Female Genital Cutting and Long-Term Health Consequences – Nationally Representative Estimates across 13 Countries

Abstract

Using cross-sectional data from 13 African countries, I compare long-term health outcomes across cut and uncut women. This study is the first to use nationally representative data. Consistent with medical research, no evidence of general health impairments or decreased fertility induced by female genital cutting (FGC) is found; rather cut women have more children. The most pronounced long-term health impairments are a 24 per cent increase in the odds of contracting sexually transmitted infections and a 15 per cent increase in genital problems. Concomitantly, the odds that a cut woman will marry before an uncut woman are 13 per cent.

Acknowledgements

I thank Mark Rosenzweig, Jean-Louis Arcand, Arjun Bedi, Lore Vandewalle, two anonymous referees and the editor, Howard White, as well as seminar participants at the German Development Economics Conference in Bonn, the Graduate Institute, Geneva and the Journées de Microéconomie Appliquée in Brest for their valuable comments on this article. All remaining sins of omission or commission are my own.

Notes

1. It has to be noted that WHO adopts the term female genital mutilation (FGM), while I use female genital cutting (FGC). Yet, this difference refers only to the terminology, it does not apply for the definition of the practice.

2. For more detailed information, see WHO (2010).

3. Comprehensive anthropological reviews about FGC are found in the two collections of articles by Shell-Duncan and Hernlund (2000, 2007).

4. Replacing this variable by completed years of education does not affect the results.

5. The DHS wealth index was constructed as a measure of economic status that consists of easily measurable components and depicts a more permanent situation than measures of income/consumption. The index results from a principal component analysis that includes existing data about (1) the type of flooring, (2) refrigerator, (3) water supply and sanitation facilities, (4v) type of vehicle, (5) ownership of agricultural land, (6) persons per sleeping room, (7) electricity, (8) radio, television and telephone, (9) domestic servant, and (10) country-specific items. Based on the index, households are classified into five wealth categories, which are represented in the wealth quintiles used in the analysis at hand. Rutstein and Johnson (2004) present detailed information about the philosophy and construction of the DHS wealth index.

6. For the sake of brevity, coefficient estimates associated with religious adherence are not presented in Table 6.

7. Detailed results of coefficients associated with religion and ethnicity are made available by the author upon request.

8. The hazard ratio is derived from the coefficient estimate in similar fashion as the odds ratio, namely HR= exp(), where k = 1,2, …, K, given the other explanatory variables are held fixed. Also the interpretation of the hazard ratio is analogous to the odds ratio with a hazard ratio greater than one associated with FGC status indicating that cut women marry earlier.

9. Results are made available by the author upon request.