Climatic shocks and multidimensional energy poverty in Ugandan households: does women empowerment play a moderating role?

Figures & data

Table 1. Dimensions, indicators and deprivation cut-offs for multidimensional energy poverty.

Dimension(i)s	Indicator(j)	Variable	Deprivation cut-off (energy poor if) (1/0)	Weight
Electricity Energy supply (energy access)	Capacity	Power capacity ratings	Household has a power capacity rating of less than 500 kWh per day, uses a solar lantern, or has no access to electricity.	1/9
(1/3)	Quality	Quality of power	Household experiences low voltage of electricity, damage to appliances or uses a solar system that cannot power large devices or households use devices such as candles, dry cell batteries, kerosene lamps, etc., due to lack of access.	1/9
	Affordability	Affordability of power	Household spends >5% of total income on energy or household cannot afford to pay connection fee, and even if they are waived, they cannot afford the electricity bills.	1/9
Cooking solutions	Cooking exposure	Cooking exposure	Household uses a three-stone/open-fire stove or an unimproved cooking stove with traditional fuels in a poorly ventilated area.	1/6
(1/3)	Affordability	Expenditure on cooking fuel	Household’s expenditure on cooking fuel is more than 5% of the total income or cannot afford an Improved Cooking Stove (ICS).	1/6
Electricity service and appliance	Information, communication, and entertainment	Information, communication, and entertainment appliance ownership	Household neither has a radio, mobile phone, phone charger, TV nor VCD/DVD.	1/9
(1/3)	Housing	Roofing type	Household members live in a dwelling with a grass thatched roof.	1/9
	Access to community services	Existence of at least one enterprise powered with electricity in the community	Community has no milk collection centre, grain/saw/oil mill, mobile phone charging/repairing services, restaurant/tea or coffee shop powered by electricity.	1/9

Based on the Alkire – Foster (AF) approach (Alkire, Kanagaratnam, and Suppa 2020) and Ssennono et al. (2021).

Table 2. Indicators for constructing the women empowerment index.

Domains	Indicator	Definition of adequacy	Source	Weight
Production	Input in productive decisions	=1 if the respondent owns an income generating activity (either agricultural or non-agriculture or both) individually and jointly	Malapit et al. (2019), Diiro et al. (2018)	1/5
Resource	Ownership of assets	=1 if the respondent solely or jointly owns land, buildings, cattle or goats	Malapit et al. (2019)	1/10
	Access and decisions on financial services	' = 1 if the respondent has access to and participates in decision-making concerning credit or individually or jointly operates an account in a financial institution	Malapit et al. (2019), Diiro et al. (2018), Boender, Malhorta, and Schuler (2002), Alkire et al. (2013), Sen (1999)	1/10
Income	Control over the use of income	=1 if the respondent decides on the use of income from all income generating activities	Malapit et al. (2019); Iqbal et al. (2015)	1/10
	Autonomy in income	=1 if a female is employed in a managerial position	Malapit et al. (2019)	1/10
Time	Workload	=1 if the respondent worked less than 10.5 hours in the previous 24 h	Malapit et al. (2019), Diiro et al. (2018)	1/10
	Leisure	=1if the respondent participated in leisure activities like visiting neighbors, watching TV, listening to the radio, seeing movies or doing sports	Malapit et al. (2019), Diiro et al. (2018)	1/10
Leadership	Group membership	=1 if the respondent is an active member in at least one economic or social group	Alkire et al. (2013), Malapit et al. (2019)	1/10
	Membership in influential groups	=1 if the respondent is an active member of the local council committee or actively participates in the local council committee	Malapit et al. (2019)	1/10

Source: Author’s construction (2022) based on the ideas of Malapit et al. (2019), Alkire et al. (2013), Diiro et al. (2018).

Figure 1. Composition of women empowerment index.

Figure 2. Incidence of multidimensional energy poverty in Uganda. Source: Authors’ own developed based on 2014, 2016, 2019 and 2020 UNPS data.

Table 3. Climatic shocks and multidimensional energy poverty – Marginal effects from random effects probit.

	Column 1		Column 2 (4)
Variables	Marginal effects	Se	Marginal effects	Se
Climatic shocks	0.087***	(0.01)	0.054***	(0.011)
Age			−0.006***	(0.002)
Age squared			0.004**	(0.002)
Education of head [No schooling]
Primary			−0.117***	(0.017)
Secondary			−0.241***	(0.021)
Post-secondary			−0.424***	(0.029)
Employment status of HH head [Nonfarm Wage]
Farm wage employment			0.122***	(0.028)
Nonfarm self-employment			0.044**	(0.017)
On-farm employment			0.100***	(0.018)
Unemployed			0.136***	(0.025)
Poor household [non-Poor]			0.228***	(0.015)
Number of child members			0.018***	(0.003)
Share of women in employment			−0.042**	(0.019)
Year of the survey [2014]
2016			−0.116***	(0.013)
2018			−0.277***	(0.012)
2020			−0.296***	(0.011)
Household fixed effect			Yes
Ecological zones fixed effect			Yes
Observations			8124

Notes: the reference category for education is no schooling, for employment status is nonfarm wage, for poor households is non-poor household and the year of the survey is 2014. Robust standard errors in parentheses.

