It has been empirically shown that structural holes in social networks enable potential large
benefits to those individuals who bridge them (Burt, 2004). The pioneering paper Goyal and
Vega-Redondo (2007) offers a new incentives based explanation of this phenomenon. But the main
equilibrium network of their model does not display a basic empirical regularity: the architecture
of social networks is characterized by the existence of densely linked communities loosely connected
to one another (Granovetter, 1983). This paper analyzes the conditions under which agents who
benefit from bridging structural holes can be sustained in equilibrium networks constituted by
densely linked groups.