	.5		.5	.5	.5		0	1
		N1 = 11,193			N2 = 25,057

Note: The panel survey data utilized in this study did not include an aggregate data for the total household income. However, aggregate household consumption expenditure data is available. Hence, per adult equivalent consumption expenditure is used to proxy the per adult equivalent income. “HH” refers to “Household”, “Paeq.” refers to “per adult equivalent”, LCU refers to “Local currency unit”.

Table 2. Comparison of children and adults in remittance receiving and non-receiving households: an independent two-sample t-test

	Age7-14			Age15-64			Age7-14			Age15-64
	Foreign remittance						Domestic remittance
Variable	Non- Receiving (n= 10,641)	Receiving (n= 552)	diff.	Non- receiving (n= 23,360)	receiving (n= 1,697)	diff.	Non- receiving (n= 9,925)	Receiving (n= 1,268)	diff.	Non- Receiving (n= 21,765)	Receiving (n= 3292)	diff.
LPd	.40	.31	.09**	.56	.46	.10***	.39	.36	.04**	.55	.52	.03***
LPd HHAgr	.29	.38	(.09)**	.38	.25	.13***	.38	.34	.04**	.38	.29	.09**
LPdHHnonAgr	.02	.01	.01	.12	.10	.02**	.02	.02	.00	.11	.12	(.01)
LPd tempaid	.01	.01	.00	.04	.03	.01	.01	.01	.00	.04	.05	(.01)**
LPdnontempaid	.002	.002	.000	.08	.12	(.04)**	.002	.001	.001	.08	.12	(.04)**
Hours_worked	9.63	8.24	1.39**	17.52	15.79	1.73**	9.68	8.70	.98**	17.49	16.86	.630

Note: “diff.” represent the mean difference. ***, **, * refers to the difference is statistically significant at 1%, 5%, and 10% respectively. The null and alternative hypotheses are H0: diff = 0, and Ha: diff # 0, respectively.

Table 3. Fixed Effect (FE) and Random Effect (RE) logit results: Remittance receiving status (1=Yes) is the key independent variable and dependent variable: Labor Participation decision (LPd)

	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)
	Total sample FE	Total sample RE	Male RE	Female RE	Rural RE	Urban RE	Age 7-14 RE	Age 15-25 RE	Age 26-64 RE
Receiving foreign remittances	-.29***	-.19***	-.11	-.26***	.03	-.38***	-.16	-.20**	-.18**
	(.1)	(.05)	(.08)	(.07)	(.08)	(.08)	(.11)	(.09)	(.07)
Receiving domestic remittances	.03	-.02	-.01	-.04	.16***	-.21***	-.05	0.00	.05
	(.07)	(.04)	(.06)	(.05)	(.06)	(.06)	(.07)	(.06)	(.05)
sex	.83***	.77***			.92***	.52***	.6***	.64***	.85***
	(.05)	(.03)			(.04)	(.05)	(.05)	(.06)	(.04)
Age	.02***	.02***	.02***	.02***	.01***	.04***
	(0.00)	(0.00)	(0.00)	(0.00)	(0.00)	(0.00)
Years of schooling	.01	.01**	0.00	.02***	-.03***	.04***	.1***	-.02***	.01***
	(.01)	(0.00)	(.01)	(.01)	(.01)	(.01)	(.01)	(.01)	(.01)
Head Sex	.12*	.1***	.39***	-.11**	.07	.13**	.19***	.06	.04
	(.06)	(.03)	(.06)	(.05)	(.05)	(.05)	(.06)	(.05)	(.05)
Head Age	-.01***	-.01***	-.01***	-.01***	-.01***	-.02***	00**	00**	-.01***
	(0.00)	(0.00)	(0.00)	(0.00)	(0.00)	(0.00)	(0.00)	(0.00)	(0.00)
Household Size	.01	-.02**	-.03**	-.01	-.02**	-.01	-.07***	-.05***	-.02**
	(.01)	(.01)	(.01)	(.01)	(.01)	(.02)	(.01)	(.01)	(.01)
Working male	-.10***	-.08***	-.09***	-.04**	-.1***	-.04	-.02	-.03	-.06***
	(.02)	(.01)	(.02)	(.02)	(.02)	(.02)	(.02)	(.02)	(.02)
Working Female	0.00	-.01	.02	-.06***	-.02	.03	.09***	.01	.03
	(.03)	(.02)	(.03)	(.02)	(.02)	(.03)	(.03)	(.03)	(.02)
Paeq. Consumption exp.(log)	.22***	.1***	.14***	.06*	.11***	.04	.32***	.17***	.06*
	(.05)	(.02)	(.04)	(.03)	(.03)	(.04)	(.04)	(.04)	(.03)
Wealth	0.00	-.01	0.00	-.01	.05***	-.04***	-.2***	-.06***	0.00
	(.01)	(.01)	(.01)	(.01)	(.02)	(.01)	(.02)	(.01)	(.01)
Shock occurrence	.15***	.03	.06	.01	.04	-.02	.17***	.04	.01
	(.05)	(.03)	(.04)	(.04)	(.04)	(.05)	(.05)	(.05)	(.04)
Rural dummy	.42***	.66***	.81***	.55***			1.67***	.82***	.5***
	(.08)	(.04)	(.07)	(.06)			(.1)	(.07)	(.06)
Year dummy	.11***	.08***	.15***	.03	-.01	.22***	.12***	.09*	.11***
	(.04)	(.03)	(.04)	(.04)	(.04)	(.05)	(.05)	(.05)	(.04)
_cons		-1.42***	-1.22***	-.9***	-.22	-1.66***	-5.81***	-2***	-.19
		(.24)	(.37)	(.32)	(.31)	(.44)	(.41)	(.4)	(.31)
/lnsig2u		-5	-2.38*	-7.02	-3.02**	-8.24	-1.46**	-2.03*	-9.49
		(7.12)	(1.26)	(97.93)	(1.48)	(396.63)	(.57)	(1.05)	(1128.83)
Observations	8104	25057	11885	13172	16290	8767	11193	9864	15193
Pseudo R^2	.09	.055	.052	.022	.041	.069	.110	.060	.045
Wald chi2		1429.20	532.77	339.63	684.10	516.84	728.17	450.31	652.25
LR chi2	525.77
Prob > chi2	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000

