Women’s participation and cost efficiency in microfinance institutions: a Sub-Saharan study

Figures & data

Table 1. Variable definition and measurement.

	Measurement	Definition
Panel A: Dependent variable
TC: total cost	Administrative expense + interest expense + Other financial expense + Operating expense + other operating expense + Non – operating expense + Personnel expense (Abdulai & Tewari 2016; MIX, 2021; Srairi 2010)	Interest expense = interest expense on deposits + interest expense on borrowings
Panel B: Independent variables
Input prices
Labor: Price of labor	Personnel expenses divided by total number of personnel (MIX, 2021; Lu et al., 2018)	Cost paid as a salary or wage to mobilize employees’ effort
Fund: Price of fund	Interest expenses (interest paid) divided by total deposits and borrowings (MIX, 2021; Lu et al., 2018)	Interest expense = interest expense on deposits + interest expense on borrowings
PPC: Price of physical capital	Operating expense divided by net fixed assets	Operating expense includes expenses not related to personnel expenses, depreciation, amortization and administrative expenses & Net fixed assets are tangible assets net of accumulated depreciation (MIX, 2021).
Output quantities
Loan: Net loan portfolio	Value of loan portfolio net of impairment loss allowance and unearned income and discount (when applicable) (MIX, 2021; Lu et al., 2018).	Describes efficiency of MFI in minimizing adverse outcomes
OEA : Other earning Assets	Total of all other assets	Includes receivables, long-term investment, inventories, intangible assets (MIX, 2021).
Panel C: Determinants of cost inefficiency
Firm-specific variables
firmsize: Firm size	Natural logarithm of total assets (Lu et al., 2018)	Total assets (MIX, 2021)
ROA: Profitability	ROA (Abdulai & Tewari, 2016; Srairi, 2010)	Return on asset ratio (MIX, 2021)
Boardg: Broad gender diversity	Percentage of female as board members (MIX, 2021; Mori et al., 2015)	Gender diversity in corporate governance
Mgmtg: Management gender diversity	Percentage of female in management (MIX, 2021; Mori et al., 2015)
Femaleb: Female borrower	Percent of female as active borrowers (Abdulai & Tewari 2016; Tadesse Abate et al., 2014)	Efficiency in outreach (depth of outreach)
Femalestaf: female personnel	Number of female in personnel (MIX, 2021)	Female employees acting as regular personnel in microfinance institution.
Femaleof: Female loan official	Percentage of female loan officer (MIX, 2021)	Female employees acting as loan officer in microfinance institution
Panel D: Controlling and additional variables
Time trend in year (T)	Labeled as “1” if 2009, “2” if 2010, “3” if 2011 … and “10” if 2018.	Added as dummy variable (Lu et al., 2018)
Income: Income category	1 = Upper middle income 2 = Lower middle income 3 = Low income	MFI’s country income category (dummy variable) (World Bank, 2022).
Type: Type of MFIs	1 = NBFI, 2 = Bank, 3 = NGO, 4 = Credit union/cooperative (Tilahun, 2021)	Legal status of MFI (MIX, 2021)

Table 2. Summary statistic for cost efficiency scores.

	Cost efficiency scores (%)
	Mean	Std.Dev
Panel A: Types of MFI
NBFI	14.22	0.131
Bank	9.40	0.065
NGO	14.81	0.092
Credit union/cooperative	11.66	0.069
Overall efficiency	13.07	0.099
Panel B: Income category of MFI’s country
Upper-middle income	9.56	0.043
Lower-middle income	13.62	0.110
Low income	12.46	0.083
Overall efficiency	13.07	0.099
Panel C: Time trend (2009–2018)
2009	12.50	0.109
2010	12.70	0.105
2011	13.25	0.106
2012	12.77	0.068
2013	12.96	0.098
2014	13.31	0.105
2015	13.29	0.104
2016	13.37	0.111
2017	13.15	0.083
2018	13.59	0.089
Overall efficiency	13.07	0.099
Panel D: Top Three Countries with most cost efficient MFIs
	Score (%)	Rank
Angola	87%	1
Mali	24%	2
Congo Republic	17%	3
Egypt	17%	3
Panel E: Countries with smallest cost efficient MFIs
Uganda	8%	3
Malawi	7%	2
Sudan *	2%	1

Source: Authors’ computation.

*MFIs operating in Sudan have scored the smallest cost efficiency result, which means that they are efficient only for 2%.

Table 3. Stochastic frontier regression results.

