Social Capital: Experimental validation of survey measures

The social fabric, generally recognized as essential for economic and social transactions, is often referred to as Social Capital (SC). In this paper, we explore to what extent inexpensive survey data can be a substitute for more expensive experimental data as a metric of SC, using a cross-country design. We use data from two standard subject pools (located in Spain and France) and a mixed-method approach in the sense of presenting validated survey questions from the SC section of the latest wave of World Values Survey (WVS) to our participants, in addition to games for eliciting SC through actions and beliefs. Our data can be compared to publicly available WVS data at the relevant regional level as well as the national level. The main takeaway from our study is that SC measured by survey items consistently is higher in Spain than in France regardless of item and spatial resolution (nation, region, lab), whereas SC measured by choices and beliefs in incentivised games consistently is higher in France. This may confirm that there is reason for scepticism concerning the validity of survey measures in the context of social capital, not least since we, as opposed to in earlier studies, have data on group specific items used in the latest wave of WVS pertaining to trust in personal relations as well as more distant relations, all consistently pointing in the same direction regardless of spatial resolution. In this version of the paper we are concentrating on aggregates. Work remain to be done on the individual level.