Standardized coefficients in brackets. Level of sig. *** p<0.01, ** p<0.05, * p<0.1.

Table 4. Robustness check (energy poverty by type of climatic Shock).

VARIABLES	Marginal effects
Flood	0.062*
	(0.037)
Drought	0.049***
	(0.018)
Irregular rainfall	0.077***
	(0.027)
Control variables	Yes
Year fixed effects	Yes
Household fixed effects	Yes
Cluster fixed effects	Yes
Observations
Observations	8124

Robust standard errors in parentheses.

*** p<0.01, ** p<0.05, * p<0.1.

Table 5. Moderating effect of women empowerment on climatic shocks and energy poverty.

Variables	Marginal effects	Se
Climatic shocks	0.038***	(0.018)
Female is empowered	−0.064***	(0.018)
Shock* Female is empowered	−0.087***	(0.033)
Covariates	Yes
Year Fixed Effect	Yes
Cluster fixed effect	Yes
Household fixed effects	Yes
Observations	8124

Robust standard errors in parentheses; *** p<0.01, ** p<0.05, * p<0.1.

Table 6. Moderating effect of women empowerment on climatic shocks and multidimensional energy poverty.

	Model (1)	Model (2)
Variables	Marginal	Marginal
Climatic Shocks	0.053***	0.050**
	(0.017)	(0.023)
Empowerment Category (neither Male nor Female is empowered)
Only female is empowered	−0.081***	−0.081***
	(0.018)	(0.020)
Only male is empowered	−0.062***	−0.059**
	(0.022)	(0.025)
Both males and females are empowered	−0.113***	−0.110***
	(0.026)	(0.030)
Shock* Only male is empowered		−0.001
		(0.038)
Shock* Only female is empowered		−0.010
		(0.045)
Shock*Both Males and females are empowered		−0.081***
		(0.013)
Covariates	Yes	Yes
Year Fixed Effect	Yes	Yes
Cluster fixed effect	Yes	Yes
Household fixed effects	Yes	Yes
Observations	8124	8124

Robust standard errors in parentheses; *** p<0.01, ** p<0.05, * p<0.1.

Table 7. Women empowerment and multidimensional energy poverty (Lewbel 2SLS Regression).

	(1)
Variables	Internal Instruments
Women empowerment	−0.118**
	(0.047)
HH head x-tics
Household variables	Yes
Year fixed effects	Yes
Regional fixed effects	Yes
F-statistic	69.63
J-p value	.0344
Constant	0.330***
	(0.015)
Observations	8124
R-squared	0.109

All regressions include relevant covariates consistent with Table 3.

Standard errors are in parentheses.

***p < 0.01, **p < 0.05, *p < 0.1.

Table 8. PSM results with different matching methods.

	Column 1	Column 2
Matching Method	(Climatic Shocks ATT)	(Women Empowerment ATT)
1 – Nearest Neighbor (one-to-one)	0.096***	−0.078***
	(0.027)	(0.025)
5 – Nearest Neighbor	0.080***	−0.082***
	(0.023)	(0.018)
Radius	0.125***	−0.124***
	(0.019)	(0.017)
Kernel	0.090***	−0.083***
	(0.017)	(0.015)
Local liner regression	0.080***	−0.090***
	(0.018)	(0.015)
Observations	8124	8124

Robust standard errors in parentheses; *** p<0.01, ** p<0.05, * p<0.1.

The number of Bootstrap for all estimates is 50.

Table 9. The effect of percent deviation in rainfall days on multidimensional energy poverty.

VARIABLES	Marginal effects	Se
Percent deviation in rainfall days	0.476***	(0.065)
Year of survey fixed effects	Yes
Household fixed effect	Yes
Ecological zones fixed effect	Yes
Observations	8124

Robust standard errors in parentheses.

*** p<0.01, ** p<0.05, * p<0.1.

Table 10. Estimated marginal effects on climatic shocks, women empowerment and multidimensional energy poverty using sub-sample analysis.

	Column (1)	Column (2)	Column (3)
	75% sample	50% sample	25% sample
Variables	Marginal effect	Marginal effect	Marginal effect
Climatic shocks	0.061***	0.064***	0.059***
	(0.019)	(0.024)	(0.013)
Women empowerment	−0.069***	−0.124***	−0.081**
	(0.019)	(0.024)	(0.034)
Covariates	Yes	Yes	Yes
Year fixed effect	Yes	Yes	Yes
Cluster fixed effect	Yes	Yes	Yes
Household fixed effects	Yes	Yes	Yes
Observations	6093	4062	2031

Robust standard errors in parentheses.

*** p<0.01, ** p<0.05, * p<0.1.

Table 11. Robustness checks (Alternatives cut-offs for energy poverty).

	(1)	(2)
VARIABLES	25% cut-off	40% cut-off
Climatic shocks	0.123***	0.069***
	(0.018)	(0.015)
Women empowerment	−0.048***	−0.054***
	(0.017)	(0.015)
Control variables	Yes	Yes
Year fixed effects	Yes	Yes
Household fixed effects	Yes	Yes
Cluster fixed effects	Yes	Yes
Observations	10,940	10,940

Standard errors in parentheses.