Dependent variable – total cost		Cost inefficiency
Panel A: Input, Outputs and Cross terms
Notation	Parameter	Coef	z-value
ln(Labor/PPC)	${\alpha }_1$	0.5622	3.17***
ln(Fund/PPC)	${\alpha }_2$	15.1082	2.55**
ln(Loan)	${\alpha }_3$	0.2265	2.01**
ln(OEA)	${\alpha }_4$	0.0108	0.11
½ ln(Labor/PPC)^2	${\alpha }_5$	0.1024	11.82***
½ ln(Fund/PPC)^2	${\alpha }_6$	−20.2819	−2.97***
½ ln(Loan)^2	${\alpha }_7$	0.0514	11.81***
½ ln(OEA)^2	${\alpha }_8$	0.0079	1.55
ln(Labor/PPC)* ln(Fund/PPC)	${\alpha }_9$	−2.3964	−7.73***
ln(Labor/PPC)* ln(Loan)	${\alpha }_{10}$	−0.0523	−4.17***
ln(Labor/PPC)* ln(OEA)	${\alpha }_{11}$	0.0052	0.64
ln(Fund/PPC)* ln(Loan)	${\alpha }_{12}$	−0.0601	−0.14
ln(Fund/PPC)* ln(OEA)	${\alpha }_{13}$	0.7328	2.84***
ln(Loan)* ln(OEA)	${\alpha }_{14}$	−0.0082	−1.24
ln(Labor/PPC)*T	${\alpha }_{15}$	0.0135	3.34***
ln(Fund/PPC)*T	${\alpha }_{16}$	0.2304	1.60
ln(Loan)*T	${\alpha }_{17}$	−0.0058	−1.35
ln(OEA)*T	${\alpha }_{18}$	0.0007	0.20
½(T)^2	${\alpha }_{19}$	−0.0018	−0.50
T	${\alpha }_{20}$	0.0904	1.70*
Constant	${\alpha }_0$	0.4095	0.27
Wald Chi-square	2043.13		000***
Sigma squared	0.4216		8.03***
Gamma (γ)	0.7694		24.54***
Eta (η)	0.0337		000***
Log-likelihood function	−456.7440
Number of observation	930
Number of group	128

***p < 1%, **p < 5%, *p < 10%.

Source: Authors’ regression model outputs.

Table 4. Factors affecting cost inefficiency of MFIs in SSA.

Dependent variable: Cost inefficiency		Model 1		Model 2		Model 3
Independent variable	Parameter	Coef	t-value	Coef	t-value	Coef	t-value
lnfirmsize	${\beta }_1$	0.066	7.12***	0.085	9.79***
lnROA	${\beta }_2$	0.030	0.81	0.040	1.06
lnCostperbor	${\beta }_3$	0.057	3.70***	0.056	3.64***
Boardg	${\beta }_4$	−0.171	−2.73***			−0.223	−3.28***
Mgmtg	${\beta }_5$	0.079	1.16			0.168	2.32**
Femalebr	${\beta }_6$	0.030	0.42			−0.329	−4.83***
lnFemalestaff	${\beta }_7$	0.063	5.70***			0.098	8.64***
Femaleof	${\beta }_8$	−0.212	−2.81***			−0.252	−3.16***
Upper_middle_income	${\beta }_9$	−0.072	−0.57	0.006	0.05
Lower_middle_income	${\beta }_9$	−0.179	−5.41***	−0.151	−4.60***
NBFI	${\beta }_{10}$	−0.241	−5.19***	−0.233	−4.95***
Bank	${\beta }_{10}$	0.002	0.04	0.003	0.06
NGO	${\beta }_{10}$	−0.229	−4.83***	−0.220	−4.75***
Constant	${\beta }_0$	1.051	4.36***	0.912	3.90***	2.37	21.28***
Number of obs.		930		930		930
Chi-square		313.66***		253.64***		121.89***
LR test of the one-sided error		3.68**		11.64 ***		1.66 *
Log likelihood function		−601.44		−622.95		−683.05

*** p < 1%, ** p < 5%, * p < 10%.

Source: Authors’ regression model outputs.

Abdulai, A., & Tewari, D. D. (2016). Efficiency of microfinance institutions in Sub – Saharan Africa: A stochastic frontier approach. Ghana Journal of Development Studies, 13(2), 1–13. https://doi.org/10.4314/gjds.v13i2.7

 Google Scholar

MIX (2021). World Bank Group. 11 5, https://databank.worldbank.org/source/mix-market#.

 Google Scholar

Srairi, S. A. (2010). Cost and profit efficiency of conventional and Islamic banks in GCC countries. Journal of Productivity Analysis, 34(1), 45–62. https://doi.org/10.1007/s11123-009-0161-7

 Web of Science ®Google Scholar

Lu, Y. F., Gan, C., Hu, B., Toh, M. Y., & Cohen, D. A. (2018). Bank efficiency in New Zealand: A stochastic frontier approach. New Zealand Economic Papers, 53(2), 166–183. https://doi.org/10.1080/00779954.2018.1455728

 Google Scholar

Mori, N., Golesorkhi, S., Randø, T., & Hermes, N. (2015). Board composition and outreach performance of microfinance institutions. Strategic Change, 24(1), 99–113. https://doi.org/10.1002/jsc.2000

 Google Scholar

Tadesse Abate, G., Borzaga, C., & Getnet, K. (2014). Cost-efficiency and outreach of microfinance institutions: Tradeoffs and the role of ownership. Journal of International Development, 26(6), 923–932. https://doi.org/10.1002/jid.2981

 Web of Science ®Google Scholar

World Bank. (2022, May 5). World Bank Group. https://datahelpdesk.worldbank.org/knowledgebase/articles/906519-world-bank-country-and-lending-groups.

 Google Scholar

Tilahun, A. T. (2021). Do location and legal status matter in microfinance institutions’ performance? Evidence from sub-Saharan Africa. Development in Practice, 31(3), 404–420. https://doi.org/10.1080/09614524.2020.1853060

 Web of Science ®Google Scholar

Data availability statement

We used data available on MIX-market database accessed through https://databank.worldbank.org/source/mix-market#.