A Reformulation of the Polygyny-Fertility Hypothesis

Article excerpt


Although the proportion of women in polygynous marriages has generally declined, over the last few years the effect of polygyny on fertility has drawn an extensive amount of research (Ahmed, 1986; Bean and Mineau, 1986; Bhatia, 1985; Borgerhoff Mulder, 1989; Garenne and van de Walle, 1989; Hem, 1992; Johnson and Elmi, 1989; Olusanya, 1971; Shaikh, Aziz and Chowdhury, 1987; Sichona, 1993; Ukaegbu, 1977). Recent evidence shows that Ghana, like most African countries, is experiencing a decline in the proportion of women who are in polygynous marriages. For example, the proportion of married women in polygynous marriages in Ghana decreased from 32.6 per cent in 1988 (GSS and IRD, 1989) to 27.7 per cent in. 1993 (GSS and MI, 1994), while that of women in polygynous marriages in Kenya decreased from 23.4 per cent in 1989 (NCPD and IRD/MI, 1989) to 19.5 per cent in 1993 (NCPD, CBS and MI, 1994). The rates for Ghana and Kenya are lower than those for Burkina Faso, Mali, Nigeria, Senegal and Togo, where over 40 per cent of women are in polygynous marriages (Carr and Way, 1994). The polygyny-fertility hypothesis states that women in polygynous marriages have lower fertility than women in monogamous marriages (Garenne and van de Walle, 1989; Muhsam, 1956). The burgeoning literature, however, has shown that the relationship between polygyny and fertility appears to be more complex than has been generally suggested (Borgerhoff Mulder, 1989; Olusanya, 1971).

Much of the literature that addresses the polygyny-fertility hypothesis has methodological and conceptual problems. Welch and Glick (1981) are concerned about the legislative prohibitions against the contract of plural marriages because they prevent women in polygynous marriages from reporting their marital status in either censuses or surveys for fear of sanctions. Analysis based on surveys may not be able to establish the precise relationship between polygyny and fertility. The validity of the polygyny-fertility hypothesis has been questioned by Olusanya (1971). One issue Olusanya (1971) raised is whether we can find families which have always been monogamous or polygynous. Marital status may change over time and confound the relationship between the type of marriage and fertility. In fact, women who have previously married may have had some of their children, or all of them, during the previous marriage or vice versa. Current marital status, therefore, may not have influenced the person's fertility in any significant way.

As a result of these fundamental questions, it has been argued that polygyny has probably lost its independent effect on fertility (Aborampah, 1987). Apart from the methodological problems associated with these studies, no systematic effort has been expended to conceptualize polygyny in order to estimate its effect on fertility in Ghana. This paper reformulates the polygyny-fertility hypothesis, taking into account the effect of previous marriage. That is, women in polygynous marriages are more likely to have lower fertility if they have previously married. We argue that in the absence of the effect of previous marriage, women in polygynous and monogamous marriages will have similar fertility.


The kinship system is central to understanding nuptiality patterns in Ghana. Individuals reckon descent through either the matrilineage (for the Akan ethnic group) or the patrilineage (for non-Akan ethnic groups). Divorce tends to be very frequent in matrilineal societies where the children belong to the woman's lineage and a low bride-price exists (Burch, 1983). In the patrilineal societies where bride-price is high and children belong to the man's lineage, divorce tends to be less frequent (Burch, 1983). Because the perpetuity of the lineage depends on the fertility of the members of the lineage, marriage is viewed as a relationship between two lineages and not just an individual affair (Assimeng, 1981). …


An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.