Social capital and the effect of female representation: Expressive and instrumental voting

Academic and public discussion on female representation in politics has focused on whether and how such representation affects public policy.  The question has been asked whether the number of female representatives (so-called descriptive representation) results in women’s preferences being better accounted for (substantive representation).  The empirical evidence is at best mixed. 

Jottier and Heyndels (2011) demonstrate that social capital lowers the relative price of instrumental – as opposed to expressive – voting where voting behavior is seen as a mixture of both instrumental and expressive considerations.  Depending on context either instrumental or expressive considerations dominate.   Jottier and Heyndels (2011) find that the link between the quality of public policy and the election outcome is stronger to the extent that social capital is higher.  In other words: instrumental voting – made more attractive by the presence of social capital - is effective in bringing public policy in line with voters’ wishes.   If social capital increases the likelihood of instrumental voting, it will increase the link between voters’ preferences and public policy.  Public policy can then be expected to reflect this. 

Ashworth, Geys and Heyndels (2013) find that women tend to have a preference for higher levels of public expenditures.  Thus, in a context of pure instrumental voting, this preference is expected to be translated into actual policies.  The presence of larger shares of women in the electorate results in higher spending.  If expressive voting is more prominent we expect the translation of the electorate’s preferences into policy to be biased.  Politicians have less incentive to follow the voter.  As a result their identity becomes relevant for actual policy. 

Thus: controlling for the effect of female preferences in the population (captured by the share of women in the electorate) in a municipality high in social capital we expect to find that the presence of women in the council and/or government translates into significantly higher expenditures.   In municipalities low in social capital we expect to find the presence of women in the council and/or government to be irrelevant.

Using a panel from Flemish municipalities, this proposition is tested and is found to hold for women in government and, indeed, for women mayors, but there is no significant effect for women in parliament.  Given that it is only those in power who make the ultimate decisions we conclude that social capital does indeed increase the instrumental and decrease expressive effects.