This paper by Bauchet and Morduch does a pretty good job at separating two different sectors: that of the Small and Medium Enterprises (SME) and that of the microfinance sector. The authors say in the abstract:
We integrate evidence from three surveys in Bangladesh to show that, compared to microcredit customers, typical SME employees in Bangladesh are far more likely to be men, have more education and professional skills, and live in households that are notably less poor. SME jobs also require long work weeks, clashing with family responsibilities. The evidence from Bangladesh rejects the idea that SME finance more efficiently creates jobs for the population served by microcredit.
SME finance is gaining attention as a possible alternative to microcredit investment in the fight for poverty reduction, notably because SMEs provide employment on a much larger scale than microenterprises supported by microcredit. Our evidence suggests that in Bangladesh the two forms of support are complements, not substitutes.
A survey of employee of Bangladeshi SMEs shows that the typical employee is a young, educated male, whose household tends to be better off than the typical households of microcredit borrowers. Working in a small enterprise is hard: employees work 70 hours per week on average, for very low hourly wages. Men and women work similar hours, but women receive hourly wages one‐third lower than men on average. The long hours are not particularly friendly for workers with substantial family responsibilities.
The figure is taken from the paper.