Standard errors are in parenthesis and for the random effect models, they are cluster-robust standard errors. *** p < .01, ** p < .05, * p < .1

Table 4. Fixed Effect (FE) and Random Effect (RE) logit result by occupation type: remittance receiving status (1 = yes) is the key independent variable and dependent variable: Labor Participation decision(LPd)

	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)
	HHAgr FE	HHAgr RE	HHnon-Agr FE	HHnon-Agr RE	tempaid FE	tempaid RE	Non tempaid FE	Non tempaid RE
Receiving foreign remittances	-.107	.012	-.203	-.444***	-.326	-.098	-.422**	-.263***
	(.147)	(.07)	(.165)	(.092)	(.251)	(.146)	(.17)	(.096)
Receiving domestic remittances	.162	.04	.058	-.182***	-.019	.237***	.092	.008
	(.099)	(.049)	(.122)	(.064)	(.174)	(.09)	(.143)	(.071)
sex	.873***	.82***	.041	.073	.832***	.846***	.566***	.567***
	(.066)	(.036)	(.091)	(.046)	(.146)	(.076)	(.115)	(.062)
Age	.003	.005***	.015***	.018***	-.003	.004	.032***	.037***
	(.003)	(.001)	(.004)	(.002)	(.006)	(.003)	(.005)	(.002)
Years of schooling	-.056***	-.059***	-.038***	-.051***	-.016	-.007	.136***	.173***
	(.01)	(.005)	(.012)	(.006)	(.018)	(.009)	(.014)	(.008)
Head Sex	.112	.156***	.023	-.013	-.45***	-.227***	.019	-.024
	(.083)	(.042)	(.107)	(.053)	(.163)	(.082)	(.128)	(.064)
Head Age	0	-.003**	-.018***	-.024***	-.006	-.012***	-.017***	-.015***
	(.003)	(.001)	(.004)	(.002)	(.005)	(.003)	(.004)	(.002)
Household Size	.023	.011	.035	.021	-.046	-.067***	-.083**	-.103***
	(.018)	(.009)	(.027)	(.013)	(.041)	(.02)	(.039)	(.019)
Working Male	-.038	-.056***	-.076*	-.027	-.135**	-.069**	-.11*	-.102***
	(.029)	(.015)	(.041)	(.021)	(.066)	(.034)	(.06)	(.031)
Working Female	.052	.024	.044	.013	-.138	-.07*	.046	-.019
	(.037)	(.02)	(.052)	(.026)	(.09)	(.042)	(.067)	(.034)
Paeq. Consumption exp.(log)	.08	.039	.247***	.224***	.045	-.034	.067	.002
	(.061)	(.028)	(.081)	(.038)	(.122)	(.056)	(.104)	(.05)
Wealth	-.071***	-.114***	.123***	.055***	-.038	-.042**	-.036	.042***
	(.025)	(.012)	(.025)	(.011)	(.038)	(.019)	(.028)	(.013)
Shock occurrence	.216***	.004	.016	-.048	.719***	.256***	-.089	.04
	(.066)	(.032)	(.09)	(.044)	(.135)	(.069)	(.109)	(.055)
Rural dummy	2.035***	2.184***	-1.519***	-1.112***	-.131	-.315***	-1.681***	-1.215***
	(.128)	(.066)	(.141)	(.069)	(.199)	(.102)	(.204)	(.088)
Year dummy	.112**	.079**	-.013	.039	-.166*	-.219***	-.022	.094*
	(.048)	(.032)	(.066)	(.043)	(.098)	(.069)	(.082)	(.056)
_cons		-3.227***		-2.748***		-2.127***		-3.192***
		(.276)		(.38)		(.576)		(.51)
/lnsig2u		-2.941**		-1.205***		-10.547		-2.165*
		(1.205)		(.442)		(10252.34)		(1.186)
Observations	6646	25057	3292	25057	1300	25057	2492	25057
Pseudo R^2	.261	.219	.218	.069	.142	.037	.363	.260
Wald chi2		2871.03		830.37		274.97		1724.75
LR chi2	1201.21		496.82		127.56		626.27
Prob > chi2	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000

