(Not) in my backyard? : the impact of citizen initiatives on housing supply in Germany

Recently, the channels of citizen participation in Germany with regard to governmental processes are subject to change. Classical interaction between politicians and citizens still takes place. However, there are new tools for participation at hand as well. One interesting channel that attracted attention recently consists of citizen initiatives trying to obtain a referendum at a local level. These local referendums received much attention, even at the federal level in Germany. Interestingly, in a referendum on housing construction in a green area in Berlin, proponents were arguing that there is a lack of developable areas in terms of living space and a scarcity of housing areas. The shortage of housing supply became an apparent phenomenon in urbanized areas all over Germany. Therefore, it is worth to explore the interplay of local initiatives and the amount of provided housing areas. This paper contributes to this discussion by estimating the impact of local initiatives on the amount of approved and finished residential areas in Bavaria from 2003 until 2012. Since citizen initiatives were introduced in 1995 in Bavaria, it has become the federal state within Germany where most citizen initiatives took place. We use the introduction and the spatial expansion of initiatives as a natural experiment to obtain a causal effect of initiatives. We suppose two possible theoretical mechanisms how local initiatives affect the urban development process. First, the median voter in the respective municipality might be a homevoter (Fischel (2001)). Briefly, a homevoter is described as an inhabitant that allocated all his investments into his own house. Therefore, homevoters have an incentive to use local initiatives in order to maintain the existing conditions and avoid a potential loss in their home's value. A second mechanism might be the loss of reputation for local politicians with the presence of an initiative. Therefore, local governments that already experienced an initiative might be more hesitant with urban development policies in the future. We start our empirical assessment by thoroughly exploring the time frame where initiatives might have an impact on local land use. It seems that initiatives have an effect on local land use within five to nine years after they took place. Even though there is a negative trend of housing construction in Bavaria during our observational period, a local initiative causes a moderate negative dip of around 10% in finished residential area per capita. Interestingly enough, initiatives only influence the provision of residential areas but do not have any effect at all on the provision of non-residential areas. We consider this as evidence for our hypothesis, since voters are more directly affected by housing construction in their immediate neighborhood. Our results remain robust through a variety of robustness checks.