*** p<0.01, ** p<0.05, * p<0.1.

Table A1. Descriptive statistics.

Variable	Description	Mean	SD
MEPI	1 = if household energy poor based on the multidimensional energy poverty cuff-off 30%	0.237	0.425
WEATHER	1  = if the household experienced climatic shocks drought, irregular rains, floods, erosions, or landslides in the last 12 months	0.181	0.385
FLOOD	1  = if the household experienced floods in the last 12 months	0.026	0.16
DROUGHT	1  = if the household experienced droughts in the last 12 months	0.118	0.323
IRRAIN	1  = if the household experienced irregular rains in the last 12 months	0.05	0.218
RAINFALL DAYS	The number of days with at least 1 mm of rain for the year	.457	1.79
EMPOWERED_F	1 = if Female is empowered	0.287	0.452
EMPOWERED_M	1 = if male is empowered	0.426	0.495
EMPOWERED_1	1 = if both females and males are not empowered in the household	0.282	0.45
EMPOWERED_2	1 = if only males are empowered in the household	0.138	0.345
EMPOWERED_3	1 = if only females are empowered in the household	0.139	0.346
EMPOWERED_4	1 = if both females and males are empowered in the household	0.44	0.496
AGEHEAD	Age of household head	45.362	14.123
NOEDUC	1 =  if the household head has no formal education	0.091	0.287
PRIMARY	1 = if household head's highest education level achieved is primary education	0.528	0.499
SECONDARY	1 = if household head's highest education level achieved is secondary education	0.25	0.433
TERTIARY	1 = if household head's highest education level achieved is a Diploma or bachelor's and above	0.132	0.338
POOR	1 = if household is below the income poverty line for their region	0.113	0.316
HHCHILDREN	Number of children in the household aged 0–18 years	2.167	1.731
FLABORSHARE	Female share of working population in the households	0.519	0.273
RURAL	1 = if household resides in urban areas	0.688	0.463
URBAN	1 = if household resides in rural areas	0.312	0.463
CENTRAL	1 = if household resides in the central region	0.325	0.468
EASTERN	1 = if household resides in the Eastern region	0.223	0.416
NORTHERN	1 = if household resides in the Northern region	0.189	0.391
WESTERN	1 = if household resides in the western region	0.264	0.441
NONFARM WAGE	1 =  if the household head has a nonfarm wage job	0.203	0.402
FARM WAGE EMPLOYMENT	1 =  if the household head has a farm wage job	0.052	0.222
NONFARM SELF-EMPLOYMENT	1 =  if the household head has a nonfarm self-employed job	0.221	0.415
ON FARM EMPLOYMENT	1 =  if the household head has on farm employment job	0.441	0.496
UNEMPLOYED	1 =  if the household head is unemployed./ not in the labour force	0.041	0.198

Alkire, S., U. Kanagaratnam, and N. Suppa. 2020. The global Multidimensional Poverty Index (MPI) 2020. Tech. rep., OPH I MPI Methodological Note 49. Oxford Poverty and Human Development.

 Google Scholar

Ssennono, V. F., J. M. Ntayi, F. Buyinza, F. Wasswa, S. M. Aarakit, and C. N. Mukiza. 2021. “Energy Poverty in Uganda: Evidence from a Multidimensional Approach.” Energy Economics 101: 105445.

 Web of Science ®Google Scholar

Malapit, H., A. Quisumbing, R. Meinzen-Dick, G. Seymour, E. M. Martinez, J. Heckert, and S. Team. 2019. “Development of the Project-Level Women’s Empowerment in Agriculture Index (pro-WEAI).” World Development 122: 675–692.

 PubMed Web of Science ®Google Scholar

Diiro, G. M., G. Seymour, M. Kassie, G. Muricho, and B. W. Muriithi. 2018. “Women’s Empowerment in Agriculture and Agricultural Productivity: Evidence from Rural Maize Farmer Households in Western Kenya.” PLoS One 13 (5): e0197995.

 PubMed Web of Science ®Google Scholar

Boender, C., A. Malhorta, and S. Schuler. 2002. Measuring Women’s Empowerment as a Variable in International Development: Background Paper Prepared for the World Bank Workshop on Poverty and Gender: New Perspectives.

 Google Scholar

Alkire, S., R. Meinzen-Dick, A. Peterman, A. Quisumbing, G. Seymour, and A. Vaz. 2013. “The Women’s Empowerment in Agriculture Index.” World Development 52: 71–91.

 Web of Science ®Google Scholar

Sen, A. 1999. “Capability and Wellbeing.” Quality Life 30.

 Google Scholar

Iqbal, M. A., Q. Ping, U. I. Ahmed, and A. Nazir. 2015. “Determinants of Off-Farm Activity Participation Among Cotton Farmers in Punjab, Pakistan.” International Journal of Management Accounting and Economics 2 (7): 707–718.

 Google Scholar