Standard errors are in parenthesis and for the random effect models, they are cluster-robust standard errors. *** p < .01, ** p < .05, * p < .1

Table 5. Left censored Tobit estimation with Random Effect (RE) approach: remittance receiving status (1 = yes) is the key independent variable and dependent variable: hours worked

	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)
	Total Sample	Male	Female	Rural	Urban	Age 7–14	Age 15–25	Age 26–64
Receiving foreign remittances	−4.941***	−3.79***	−6.126***	−.11	−10.702***	−1.648	−5.625***	−4.279***
	(1.007)	(1.372)	(1.488)	(1.191)	(1.944)	(1.773)	(1.734)	(1.231)
Receiving domestic remittances	−1.527**	−1.469	−1.88*	.466	−5.181***	−1.591	−1.474	−.163
	(.736)	(1.034)	(1.059)	(.833)	(1.493)	(1.169)	(1.263)	(.902)
sex	15.759***			15.782***	15.076***	10.243***	13.546***	16.532***
	(.536)			(.539)	(1.295)	(.71)	(1.048)	(.639)
Age	.361***	.309***	.331***	.096***	1.048***
	(.021)	(.028)	(.035)	(.021)	(.052)
Years of schooling	.086	−.178*	.363***	−.609***	.959***	1.106***	−.609***	.161*
	(.071)	(.092)	(.111)	(.083)	(.141)	(.175)	(.125)	(.084)
Head Sex	1.027	6.825***	−3.613***	−.114	2.26*	2.693***	.483	−.342
	(.629)	(.981)	(.911)	(.694)	(1.325)	(.981)	(1.071)	(.798)
Head Age	−.257***	−.243***	−.221***	−.154***	−.525***	.067**	−.108***	−.235***
	(.02)	(.028)	(.031)	(.021)	(.049)	(.03)	(.032)	(.025)
Household Size	−.34**	−.431**	−.132	−.276**	−.494	−1.262***	−.905***	−.424**
	(.145)	(.192)	(.222)	(.138)	(.407)	(.218)	(.262)	(.172)
Working male	−1.225***	−1.453***	−.657*	−1.419***	−.718	−.037	−.421	−.809***
	(.235)	(.312)	(.37)	(.234)	(.602)	(.372)	(.429)	(.288)
Working Female	−.027	.293	−.868**	−.344	1.163*	1.581***	.66	.577
	(.3)	(.423)	(.442)	(.313)	(.704)	(.468)	(.529)	(.374)
Paeq. Consumption exp.(log)	3.949***	4.418***	3.329***	3.929***	2.412**	6.496***	5.386***	3.072***
	(.441)	(.593)	(.66)	(.44)	(1.121)	(.64)	(.803)	(.518)
Wealth	.319**	.632***	.1	1.384***	−.379	−3.327***	−.931***	.722***
	(.148)	(.202)	(.218)	(.234)	(.265)	(.29)	(.256)	(.183)
Shock occurrences	1.538***	1.847***	1.315*	2.471***	−.959	4.481***	1.784**	1.229**
	(.505)	(.68)	(.754)	(.515)	(1.191)	(.739)	(.905)	(.599)
Rural dummy	4.443***	4.326***	4.703***			27.25***	8.114***	.975
	(.79)	(1.075)	(1.17)			(1.451)	(1.403)	(.945)
Year dummy	3.345***	4.257***	2.349***	2.031***	5.872***	4.445***	4.619***	3.19***
	(.509)	(.685)	(.762)	(.523)	(1.214)	(.744)	(.915)	(.605)
_cons	−36.968***	−26.587***	−31.442***	−19.366***	−48.034***	−107.142**	−53.32***	−11.456**
	(4.358)	(5.852)	(6.557)	(4.474)	(11.228)	(6.348)	(7.843)	(5.104)
/sigma_u	3.179	5.551**	3.86	4.485***	0	8.173***	9.129***	2.988
	(2.487)	(2.618)	(4.584)	(1.709)	(10.654)	(1.899)	(2.712)	(3.692)
/sigma_e	34.878***	32.991***	36.633***	28.742***	47.125***	30.031***	36.604***	32.978***
	(.324)	(.524)	(.61)	(.346)	(.598)	(.621)	(.795)	(.427)
Observations	25,057	11,885	13,172	16,290	8767	11,193	9864	15,193
Pseudo R^2	.011	.008	.004	.012	.017	.032	.009	.012
Wald chi2	1814.06	683.69	294.67	1320.32	819.11	1320.32	492.06	1306.76
Prob > chi2	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000

