3
1) Introduction
Policy diffusion has recently attracted considerable academic interest. Scholars are
increasingly recognising the need to explicitly consider policy choices as interdependent.
Of course, the possibility that countries do not constitute independent observations has
been known for a long time in comparative politics, notably under the label “Galton’s
problem” (see e.g. Przeworski and Teune 1970: 51-53; Ross and Homer 1976). Until
recently, this was treated precisely as a “problem”, an annoyance that complicates
empirical analysis. An emerging policy diffusion literature (e.g. Meseguer 2004; Simmons
and Elkins 2004), however, is focusing precisely on the characteristics and consequences of
countries’ interdependencies. In this literature, diffusion is defined as a process where
choices are interdependent, i.e. where the choice of a government influences the choices
made by others and, conversely, the choice of a government is influenced by the choices
made by others.
The diffusion literature acknowledges that diffusion-like patterns, namely patterns of
“successive adoptions of a policy innovation” (Eyestone 1977: 441), can emerge from fully
independent decisions. The spread of policies can be driven not only by the
interdependencies among actors, but also by “internal determinants” (Berry and Berry
1990), “prerequisites” (Collier and Messick 1975), or “common contextual effects” (Van
den Bulte and Lilien 2001). This is sometimes considered as the “null hypothesis” against
which diffusion hypotheses are tested. However, since non-diffusion factors are
commonly found to matter along with diffusion mechanisms (see e.g. Simmons and Elkins
2004), which goes against the idea of a null hypothesis, we suggest that the term “spurious
diffusion” better captures the fact that a pattern may look like diffusion though it is not
driven by diffusion.
Convention