Standard errors are in parenthesis and for the random effect models, they are cluster-robust standard errors. *** p < .01, ** p < .05, * p < .1

Figure A1. Outlier examination: a box plot method on weekly hours worked.

Table A1. Fixed Effect (FE) and Random Effect (RE) logit result: per adult equivalent amount of remittances(log) is the key independent variable and dependent variable: Labor Participation decision(LPd)

	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)
	Total SampleFE	Total SampleRE	MaleRE	FemaleRE	RuralRE	UrbanRE	Age 7_14RE	Age 15_25RE	Age 26_64RE
Paeq.foreign remittances(log)	−.043***	−.031***	−.018	−.042***	.003	−.053***	−.023	−.03**	−.031***
	(.014)	(.008)	(.012)	(.01)	(.012)	(.01)	(.016)	(.012)	(.01)
Paeq.domestic remittances(log)	−.01	−.018***	−.015	−.023***	.02*	−.043***	−.022	−.016	−.001
	(.012)	(.007)	(.011)	(.009)	(.011)	(.009)	(.014)	(.01)	(.009)
Year dummy	.111***	.082***	.151***	.034	−.009	.219***	.123***	.09*	.113***
	(.038)	(.028)	(.043)	(.038)	(.036)	(.048)	(.046)	(.046)	(.036)
_cons		−1.413***	−1.215***	−.887***	−.239	−1.629***	−5.818***	−1.992***	−.193
		(.241)	(.374)	(.322)	(.307)	(.444)	(.411)	(.399)	(.309)
/lnsig2u		−5.197	−2.4*	−7.095	−3.023**	−8.279	−1.435**	−2.021*	−9.497
		(8.663)	(1.281)	(105.826)	(1.481)	(414.061)	(.558)	(1.04)	(1134.328)
Observations	8104	25,057	11,885	13,172	16,290	8767	11,193	9864	15,193
Pseudo R^2	.094	.055	.052	.022	.041	.070	.110	.060	.046
Wald chi2		1438.13	534.38	346.17	682.32	520.91	728.82	450.03	654.28
LR chi2	527.63
Prob > chi2	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000

Standard errors are in parenthesis and for the random effect models, they are cluster-robust standard errors. *** p < .01, ** p < .05, * p < .1

Note: All control variables included in Table 3 are also included to estimate Table A1 but removed from the table to save space.

Table A2. Fixed Effect (FE) and Random Effect (RE) logit estimation by occupation type: Per adult equivalent amount of remittances(log) is the key independent variable and dependent variable: Labor Participation decision(LPd)

	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)
	HHAgrFE	HHAgrRE	HHnonAgrFE	HHnonAgrRE	TempaidFE	TempaidRE	Non tempaidFE	Non tempaidRE
Paeq. foreign remittances(log)	−.013	−.002	−.029	−.062***	−.047	−.018	−.061***	−.038***
	(.022)	(.01)	(.023)	(.012)	(.034)	(.021)	(.022)	(.013)
Paeq. Domestic remittances(log)	.024	−.002	−.015	−.048***	−.028	.021	.003	−.009
	(.018)	(.009)	(.02)	(.01)	(.03)	(.014)	(.022)	(.011)
Year dummy	.112**	.08**	−.012	.04	−.168*	−.216***	−.021	.095*
	(.048)	(.032)	(.066)	(.043)	(.098)	(.069)	(.082)	(.056)
_cons		−3.233***		−2.718***		−2.122***		−3.164***
		(.276)		(.38)		(.576)		(.51)
/lnsig2u		−2.936**		−1.209***		−10.452		−2.187*
		(1.199)		(.441)		(9415.522)		(1.212)
Observations	6646	25,057	3292	25,057	1300	25,057	2492	25,057
Pseudo R^2	.261	.219	.218	.070	.143	.036	.363	.260
Wald chi2		2876.04		836.20		270.83		1728.08
LR chi2	1200.21		497.32		128.61		627.41
Prob > chi2	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.0000

Standard errors are in parenthesis and for the random effect models, they are cluster-robust standard errors. *** p < .01, ** p < .05, * p < .1

Note: All control variables included in Table 3 are also included to estimate Table A2 but removed from the table to save space.

Table A3. Left censored Tobit estimation with Random Effect (RE) approach: Per adult equivalent amount of remittances (log) is the key independent variable and dependent variable: Hours worked

	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)
	Total Sample	Male	Female	rural	Urban	Age 7_14	Age 15_25	Age 26_64
Paeq.foreign remittances(log)	−.792***	−.574***	−1.014***	−.097	−1.47***	−.241	−.893***	−.708***
	(.142)	(.195)	(.209)	(.182)	(.258)	(.264)	(.249)	(.172)
Paeq.domestic remittances(log)	−.512***	−.457**	−.629***	−.033	−1.079***	−.437*	−.521**	−.179
	(.124)	(.179)	(.175)	(.162)	(.224)	(.226)	(.208)	(.156)
Year dummy	3.348***	4.252***	2.363***	2.04***	5.895***	4.429***	4.64***	3.184***
	(.509)	(.685)	(.761)	(.522)	(1.213)	(.75)	(.914)	(.605)
_cons	−36.8***	−26.447***	−31.351***	−19.559***	−47.342***	−107.337**	−53.103***	−11.462**
	(4.356)	(5.85)	(6.551)	(4.473)	(11.22)	(6.339)	(7.839)	(5.103)
/sigma_u	3.168	5.549**	3.726	4.482***	0	0	9.128***	3.014
	(2.493)	(2.618)	(4.742)	(1.711)	(10.552)	(.031)	(2.714)	(3.662)
/sigma_e	34.862***	32.982***	36.617***	28.74***	47.085***	31.122***	36.585***	32.967***
	(.324)	(.524)	(.609)	(.346)	(.597)	(.372)	(.795)	(.427)
Observations	25,057	11,885	13,172	16,290	8767	11,193	9864	15,193
Pseudo R^2	.011	.008	.004	.012	.017	.032	.009	.012
Wald chi2	1833.16	689.28	310.30	1320.51	830.33	1320.77	498.37	1313.24
Prob > chi2	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000

Standard errors are in parentheses. *** p < .01, ** p < .05, * p < .1

Note: All control variables included in Table 3 are also included to estimate Table A3 but removed from the table to save space.

Table A4. Comparison of Years of Schooling by gender and age group: An Independent two-sample t-test

Variable	Foreign remittance			Domestic remittance
	Non-Receiving	Receiving	diff.	Non-Receiving	Receiving	diff.
Female	3.13	5.24	(2.11)***	3.11	4.27	(1.16)***
Male	4.12	6.10	(1.97)***	4.12	5.10	(.98)***
Age 7–14	2.04	2.71	(.66)***	2.05	2.26	(.21)***
Age 15–25	5.77	7.34	(1.57)***	5.68	7.09	(1.40)***
Age 26–64	3.40	5.99	(2.58)***	3.45	4.33	(.87)***

Note: “diff.” represent the mean difference. ***, **, * refers to the difference is statistically significant at 1%, 5%, and 10% respectively. The null and alternative hypotheses are H0: diff = 0, and Ha: diff # 0, respectively.